SK hynix Wraps up Dev Work on High-End PCB01 PCIe 5.0 SSD for OEMs, Launching Later This Year

SK hynix early in Friday announced that the company has finished the development of it's PCB01 PCIe Gen5 SSD, the company's forthcoming high-end SSD for OEMs. Based on the company's new Alistar platform, the PCB01 is designed to deliver chart-topping performance for client machines. And, as a sign of the times, SK hynix is positioning the PCB01 for AI PCs, looking to synergize with the overall industry interest in anything and everything AI.

The bare, OEM-focused drives have previously been shown off by SK hynix, and make no attempt to hide what's under the hood. The PCB01 relies on SK hynix's Alistar controller, which features a PCIe Gen5 x4 host interface on the front end and eight NAND channels on the back end, placing it solidly in the realm of high-end SSDs. Paired with the Alistar controller is the company's latest 238-layer TLC NAND (H25T1TD48C & H25T2TD88C), which offers a maximum transfer speed of 2400 MT/second. Being that this is a high-end client SSD, there's also a DRAM chip on board, though the company isn't disclosing its capacity.

As with other high-end PCIe 5.0 client SSDs, SK hynix is planning on hitting peak read speeds of up to 14GB/second on the drive, while peak sequential write speeds should top 12GB/second (with pSLC caching, of course) – performance figures well within the realm of possibility for an 8 channel drive. As for random performance, at Computex the company was telling attendees that the drives should be able to sustain 4K random read and write rates of 2 million IOPS, which is very high as well. The SSDs are also said to consume up to 30% less power than 'predecessors,' according to SK hynix, though the company didn't elaborate on that figure. Typically in the storage industry, energy figures are based on iso-performance (rather than peak performance) – essentially measuring energy efficiency per bit rather than toal power consumption – and that is likely the case here as well.

At least initially, SK Hynix plans to release its PCB01 in three capacities – 512 GB, 1 TB, and 2 TB. The company has previously disclosed that their 238L TLC NAND has a capacity of 512Gbit, so these are typical capacity figures for single-sided drives. And while the focus of the company's press release this week was on OEM drives, this is the same controller and NAND that is also going into the company's previously-teased retail Platinum P51 SSD, so this week's reveal offers a bit more detail into what to expect from that drive family as well.

Specs aside, Ahn Hyun, the Head of the N-S Committee at SK hynix, said that multiple global CPU providers for on-device AI PCs are seeking collaboration for the compatibility validation process, which is underway, so expect PCB01 drives inside PCs in this back-to-school and holiday seasons.

"We will work towards enhancing our leadership as the global top AI memory provider also in the NAND solution space by successfully completing the customer validation and mass production of PCB01, which will be in the limelight," Ahn Hyun said.

SSDs

HighPoint Updates NVMe RAID Cards for PCIe 5.0: 50 GBps+ Direct-Attached SSD Storage

HighPoint Technologies has updated their NVMe switch and RAID solutions with PCIe 5.0, and supporting up to eight NVMe drives. The new HighPoint Rocket 1600 (switch add-in card) and 7600 series (RAID adapters) are the successors to the SSD SSD7500 series adapter cards introduced in 2020. Similar to its predecessors, the new Rocket series cards are also based on a Broadcom PCIe switch (PEX 89048). The Rocket 7600 series runs the RAID stack on the integrated ARM processor (dual-core Cortex A15)

The PEX 89048 supports up to 48 PCIe 5.0 lanes, out of which 16 are dedicated to the host connection in the Rocket adapters. The use of a true PCIe switch means that the product doesn't rely on PCIe lane bifurcation support in the host platform.

HighPoint's Gen 5 stack currently has two products each in the switch and RAID lineups - an add-in card with support for M.2 drives, and a RAID adapter with four 5.0 x8 SFF-TA-1016 (Mini Cool Edge IO or MCIO) connectors for use with backplanes / setups involving U.2 / U.3 / EDSFF drives.

The RAID adapters require HighPoint's drivers (available for Linux, macOS, and Windows), and supports RAID 0, RAID 1, and RAID 10 arrays. On the other hand, the AIC requires no custom drivers. RAID configurations with the AIC will need to be handled by software running on the host OS. On the hardware side, all members of the Rocket series come with an external power connector (as the solution can consume upwards of 75W) and integrate a heatsink. The M.2 version is actively cooled, as the drives are housed within the full-height / full-length cards.

The solution can theoretically support up to 64 GBps of throughput, but real-world performance is limited to around 56 GBps using Gen 5 drives. It must be noted that even Gen 4 drives can take advantage of the new platform and deliver better performance with the new Rocket series compared to the older SSD7500 series.

The cards are shipping now, with pricing ranging from $1500 (add-in card) to $2000 (RAID adapters). HighPoint is not alone in targeting this HEDT / workstation market. Sabrent has been teasing their Apex Gen 5.0 x16 solution involving eight M.2 SSDs for a few months now (involving a Microchip PCIe switch. Until that solution comes to the market, HighPoint appears to be the only game in town for workstation users requiring access to direct-attached storage capable of delivering 50 GBps+ speeds.

SSDs

Samsung Validates LPDDR5X Running at 10.7 GT/sec with MediaTek's Dimensity 9400 SoC

Samsung has successfully validated its new LPDDR5X-10700 memory with MediaTek's upcoming Dimensity platform. At present, 10.7 GT/s is the highest performing speed grade of LPDDR5X DRAM slated to be released this year, so the upcoming Dimensity 9400 system-on-chip will get the highest memory bandwidth available for a mobile application processor.

The verification process involved Samsung's 16 GB LPDDR5X package and MediaTek's soon-to-be-announced Dimensity 9400 SoC for high-end 5G smartphones. Usage of LPDDR5X-10700 provides a memory bandwidth of 85.6 GB/second over a 64-bit interface, which will be available for bandwidth-hungry applications like graphics and generative AI.

"Working together with Samsung Electronics has made it possible for MediaTek's next-generation Dimensity chipset to become the world's first to be validated at LPDDR5X operating speeds up to 10.7Gbps, enabling upcoming devices to deliver AI functionality and mobile performance at a level we have never seen before," said JC Hsu, Corporate Senior Vice President at MediaTek. "This updated architecture will make it easier for developers and users to leverage more AI capabilities and take advantage of more features with less impact on battery life."

Samsung's LPDDR5X 10.7 GT/s memory in made on the company's 12nm-class DRAM process technology and is said to provide a more than 25% improvement in power efficiency over previous-generation LPDDR5X, in addition to extra performance. This will positively affect improved user experience, including enhanced on-device AI capabilities, such as faster voice-to-text conversion, and better quality graphics.

Overall, the two companies completed this process in just three months. Though it remains to be seen when smartphones based on the Dimensity 9400 application processor and LPDDR5X memory are set to be available on the market, as MediaTek has not yet even formally announced the SoC itself.

"Through our strategic cooperation with MediaTek, Samsung has verified the industry's fastest LPDDR5X DRAM that is poised to lead the AI smartphone market," said YongCheol Bae, Executive Vice President of Memory Product Planning at Samsung Electronics. "Samsung will continue to innovate through active collaboration with customers and provide optimum solutions for the on-device AI era."

Memory

Crucial P310 NVMe SSD Unveiled: Micron's Play in the M.2 2230 Market

Hand-held gaming consoles based on notebook platforms (such as the Valve SteamDeck, ASUS ROG Ally, and the MSI Claw) are one of the fastest growing segments in the PC gaming market. The form-factor of such systems has created a demand for M.2 2230 NVMe SSDs. Almost all vendors have a play in this market, and even Micron has OEM SSDs (such as the Micron 2400, 2550, and 2500 series) in this form-factor. Crucial has strangely not had an offering with its own brand name to target this segment, but that changes today with the launch of the Crucial P310 NVMe SSD.

The Crucial P310 is a family of M.2 2230 PCIe Gen4 NVMe SSDs boasting class-leading read/write speeds of 7.1 GBps and 6 GBps. The family currently has two capacity points - 1 TB and 2 TB. Micron claims that the use of its 232L 3D NAND and Phison's latest E27T DRAM-less controller (fabricated in TSMC's 12nm process) help in reducing power consumption under active use compared to the competition - directly translating to better battery life for the primary use-case involving gaming handheld consoles.

Based on the specifications, it appears that the drives are using 232L 3D QLC. Compared to the recently-released Micron 2550 SSD series in the same form-factor, a swap in the controller has enabled some improvements in both power efficiency and performance. The other specifications are summarized in the table below.

Crucial P310 SSD Specifications
Capacity	2 TB	1 TB
Controller	Phison E27T (DRAM-less)
NAND Flash	Micron 232L 3D QLC NAND
Form-Factor, Interface	Single-Sided M.2-2230 PCIe 4.0 x4, NVMe
Sequential Read	7100 MB/s
Sequential Write	6000 MB/s
Random Read IOPS	1 M
Random Write IOPS	1.2 M
SLC Caching	Yes
TCG Pyrite Encryption	Yes
Warranty	5 Years
Write Endurance	440 TBW 0.12 DWPD	220 TBW 0.12 DWPD
MSRP	$215	$115

The power efficiency, cost, and capacity points are plus points for the Crucial P310 family. However, the endurance ratings are quite low. Gaming workloads are inherently read-heavy, and this may not be a concern for the average consumer. However, a 0.12 DWPD rating may turn out to be a negative aspect when compared against the competition's 0.33 DWPD offerings in the same segment.

Storage

Samsung Joins The 60 TB SSD Club, Looking Forward To 120 TB Drives

Multiple companies offer high-capacity SSDs, but until recently, only one company offered high-performance 60 TB-class drives with a PCIe interface: Solidigm. As our colleagues from Blocks & Files discovered, Samsung quietly rolled out its BM1743 61.44 TB solid-state drive in mid-June and now envisions 120 TB-class SSDs based on the same platform.

Samsung's BM1743 61.44 TB features a proprietary controller and relies on Samsung's 7^th Generation V-NAND (3D NAND) QLC memory. Moreover, Samsung believes that its 7^th Gen V-NAND 'has the potential to accommodate up to 122.88 TB,' 

Samsung plans to offer the BM1743 in two form factors: U.2 for PCIe 4.0 x4 to address traditional servers and E3.S for PCIe 5.0 x4 interfaces to address machines designed to offer maximum storage density. BM1743 can address various applications, including AI training and inference, content delivery networks, and read-intensive workloads. To that end, its write endurance is 0.26 drive writes per day (DWPD) over five years.

Regarding performance, Samsung's BM1743 is hardly a champion compared to high-end drives for gaming machines and workstations. The drive can sustainably achieve sequential read speeds of 7,200 MB/s and write speeds of 2,000 MB/s. It can handle up to 1.6 million 4K random reads and 110,000 4K random writes for random operations.

Power consumption details for the BM1743 have not been disclosed, though it is expected to be high. Meanwhile, the drive's key selling point is its massive storage density, which likely outweighs concerns over its absolute power efficiency for intended applications, as a 60 TB SSD still consumes less than multiple storage devices offering similar capacity and performance.

As noted above, Samsung's BM1743 61.44 TB faces limited competition in the market, so its price will be quite high. For example, Solidigm's D5-P5336 61.44 TB SSD costs $6,905. Other companies, such as Kioxia, Micron, and SK Hynix, have not yet introduced their 60TB-class SSDs, which gives Samsung and Solidigm an edge for now.

UPDATE 7/25: We removed mention of Western Digital's 60 TB-class SSDs, as the company does not currently list any such drives on their website

SSDs

JEDEC Plans LPDDR6-Based CAMM, DDR5 MRDIMM Specifications

Following a relative lull in the desktop memory industry in the previous decade, the past few years have seen a flurry of new memory standards and form factors enter development. Joining the traditional DIMM/SO-DIMM form factors, we've seen the introduction of space-efficient DDR5 CAMM2s, their LPDDR5-based counterpart the LPCAMM2, and the high-clockspeed optimized CUDIMM. But JEDEC, the industry organization behind these efforts, is not done there. In a press release sent out at the start of the week, the group announced that it is working on standards for DDR5 Multiplexed Rank DIMMs (MRDIMM) for servers, as well as an updated LPCAMM standard to go with next-generation LPDDR6 memory.

Just last week Micron introduced the industry's first DDR5 MRDIMMs, which are timed to launch alongside Intel's Xeon 6 server platforms. But while Intel and its partners are moving full steam ahead on MRDIMMs, the MRDIMM specification has not been fully ratified by JEDEC itself. All told, it's not unusual to see Intel pushing the envelope here on new memory technologies (the company is big enough to bootstrap its own ecosystem). But as MRDIMMs are ultimately meant to be more than just a tool for Intel, a proper industry standard is still needed – even if that takes a bit longer.

Under the hood, MRDIMMs continue to use DDR5 components, form-factor, pinout, SPD, power management ICs (PMICs), and thermal sensors. The major change with the technology is the introduction of multiplexing, which combines multiple data signals over a single channel. The MRDIMM standard also adds RCD/DB logic in a bid to boost performance, increase capacity of memory modules up to 256 GB (for now), shrink latencies, and reduce power consumption of high-end memory subsystems. And, perhaps key to MRDIMM adoption, the standard is being implemented as a backwards-compatible extension to traditional DDR5 RDIMMs, meaning that MRDIMM-capable servers can use either RDIMMs or MRDIMMs, depending on how the operator opts to configure the system.

The MRDIMM standard aims to double the peak bandwidth to 12.8 Gbps, increasing pin speed and supporting more than two ranks. Additionally, a "Tall MRDIMM" form factor is in the works (and pictured above), which is designed to allow for higher capacity DIMMs by providing more area for laying down memory chips. Currently, ultra high capacity DIMMs require using expensive, multi-layer DRAM packages that use through-silicon vias (3DS packaging) to attach the individual DRAM dies; a Tall MRDIMM, on the other hand, can just use a larger number of commodity DRAM chips. Overall, the Tall MRDIMM form factor enables twice the number of DRAM single-die packages on the DIMM.

Meanwhile, this week's announcement from JEDEC offers the first significant insight into what to expect from LPDDR6 CAMMs. And despite LPDDR5 CAMMs having barely made it out the door, some significant shifts with LPDDR6 itself means that JEDEC will need to make some major changes to the CAMM standard to accommodate the newer memory type.

JEDEC Presentation: The CAMM2 Journey and Future Potential

Besides the higher memory clockspeeds allowed by LPDDR6 – JEDEC is targeting data transfer rates of 14.4 GT/s and higher – the new memory form-factor will also incorporate an altogether new connector array. This is to accommodate LPDDR6's wider memory bus, which sees the channel width of an individual memory chip grow from 16-bits wide to 24-bits wide. As a result, the current LPCAMM design, which is intended to match the PC standard of a cumulative 128-bit (16x8) design needs to be reconfigured to match LPDDR6's alterations.

Ultimately, JEDEC is targeting a 24-bit subhannel/48-bit channel design, which will result in a 192-bit wide LPCAMM. While the LPCAMM connector itself is set to grow from 14 rows of pins to possibly as high as 20. New memory technologies typically require new DIMMs to begin with, so it's important to clarify that this is not unexpected, but at th... Memory

Micron Ships Denser & Faster 276 Layer TLC NAND, Arriving First In Micron 2650 Client SSDs

Micron on Tuesday announced that the company has begun shipping its 9th Generation (G9) 276 layer TLC NAND. The next generation of NAND from the prolific memory maker, Micron's latest NAND is designed to further push the envelope on TLC NAND performance, offering significant density and performance improvements over its existing NAND technology.

Micron's G9 TLC NAND memory features 276 active layers, which is up from 232-layers in case of Micron's previous generation TLC NAND. At this point the company is being light on technical details in their official material. However in a brief interview with Blocks & Files, the company confirmed that their 276L NAND still uses a six plane architecture, which was first introduced with the 232L generation. At this point we're assuming Micron is also string-stacking two decks of NAND together, as they have been for the past couple of generations, which means we're looking at 138 layer decks.

Micron TLC NAND Flash Memory
	276L	232L (B58R)	176L (B47R)
Layers	276	232	176
Decks	2 (x138)?	2 (x116)	2 (x88)
Die Capacity	1 Tbit	1 Tbit	512 Gbit
Die Size (mm2)	~48.9mm2	~70.1mm2	~49.8mm2
Density (Gbit/mm2)	~21	14.6	10.3
I/O Speed	3.6 GT/s (ONFi 5.1)	2.4 GT/s (ONFi 5.0)	1.6 GT/s (ONFI 4.2)
Planes	6	6	4
CuA / PuC	Yes	Yes	Yes

On the density front, Micron told Blocks & Files that they have improved their NAND density by 44% over their 232L generation. Which, given what we know about that generation, would put the density at around 21 Gbit/mm². Or for a 1Tbit die of TLC NAND, that works out to a die size of roughly 48.9mm², comparable to the die size of a 512Gbit TLC die from Micron's older 176L NAND.

Besides improving density, the other big push with Micron's newest generation of NAND was further improving its throughput. While the company's 232L NAND was built against the ONFi 5.0 specification, which topped out at transfer rates of 2400 MT/sec, their new 276L NAND can hit 3600 MT/sec, which is consistent with the ONFi 5.1 spec.

Meanwhile, the eagle-eyed will likely also pick up on Micron's ninth-generation/G9 branding, which is new to the company. Micron's has not previously used this kind of generational branding for their NAND, which up until now has simply been identified by its layer count (and before the 3D era, its feature size). Internally, this is believed to be Micron's 7th generation 3D NAND architecture. However, taking a page from the logic fab industry, Micron seems to be branding it as ninth-generation in order to keep generational parity with its competitors, who are preparing their own 8th/9th generation NAND (and thus cliam that they are the first NAND maker to ship 9th gen NAND).

And while this NAND will eventually end up in ... SSDs

Samsung Validates LPDDR5X Running at 10.7 GT/sec with MediaTek's Dimensity 9400 SoC

Samsung has successfully validated its new LPDDR5X-10700 memory with MediaTek's upcoming Dimensity platform. At present, 10.7 GT/s is the highest performing speed grade of LPDDR5X DRAM slated to be released this year, so the upcoming Dimensity 9400 system-on-chip will get the highest memory bandwidth available for a mobile application processor.

The verification process involved Samsung's 16 GB LPDDR5X package and MediaTek's soon-to-be-announced Dimensity 9400 SoC for high-end 5G smartphones. Usage of LPDDR5X-10700 provides a memory bandwidth of 85.6 GB/second over a 64-bit interface, which will be available for bandwidth-hungry applications like graphics and generative AI.

"Working together with Samsung Electronics has made it possible for MediaTek's next-generation Dimensity chipset to become the world's first to be validated at LPDDR5X operating speeds up to 10.7Gbps, enabling upcoming devices to deliver AI functionality and mobile performance at a level we have never seen before," said JC Hsu, Corporate Senior Vice President at MediaTek. "This updated architecture will make it easier for developers and users to leverage more AI capabilities and take advantage of more features with less impact on battery life."

Samsung's LPDDR5X 10.7 GT/s memory in made on the company's 12nm-class DRAM process technology and is said to provide a more than 25% improvement in power efficiency over previous-generation LPDDR5X, in addition to extra performance. This will positively affect improved user experience, including enhanced on-device AI capabilities, such as faster voice-to-text conversion, and better quality graphics.

Overall, the two companies completed this process in just three months. Though it remains to be seen when smartphones based on the Dimensity 9400 application processor and LPDDR5X memory are set to be available on the market, as MediaTek has not yet even formally announced the SoC itself.

"Through our strategic cooperation with MediaTek, Samsung has verified the industry's fastest LPDDR5X DRAM that is poised to lead the AI smartphone market," said YongCheol Bae, Executive Vice President of Memory Product Planning at Samsung Electronics. "Samsung will continue to innovate through active collaboration with customers and provide optimum solutions for the on-device AI era."

Memory

Intel Addresses Desktop Raptor Lake Instability Issues: Faults Excessive Voltage from Microcode, Fix Coming in August

In what started last year as a handful of reports about instability with Intel's Raptor Lake desktop chips has, over the last several months, grown into a much larger saga. Facing their biggest client chip instability impediment in decades, Intel has been under increasing pressure to figure out the root cause of the issue and fix it, as claims of damaged chips have stacked up and rumors have swirled amidst the silence from Intel. But, at long last, it looks like Intel's latest saga is about to reach its end, as today the company has announced that they've found the cause of the issue, and will be rolling out a microcode fix next month to resolve it.

Officially, Intel has been working to identify the cause of desktop Raptor Lake’s instability issues since at least February of this year, if not sooner. In the interim they have discovered a couple of correlating factors – telling motherboard vendors to stop using ridiculous power settings for their out-of-the-box configurations, and finding a voltage-related bug in Enhanced Thermal Velocity Boost (eTVB) – but neither factor was the smoking gun that set all of this into motion. All of which had left Intel to continue searching for the root cause in private, and lots of awkward silence to fill the gaps in the public.

But it looks like Intel’s search has finally come to an end – even if Intel isn’t putting the smoking gun on public display quite yet. According to a fresh update posted to the company’s community website, Intel has determined the root cause at last, and has a fix in the works.

Per the company’s announcement, Intel has tracked down the cause of the instability issue to “elevated operating voltages”, that at its heart, stems from a flawed algorithm in Intel’s microcode that requested the wrong voltage. Consequently, Intel will be able to resolve the issue through a new microcode update, which pending validation, is expected to be released in the middle of August.

Based on extensive analysis of Intel Core 13th/14th Gen desktop processors returned to us due to instability issues, we have determined that elevated operating voltage is causing instability issues in some 13th/14th Gen desktop processors. Our analysis of returned processors confirms that the elevated operating voltage is stemming from a microcode algorithm resulting in incorrect voltage requests to the processor.

Intel is delivering a microcode patch which addresses the root cause of exposure to elevated voltages. We are continuing validation to ensure that scenarios of instability reported to Intel regarding its Core 13th/14th Gen desktop processors are addressed. Intel is currently targeting mid-August for patch release to partners following full validation.

Intel is committed to making this right with our customers, and we continue asking any customers currently experiencing instability issues on their Intel Core 13th/14th Gen desktop processors reach out to Intel Customer Support for further assistance.
-Intel Community Post

And while there’s nothing good for Intel about Raptor Lake’s instability issues or the need to fix them, that the problem can be ascribed to (or at least fixed by) microcode is about the best possible outcome the company could hope for. Across the full spectrum of potential causes, microcode is the easiest to fix at scale – microcode updates are already distributed through OS updates, and all chips of a given stepping (millions in all) run the same microcode. Even a motherboard BIOS-related issue would be much harder to fix given the vast number of different boards out there, never mind a true hardware flaw that would require Intel to replace even more chips than they already have.

Still, we’d also be remiss if we didn’t note that microcode is regularly used to paper over issues further down in the processor, as we’ve most famously seen with the Meltdown/Spectre fixes several years ago. So while Intel is publicly attributing the issue to microcode bugs, there are several more layers to the onion that is modern CPUs that could be playing a part. In that respect, a microcode fix grants the least amoun... CPUs

Samsung Joins The 60 TB SSD Club, Looking Forward To 120 TB Drives

Multiple companies offer high-capacity SSDs, but until recently, only one company offered high-performance 60 TB-class drives with a PCIe interface: Solidigm. As our colleagues from Blocks & Files discovered, Samsung quietly rolled out its BM1743 61.44 TB solid-state drive in mid-June and now envisions 120 TB-class SSDs based on the same platform.

Samsung's BM1743 61.44 TB features a proprietary controller and relies on Samsung's 7^th Generation V-NAND (3D NAND) QLC memory. Moreover, Samsung believes that its 7^th Gen V-NAND 'has the potential to accommodate up to 122.88 TB,' 

Samsung plans to offer the BM1743 in two form factors: U.2 for PCIe 4.0 x4 to address traditional servers and E3.S for PCIe 5.0 x4 interfaces to address machines designed to offer maximum storage density. BM1743 can address various applications, including AI training and inference, content delivery networks, and read-intensive workloads. To that end, its write endurance is 0.26 drive writes per day (DWPD) over five years.

Regarding performance, Samsung's BM1743 is hardly a champion compared to high-end drives for gaming machines and workstations. The drive can sustainably achieve sequential read speeds of 7,200 MB/s and write speeds of 2,000 MB/s. It can handle up to 1.6 million 4K random reads and 110,000 4K random writes for random operations.

Power consumption details for the BM1743 have not been disclosed, though it is expected to be high. Meanwhile, the drive's key selling point is its massive storage density, which likely outweighs concerns over its absolute power efficiency for intended applications, as a 60 TB SSD still consumes less than multiple storage devices offering similar capacity and performance.

As noted above, Samsung's BM1743 61.44 TB faces limited competition in the market, so its price will be quite high. For example, Solidigm's D5-P5336 61.44 TB SSD costs $6,905. Other companies, such as Kioxia, Micron, and SK Hynix, have not yet introduced their 60TB-class SSDs, which gives Samsung and Solidigm an edge for now.

UPDATE 7/25: We removed mention of Western Digital's 60 TB-class SSDs, as the company does not currently list any such drives on their website

SSDs

Best CPUs for Gaming: July 2024

As the second quarter of 2024 is soon set to unfold, there are many things to be excited about, especially as Computex 2024 has been and gone. We now know that AMD's upcoming Ryzen 9000 series desktop processors using the new Zen 5 cores will be hitting shelves at the end of the month (31st July), and on top of this, AMD also recently slashed pricing on their Zen 4 (Ryzen 8000) processors. Intel still needs to follow suit with their 14th or 13th Gen Core series processors, but right now from a cost standpoint, AMD is in a much better position.

Since the publication of our last guide, the only notable CPU to be launched was Intel's special binned Core i9-14900KS, which not only pushes clock speeds up to 6.2 GHz but is the last processor to feature Intel's iconic Core I series nomenclature. The other big news in the CPU world was from Intel, with a statement issued about pushing users to use the Intel Default Specification on Intel's 14th and 13th Gen processors, which ultimately limits the performance compared to published data. We're still in the process of 

While the CPU market has been relatively quiet so far this year, and things are set to pick up once AMD's Zen 5 and Intel's Arrow Lake desktop chips are all launched onto the market, it means today we are working for the same hymn sheet as our previous guide. With AMD's price drops on Ryzen 7000 series processors, much of the guide reflects this as AMD and Intel's performance is neck and neck in many use cases, but cost certainly plays a big factor in selecting a new CPU. As we move into the rest of 2024, the CPU market looks set to see the rise of the 'AI PC,' which is looking set to be something that many companies will focus on by the end of 2024, both on mobile and desktop platforms.

Guides

Micron Ships Denser & Faster 276 Layer TLC NAND, Arriving First In Micron 2650 Client SSDs

Micron on Tuesday announced that the company has begun shipping its 9th Generation (G9) 276 layer TLC NAND. The next generation of NAND from the prolific memory maker, Micron's latest NAND is designed to further push the envelope on TLC NAND performance, offering significant density and performance improvements over its existing NAND technology.

Micron's G9 TLC NAND memory features 276 active layers, which is up from 232-layers in case of Micron's previous generation TLC NAND. At this point the company is being light on technical details in their official material. However in a brief interview with Blocks & Files, the company confirmed that their 276L NAND still uses a six plane architecture, which was first introduced with the 232L generation. At this point we're assuming Micron is also string-stacking two decks of NAND together, as they have been for the past couple of generations, which means we're looking at 138 layer decks.

Micron TLC NAND Flash Memory
	276L	232L (B58R)	176L (B47R)
Layers	276	232	176
Decks	2 (x138)?	2 (x116)	2 (x88)
Die Capacity	1 Tbit	1 Tbit	512 Gbit
Die Size (mm2)	~48.9mm2	~70.1mm2	~49.8mm2
Density (Gbit/mm2)	~21	14.6	10.3
I/O Speed	3.6 GT/s (ONFi 5.1)	2.4 GT/s (ONFi 5.0)	1.6 GT/s (ONFI 4.2)
Planes	6	6	4
CuA / PuC	Yes	Yes	Yes

On the density front, Micron told Blocks & Files that they have improved their NAND density by 44% over their 232L generation. Which, given what we know about that generation, would put the density at around 21 Gbit/mm². Or for a 1Tbit die of TLC NAND, that works out to a die size of roughly 48.9mm², comparable to the die size of a 512Gbit TLC die from Micron's older 176L NAND.

Besides improving density, the other big push with Micron's newest generation of NAND was further improving its throughput. While the company's 232L NAND was built against the ONFi 5.0 specification, which topped out at transfer rates of 2400 MT/sec, their new 276L NAND can hit 3600 MT/sec, which is consistent with the ONFi 5.1 spec.

Meanwhile, the eagle-eyed will likely also pick up on Micron's ninth-generation/G9 branding, which is new to the company. Micron's has not previously used this kind of generational branding for their NAND, which up until now has simply been identified by its layer count (and before the 3D era, its feature size). Internally, this is believed to be Micron's 7th generation 3D NAND architecture. However, taking a page from the logic fab industry, Micron seems to be branding it as ninth-generation in order to keep generational parity with its competitors, who are preparing their own 8th/9th generation NAND (and thus cliam that they are the first NAND maker to ship 9th gen NAND).

And while this NAND will eventually end up in ... SSDs

Kioxia's High-Performance 3D QLC NAND Enables High-End High-Capacity SSDs

This week, Kioxia introduced its new 3D QLC NAND devices aimed at high-performance, high-capacity drives that could redefine what we typically expect from QLC-based SSDs. The components are 1 Tb and 2 Tb 3D QLC NAND ICs with a 3600 MT/s interface speed that could enable M.2-2230 SSDs with a 4 TB capacity and decent performance.

Kioxia's 1 Tb (128 MB) and 2 Tb (256 TB) 3D QLC NAND devices are made on the company's BICS 8 process technology and feature 238 active layers as well as CMOS directly Bonded to Array (CBA) design, which implies that CMOS (including interface and buffers circuitry) is built on a specialized node and bonded to the memory array. Such a manufacturing process enabled Kioxia (and its manufacturing partner Western Digital) to achieve a particularly high interface speed of 3600 MT/s.

In addition to being one of the industry's first 2 Tb QLC NAND devices, the component features a 70% higher write power efficiency compared to Kioxia's BICS 5 3D QLC NAND devices, which is a bit vague statement as the new ICs have higher capacity and performance in general. This feature will be valuable for data centre applications, though I do not expect someone to use 3D QLC memory for write-intensive applications in general. Yet, these devices will be just what the doctor ordered for AI: read-intensive, content distribution, and backup storage.

It is interesting to note that Kioxia's 1 Tb 3D QLC NAND, optimized for performance, has a 30% faster sequential write performance and a 15% lower read latency than the 2 Tb 3D QLC component. These qualities (alongside a 3600 MT/s interface) promise to make Kioxia's 1 Tb 3D QLC competitive even for higher-end PCIe Gen5 x4 SSDs, which currently exclusively use 3D TLC memory.

The remarkable storage density of Kioxia's 2Tb 3D QLC NAND devices will allow customers to create high-capacity SSDs in compact form factors. For instance, a 16-Hi stacked package (measuring 11.5 mm × 13.5 mm × 1.5 mm) can be used to build a 4TB M.2-2230 drive or a 16TB M.2-2280 drive. Even a single 16-Hi package could be enough to build a particularly fast client SSD.

Kioxia is now sampling its 2 Tb 3D QLC NAND BiCS 8 memory with customers, such as Pure Storage.

"We have a long-standing relationship with Kioxia and are delighted to incorporate their eighth-generation BiCS Flash 2Tb QLC flash memory products to enhance the performance and efficiency of our all-flash storage solutions," said Charles Giancarlo, CEO of Pure Storage. "Pure's unified all-flash data storage platform is able to meet the demanding needs of artificial intelligence as well as the aggressive costs of backup storage. Backed by Kioxia technology, Pure Storage will continue to offer unmatched performance, power efficiency, and reliability, delivering exceptional value to our customers."

"We are pleased to be shipping samples of our new 2Tb QLC with the new eighth-generation BiCS flash technology," said Hideshi Miyajima, CTO of Kioxia. "With its industry-leading high bit density, high speed data transfer, and superior power efficiency, the 2Tb QLC product will offer new value for rapidly emerging AI applications and large storage applications demanding power and space savings."

There is no word on when the 1 Tb 3D QLC BiCS 8 memory will be sampled or released to the market.

SSDs

Best CPUs for Gaming: July 2024

As the second quarter of 2024 is soon set to unfold, there are many things to be excited about, especially as Computex 2024 has been and gone. We now know that AMD's upcoming Ryzen 9000 series desktop processors using the new Zen 5 cores will be hitting shelves at the end of the month (31st July), and on top of this, AMD also recently slashed pricing on their Zen 4 (Ryzen 8000) processors. Intel still needs to follow suit with their 14th or 13th Gen Core series processors, but right now from a cost standpoint, AMD is in a much better position.

Since the publication of our last guide, the only notable CPU to be launched was Intel's special binned Core i9-14900KS, which not only pushes clock speeds up to 6.2 GHz but is the last processor to feature Intel's iconic Core I series nomenclature. The other big news in the CPU world was from Intel, with a statement issued about pushing users to use the Intel Default Specification on Intel's 14th and 13th Gen processors, which ultimately limits the performance compared to published data. We're still in the process of 

While the CPU market has been relatively quiet so far this year, and things are set to pick up once AMD's Zen 5 and Intel's Arrow Lake desktop chips are all launched onto the market, it means today we are working for the same hymn sheet as our previous guide. With AMD's price drops on Ryzen 7000 series processors, much of the guide reflects this as AMD and Intel's performance is neck and neck in many use cases, but cost certainly plays a big factor in selecting a new CPU. As we move into the rest of 2024, the CPU market looks set to see the rise of the 'AI PC,' which is looking set to be something that many companies will focus on by the end of 2024, both on mobile and desktop platforms.

Guides

Intel to Launch "Lunar Lake" Core Ultra Chips on September 3rd

Intel’s next-generation Core Ultra laptop chips finally have a launch date: September 3^rd.

Codenamed Lunar Lake, Intel has been touting the chips for nearly a year now. Most recently, Intel offered the press a deep dive briefing on the chips and their underlying architectures at Computex back in June, along with a public preview during the company’s Computex keynote. At the time Intel was preparing for Q3’2024 launch, and that window has finally been narrowed down to a single date – September 3^rd – when Intel will be hosting their Lunar Lake launch event ahead of IFA.

Intel’s second stab at a high volume chiplet-based processor for laptop users, Lunar Lake is aimed particularly at ultrabooks and other low-power mobile devices, with Intel looking to wrestle back the title of the most efficient PC laptop SoC. Lunar Lake is significant in this respect as Intel has never previously developed a whole chip architecture specifically for low power mobile devices before – it’s always been a scaled-down version of a wider-range architecture, such as the current Meteor Lake (Core Ultra 100 series). Consequently, Intel has been touting that they’ve made some serious efficiency advancements with their highly targeted chip, which they believe will vault them over the competition.

All told, Lunar Lake is slated to bring a significant series of updates to Intel’s chip architectures and chip design strategies. Of particular interest is the switch to on-package LPDDR5X memory, which is a first for a high-volume Core chip. As well, Lunar Lake incorporates updated versions of virtually every one of Intel’s architecture, from the CPU P and E cores – Lion Cove and Skymont respectively – to the Xe2 GPU and 4^th generation NPU (aptly named NPU 4). And, in a scandalous twist, both of the chiplets/tiles on the CPU are being made by TSMC. Intel isn’t providing any of the active silicon for the chip – though they are providing the Foveros packaging needed to put it together.

Intel CPU Architecture Generations
	Alder/Raptor Lake	Meteor Lake	Lunar Lake	Arrow Lake	Panther Lake
P-Core Architecture	Golden Cove/ Raptor Cove	Redwood Cove	Lion Cove	Lion Cove	Cougar Cove?
E-Core Architecture	Gracemont	Crestmont	Skymont	Crestmont?	Darkmont?
GPU Architecture	Xe-LP	Xe-LPG	Xe2	Xe2?	?
NPU Architecture	N/A	NPU 3720	NPU 4	?	?
Active Tiles	1 (Monolithic)	4	2	4?	?
Manufacturing Processes	Intel 7	Intel 4 + TSMC N6 + TSMC N5	TSMC N3B + TSMC N6	Intel 20A + More	Intel 18A
Segment	Mobile + Desktop	Mobile	LP Mobile	HP Mobile + Desktop	Mobile?
Release Date (OEM)	Q4'2021	Q4'2023	Q3'2024	Q4'2024	2025

Suffice it to say, no matter what happens, Lunar Lake and the Core Ultra 200 series should prove to be an interesting launch.

It’s worth noting, however, that while Intel’s announcement of their livestreamed event is being labeled a “launch event&rd... CPUs

The Lian Li Hydroshift LCD 360S AIO Cooler Review: Sleek, Stylish, and Lively

Among the packed field of PC hardware manufacturers, Lian Li is a company that arguably shouldn't even need an introduction. The quirky company has developed a devout following thanks to its focus on premium-quality aluminum computer cases that, more often that not, and come with in some rather unique designs. Over the years, the company has developed a solid reputation for its meticulous craftsmanship, durability, and elegant aesthetics. And consequently, when the company made the decision to expand beyond aluminum cases and in to other PC peripherals, that development attracted quite a bit of attention to see what kind of a touch Lian Li could bring to the rest of the PC ecosystem.

Lian Li's focus on premium products means that the company doesn't really make much in the way of products that are merely basic, and that kind of mentality has extended beyond cases and into the rest of their peripherals. Case in point is the subject of today's review: Lian Li's new all-in-one CPU cooler, the HydroShift LCD 360S AIO. Not content just to make a powerful 360 mm cooler, Lian Li has gone a step above by integrating recesses and other features to help hide the tubing around the cooler, and then for the coup de grace, added a high-quality 2.88-inch IPS display to the pump block.

Overall, this new product marks a significant milestone for Lian Li, as it combines advanced cooling technology with the company's signature aesthetic appeal, making for a cooler that's aimed at both enthusiasts and professional users seeking high-end thermal performance and visual customization. Overall, the HydroShift LCD 360S is undeniably designed first and foremost with aesthetics in mind, but the shiny pump block is backed up with one of the most powerful 360 mm cooler designs on the market today. So Lian Li is throwing everything they have at the new HydroShift coolers.

Overall, the 360S is part of a trio of HydroShift 360 mm coolers the company is launching this summer. All three share a similar design, although with some pump changes and the addition of RGB fan lighting, depending on the specific model, with the 360S effectively serving as the base model.

Cases/Cooling/PSUs

Intel to Launch "Lunar Lake" Core Ultra Chips on September 3rd

Intel’s next-generation Core Ultra laptop chips finally have a launch date: September 3^rd.

Codenamed Lunar Lake, Intel has been touting the chips for nearly a year now. Most recently, Intel offered the press a deep dive briefing on the chips and their underlying architectures at Computex back in June, along with a public preview during the company’s Computex keynote. At the time Intel was preparing for Q3’2024 launch, and that window has finally been narrowed down to a single date – September 3^rd – when Intel will be hosting their Lunar Lake launch event ahead of IFA.

Intel’s second stab at a high volume chiplet-based processor for laptop users, Lunar Lake is aimed particularly at ultrabooks and other low-power mobile devices, with Intel looking to wrestle back the title of the most efficient PC laptop SoC. Lunar Lake is significant in this respect as Intel has never previously developed a whole chip architecture specifically for low power mobile devices before – it’s always been a scaled-down version of a wider-range architecture, such as the current Meteor Lake (Core Ultra 100 series). Consequently, Intel has been touting that they’ve make some serious efficiency advancements with their highly targeted chip, which they believe will vault them over the competition.

All told, Lunar Lake is slated to bring a significant series of updates to Intel’s chip architectures and chip design strategies. Of particular interest is the switch to on-package LPDDR5X memory, which is a first for a high-volume Core chip. As well, Lunar Lake incorporates updated versions of virtually every one of Intel’s architecture, from the CPU P and E cores – Lion Cove and Skymont respectively – to the Xe2 GPU and 4^th generation NPU (aptly named NPU 4). And, in a scandalous twist, both of the chiplets/tiles on the CPU are being made by TSMC. Intel isn’t providing any of the active silicon for the chip – though they are providing the Foveros packaging needed to put it together.

Intel CPU Architecture Generations
	Alder/Raptor Lake	Meteor Lake	Lunar Lake	Arrow Lake	Panther Lake
P-Core Architecture	Golden Cove/ Raptor Cove	Redwood Cove	Lion Cove	Lion Cove	Cougar Cove?
E-Core Architecture	Gracemont	Crestmont	Skymont	Crestmont?	Darkmont?
GPU Architecture	Xe-LP	Xe-LPG	Xe2	Xe2?	?
NPU Architecture	N/A	NPU 3720	NPU 4	?	?
Active Tiles	1 (Monolithic)	4	2	4?	?
Manufacturing Processes	Intel 7	Intel 4 + TSMC N6 + TSMC N5	TSMC N3B + TSMC N6	Intel 20A + More	Intel 18A
Segment	Mobile + Desktop	Mobile	LP Mobile	HP Mobile + Desktop	Mobile?
Release Date (OEM)	Q4'2021	Q4'2023	Q3'2024	Q4'2024	2025

Suffice it to say, no matter what happens, Lunar Lake and the Core Ultra 200 series should prove to be an interesting launch.

It’s worth noting, however, that while Intel’s announcement of their livestreamed event is being labeled a “launch event&rd... CPUs

SK hynix to Enter 60 TB SSD Club Next Quarter

SK hynix this week reported its financial results for the second quarter, as well as offering a glimpse at its plans for the coming quarters. Notably among the company's plans for the year is the release of a SK hynix-branded 60 TB SSD, which will mark the firm's entry into the ultra-premium enterprise SSD league.

"SK hynix plans to expand sales of high-capacity eSSD and lead the market in the second half with 60TB products, expecting eSSD sales to be more than quadrupled compared to last year," a statement by SK hynix reads.

Currently there are only two standard form-factor 61.44 TB SSDs on the market: the Solidigm D5-P5336 (U.2/15mm and E1.L), and the Samsung BM1743 (U.2/15mm and E3.S). Both are built from a proprietary controller (Solidigm's controller still carries an Intel logotype) with a PCIe 4.0 x4 interface and use QLC NAND for storage.

SK hynix's brief mention of the drive means that tere aren't any formal specifications or capabilities to discuss just yet. But it is reasonable to assume that the company will use its own QLC memory for their ultra-high-capacity drives. What's more intriguing are which controller the company plans to use and how it is going to position its 60 TB-class SSD.

Internally, SK hynix has access to multiple controller teams, both of which have the expertise to develop an enterprise-grade controller suitable for a 60 TB drive. SK hynix technically owns Solidigm, the former Intel SSD and NAND unit, giving SK hynix the option of using Solidigm's controller, or even reselling a rebadged D5-P5336 outright. Alternatively, SK hynix has their own (original) internal SSD team, who is responsible for building their well-received Aries SSD controller, among other works.

Ultra-high-capacity SSDs for performance demanding read-intensive storage applications, such as AI inference on the edge or content delivery networks, is a promising premium market. So SK hynix is finding itself highly incentivized to enter it with a compelling offering. 

SSDs

Micron Expands Datacenter DRAM Portfolio with MR-DIMMs

The compute market has always been hungry for memory bandwidth, particularly for high-performance applications in servers and datacenters. In recent years, the explosion in core counts per socket has further accentuated this need. Despite progress in DDR speeds, the available bandwidth per core has unfortunately not seen a corresponding scaling.

The stakeholders in the industry have been attempting to address this by building additional technology on top of existing widely-adopted memory standards. With DDR5, there are currently two technologies attempting to increase the peak bandwidth beyond the official speeds. In late 2022, SK hynix introduced MCR-DIMMs meant for operating with specific Intel server platforms. On the other hand, JEDEC - the standards-setting body - also developed specifications for MR-DIMMs with a similar approach. Both of them build upon existing DDR5 technologies by attempting to combine multiple ranks to improve peak bandwidth and latency.

How MR-DIMMs Work

The MR-DIMM standard is conceptually simple - there are multiple ranks of memory modules operating at standard DDR5 speeds with a data buffer in front. The buffer operates at 2x the speed on the host interface side, allowing for essentially double the transfer rates. The challenges obviously lie in being able to operate the logic in the host memory controller at the higher speed and keeping the power consumption / thermals in check.

The first version of the JEDEC MR-DIMM standard specifies speeds of 8800 MT/s, with the next generation at 12800 MT/s. JEDEC also has a clear roadmap for this technology, keeping it in sync with the the improvements in the DDR5 standard.

Micron MR-DIMMs - Bandwidth and Capacity Plays

Micron and Intel have been working closely in the last few quarters to bring their former's first-generation MR-DIMM lineup to the market. Intel's Xeon 6 Family with P-Cores (Granite Rapids) is the first platform to bring MR-DIMM support at 8800 MT/s on the host side. Micron's standard-sized MR-DIMMs (suitable for 1U servers) and TFF (tall form-factor) MR-DIMMs (for 2U+ servers) have been qualified for use with the same.

The benefits offered by MR-DIMMs are evident from the JEDEC specifications, allowing for increased data rates and system bandwidth, with improvements in latency. On the capacity side, allowing for additional ranks on the modules has enabled Micron to offer a 256 GB capacity point. It must be noted that some vendors are also using TSV (through-silicon vias) technology to to increase the per-package capacity at standard DDR5 speeds, but this adds additional cost and complexity that are largely absent in the MR-DIMM manufacturing process.

The tall form-factor (TFF) MR-DIMMs have a larger surface area compared to the standard-sized ones. For the same airflow configuration, this allows the DIMM to have a better thermal profile. This provides benefits for energy efficiency as well by reducing the possibility of thermal throttling.

Micron is launching a comprehensive lineup of MR-DIMMs in both standard and tall form-factors today, with multiple DRAM densities and speed options as noted above.

MRDIMM Benefits - Intel Granite Rapids Gets a Performance Boost

Micron and Intel hosted a media / analyst briefing recently to demonstrate the benefits of MR-DIMMs for Xeon 6 with P-Cores (Granite Rapids). Using a 2P configuration with 96-core Xeon 6 processors, benchmarks for different ... Memory

Kioxia's High-Performance 3D QLC NAND Enables High-End High-Capacity SSDs

This week, Kioxia introduced its new 3D QLC NAND devices aimed at high-performance, high-capacity drives that could redefine what we typically expect from QLC-based SSDs. The components are 1 Tb and 2 Tb 3D QLC NAND ICs with a 3600 MT/s interface speed that could enable M.2-2230 SSDs with a 4 TB capacity and decent performance.

Kioxia's 1 Tb (128 MB) and 2 Tb (256 TB) 3D QLC NAND devices are made on the company's BICS 8 process technology and feature 238 active layers as well as CMOS directly Bonded to Array (CBA) design, which implies that CMOS (including interface and buffers circuitry) is built on a specialized node and bonded to the memory array. Such a manufacturing process enabled Kioxia (and its manufacturing partner Western Digital) to achieve a particularly high interface speed of 3600 MT/s.

In addition to being one of the industry's first 2 Tb QLC NAND devices, the component features a 70% higher write power efficiency compared to Kioxia's BICS 5 3D QLC NAND devices, which is a bit vague statement as the new ICs have higher capacity and performance in general. This feature will be valuable for data centre applications, though I do not expect someone to use 3D QLC memory for write-intensive applications in general. Yet, these devices will be just what the doctor ordered for AI: read-intensive, content distribution, and backup storage.

It is interesting to note that Kioxia's 1 Tb 3D QLC NAND, optimized for performance, has a 30% faster sequential write performance and a 15% lower read latency than the 2 Tb 3D QLC component. These qualities (alongside a 3600 MT/s interface) promise to make Kioxia's 1 Tb 3D QLC competitive even for higher-end PCIe Gen5 x4 SSDs, which currently exclusively use 3D TLC memory.

The remarkable storage density of Kioxia's 2Tb 3D QLC NAND devices will allow customers to create high-capacity SSDs in compact form factors. For instance, a 16-Hi stacked package (measuring 11.5 mm × 13.5 mm × 1.5 mm) can be used to build a 4TB M.2-2230 drive or a 16TB M.2-2280 drive. Even a single 16-Hi package could be enough to build a particularly fast client SSD.

Kioxia is now sampling its 2 Tb 3D QLC NAND BiCS 8 memory with customers, such as Pure Storage.

"We have a long-standing relationship with Kioxia and are delighted to incorporate their eighth-generation BiCS Flash 2Tb QLC flash memory products to enhance the performance and efficiency of our all-flash storage solutions," said Charles Giancarlo, CEO of Pure Storage. "Pure's unified all-flash data storage platform is able to meet the demanding needs of artificial intelligence as well as the aggressive costs of backup storage. Backed by Kioxia technology, Pure Storage will continue to offer unmatched performance, power efficiency, and reliability, delivering exceptional value to our customers."

"We are pleased to be shipping samples of our new 2Tb QLC with the new eighth-generation BiCS flash technology," said Hideshi Miyajima, CTO of Kioxia. "With its industry-leading high bit density, high speed data transfer, and superior power efficiency, the 2Tb QLC product will offer new value for rapidly emerging AI applications and large storage applications demanding power and space savings."

There is no word on when the 1 Tb 3D QLC BiCS 8 memory will be sampled or released to the market.

SSDs

Samsung Validates LPDDR5X Running at 10.7 GT/sec with MediaTek's Dimensity 9400 SoC

Samsung has successfully validated its new LPDDR5X-10700 memory with MediaTek's upcoming Dimensity platform. At present, 10.7 GT/s is the highest performing speed grade of LPDDR5X DRAM slated to be released this year, so the upcoming Dimensity 9400 system-on-chip will get the highest memory bandwidth available for a mobile application processor.

The verification process involved Samsung's 16 GB LPDDR5X package and MediaTek's soon-to-be-announced Dimensity 9400 SoC for high-end 5G smartphones. Usage of LPDDR5X-10700 provides a memory bandwidth of 85.6 GB/second over a 64-bit interface, which will be available for bandwidth-hungry applications like graphics and generative AI.

"Working together with Samsung Electronics has made it possible for MediaTek's next-generation Dimensity chipset to become the world's first to be validated at LPDDR5X operating speeds up to 10.7Gbps, enabling upcoming devices to deliver AI functionality and mobile performance at a level we have never seen before," said JC Hsu, Corporate Senior Vice President at MediaTek. "This updated architecture will make it easier for developers and users to leverage more AI capabilities and take advantage of more features with less impact on battery life."

Samsung's LPDDR5X 10.7 GT/s memory in made on the company's 12nm-class DRAM process technology and is said to provide a more than 25% improvement in power efficiency over previous-generation LPDDR5X, in addition to extra performance. This will positively affect improved user experience, including enhanced on-device AI capabilities, such as faster voice-to-text conversion, and better quality graphics.

Overall, the two companies completed this process in just three months. Though it remains to be seen when smartphones based on the Dimensity 9400 application processor and LPDDR5X memory are set to be available on the market, as MediaTek has not yet even formally announced the SoC itself.

"Through our strategic cooperation with MediaTek, Samsung has verified the industry's fastest LPDDR5X DRAM that is poised to lead the AI smartphone market," said YongCheol Bae, Executive Vice President of Memory Product Planning at Samsung Electronics. "Samsung will continue to innovate through active collaboration with customers and provide optimum solutions for the on-device AI era."

Memory

Tenstorrent Launches Wormhole AI Processors: 466 FP8 TFLOPS at 300W

Tenstorrent has unveiled its next-generation Wormhole processor for AI workloads that promises to offer decent performance at a low price. The company currently offers two add-on PCIe cards carrying one or two Wormhole processors as well as TT-LoudBox, and TT-QuietBox workstations aimed at software developers. The whole of today's release is aimed at developers rather than those who will deploy the Wormhole boards for their commercial workloads.

“It is always rewarding to get more of our products into developer hands. Releasing development systems with our Wormhole™ card helps developers scale up and work on multi-chip AI software.” said Jim Keller, CEO of Tenstorrent. “In addition to this launch, we are excited that the tape-out and power-on for our second generation, Blackhole, is going very well.”

Each Wormhole processor packs 72 Tensix cores (featuring five RISC-V cores supporting various data formats) with 108 MB of SRAM to deliver 262 FP8 TFLOPS at 1 GHz at 160W thermal design power. A single-chip Wormhole n150 card carries 12 GB of GDDR6 memory featuring a 288 GB/s bandwidth.

Wormhole processors offer flexible scalability to meet the varying needs of workloads. In a standard workstation setup with four Wormhole n300 cards, the processors can merge to function as a single unit, appearing as a unified, extensive network of Tensix cores to the software. This configuration allows the accelerators to either work on the same workload, be divided among four developers or run up to eight distinct AI models simultaneously. A crucial feature of this scalability is that it operates natively without the need for virtualization. In data center environments, Wormhole processors will scale both inside one machine using PCIe or outside of a single machine using Ethernet. 

From performance standpoint, Tenstorrent's single-chip Wormhole n150 card (72 Tensix cores at 1 GHz, 108 MB SRAM, 12 GB GDDR6 at 288 GB/s) is capable of 262 FP8 TFLOPS at 160W, whereas the dual-chip Wormhole n300 board (128 Tensix cores at 1 GHz, 192 MB SRAM, aggregated 24 GB GDDR6 at 576 GB/s) can offer up to 466 FP8 TFLOPS at 300W (according to Tom's Hardware).

To put that 466 FP8 TFLOPS at 300W number into context, let's compare it to what AI market leader Nvidia has to offer at this thermal design power. Nvidia's A100 does not support FP8, but it does support INT8 and its peak performance is 624 TOPS (1,248 TOPS with sparsity). By contrast, Nvidia's H100 supports FP8 and its peak performance is massive 1,670 TFLOPS (3,341 TFLOPS with sparsity) at 300W, which is a big difference from Tenstorrent's Wormhole n300. 

There is a big catch though. Tenstorrent's Wormhole n150 is offered for $999, whereas n300 is available for $1,399. By contrast, one Nvidia H100 card can retail for $30,000, depending on quantities. Of course, we do not know whether four or eight Wormhole processors can indeed deliver the performance of a single H300, though they will do so at 600W or 1200W TDP, respectively.

In addition to cards, Tenstorrent offers developers pre-built workstations with four n300 cards inside the less expensive Xeon-based TT-LoudBox with active cooling and a premium EPYC-powered TT-QuietBox with liquid cooling.

Sources: Tenstorrent, Tom's Hardware

AI

Frore Unveils Waterproof AirJet Mini Sport for Smartphones

Over the past couple of years, Frore Systems has demonstrated several ways that its AirJet solid-state active cooling systems can be used to improve cooling in fanless devices like laptops, tablets, SSDs, and edge computing devices. But there are a subset of those applications that need their cooling options to also be waterproof, and Frore is looking to address those as well. To that end, this week Frore introduced its AirJet Mini Sport, a waterproof, IP68-rated solid-state cooling device that is aimed at use in smartphones and action cameras.

Introduced at MWC Shanghai to attract attention of China-based handset vendors, edge and industrial computing devices, and action cameras, the AirJet Mini Sport is an enhanced version of Frore's AirJet Mini Slim. This version has been fully waterproofed, offering IP68-level protection that allows it to work while being submerged in over 1.5 meters of water for up to 30 minutes. Internally, the AirJet Mini Sport can effectively dissipate 5.25 Watts of heat by generating 1750 Pascals of back pressure, while consuming 1 Watt of energy itself.

Elsewhere, Frore claims that the AirJet Mini Sport can be used to provide 2.5 Watts of cooling capacity to smartphones. Which, although not enough to cover the complete power consumption/heat dissipation of a high-end SoC, would have a significant impact on both burst and steady-state performance by allowing those chips to run at peak clocks and power for longer periods of time.

To ensure consistent performance of Frore's AirJet Mini Sport in diverse environments, the cooling device includes features such as dust resistance and self-cleaning. In addition, just like AirJet Mini Slim, the Sport-badged version its own thermal sensor to control its own operation and maintain optimal performance. As a result, Frore claims that smartphones and action cameras with the AirJet Mini Sport can achieve up to 80% better performance.

"We are excited to announce the waterproof AirJet Mini Sport," said Dr. Seshu Madhavapeddy, founder and CEO of Frore Systems. "Consumers demand increased performance in compact devices they can use anywhere, on land or in water. AirJet unleashes device performance, now enabling users to do more with their IP68 dustproof and waterproof devices."

Air Cooling

Samsung Joins The 60 TB SSD Club, Looking Forward To 120 TB Drives

Multiple companies offer high-capacity SSDs, but until recently, only one company offered high-performance 60 TB-class drives with a PCIe interface: Solidigm. As our colleagues from Blocks & Files discovered, Samsung quietly rolled out its BM1743 61.44 TB solid-state drive in mid-June and now envisions 120 TB-class SSDs based on the same platform.

Samsung's BM1743 61.44 TB features a proprietary controller and relies on Samsung's 7^th Generation V-NAND (3D NAND) QLC memory. Moreover, Samsung believes that its 7^th Gen V-NAND 'has the potential to accommodate up to 122.88 TB,' 

Samsung plans to offer the BM1743 in two form factors: U.2 for PCIe 4.0 x4 to address traditional servers and E3.S for PCIe 5.0 x4 interfaces to address machines designed to offer maximum storage density. BM1743 can address various applications, including AI training and inference, content delivery networks, and read-intensive workloads. To that end, its write endurance is 0.26 drive writes per day (DWPD) over five years.

Regarding performance, Samsung's BM1743 is hardly a champion compared to high-end drives for gaming machines and workstations. The drive can sustainably achieve sequential read speeds of 7,200 MB/s and write speeds of 2,000 MB/s. It can handle up to 1.6 million 4K random reads and 110,000 4K random writes for random operations.

Power consumption details for the BM1743 have not been disclosed, though it is expected to be high. Meanwhile, the drive's key selling point is its massive storage density, which likely outweighs concerns over its absolute power efficiency for intended applications, as a 60 TB SSD still consumes less than multiple storage devices offering similar capacity and performance.

As noted above, Samsung's BM1743 61.44 TB faces limited competition in the market, so its price will be quite high. For example, Solidigm's D5-P5336 61.44 TB SSD costs $6,905. Other companies, such as Kioxia, Micron, and SK Hynix, have not yet introduced their 60TB-class SSDs, which gives Samsung and Solidigm an edge for now.

UPDATE 7/25: We removed mention of Western Digital's 60 TB-class SSDs, as the company does not currently list any such drives on their website

SSDs

Frore Unveils Waterproof AirJet Mini Sport for Smartphones

Over the past couple of years, Frore Systems has demonstrated several ways that its AirJet solid-state active cooling systems can be used to improve cooling in fanless devices like laptops, tablets, SSDs, and edge computing devices. But there are a subset of those applications that need their cooling options to also be waterproof, and Frore is looking to address those as well. To that end, this week Frore introduced its AirJet Mini Sport, a waterproof, IP68-rated solid-state cooling device that is aimed at use in smartphones and action cameras.

Introduced at MWC Shanghai to attract attention of China-based handset vendors, edge and industrial computing devices, and action cameras, the AirJet Mini Sport is an enhanced version of Frore's AirJet Mini Slim. This version has been fully waterproofed, offering IP68-level protection that allows it to work while being submerged in over 1.5 meters of water for up to 30 minutes. Internally, the AirJet Mini Sport can effectively dissipate 5.25 Watts of heat by generating 1750 Pascals of back pressure, while consuming 1 Watt of energy itself.

Elsewhere, Frore claims that the AirJet Mini Sport can be used to provide 2.5 Watts of cooling capacity to smartphones. Which, although not enough to cover the complete power consumption/heat dissipation of a high-end SoC, would have a significant impact on both burst and steady-state performance by allowing those chips to run at peak clocks and power for longer periods of time.

To ensure consistent performance of Frore's AirJet Mini Sport in diverse environments, the cooling device includes features such as dust resistance and self-cleaning. In addition, just like AirJet Mini Slim, the Sport-badged version its own thermal sensor to control its own operation and maintain optimal performance. As a result, Frore claims that smartphones and action cameras with the AirJet Mini Sport can achieve up to 80% better performance.

"We are excited to announce the waterproof AirJet Mini Sport," said Dr. Seshu Madhavapeddy, founder and CEO of Frore Systems. "Consumers demand increased performance in compact devices they can use anywhere, on land or in water. AirJet unleashes device performance, now enabling users to do more with their IP68 dustproof and waterproof devices."

Air Cooling

Micron: U.S. Fabs Will Start Operating in 2026 - 2029

When Micron announced plans to build two new fabs in the U.S. in 2022, the company vaguely said both would come online by the decade's end. Then, in 2023, it began to optimize its spending, which pushed production at these fabrication facilities. This week, the company outlined more precise timeframes for when its fabs in Idaho and New York will start operations: this will happen from calendar 2026 to calendar 2029.

"These fab construction investments are necessary to support supply growth for the latter half of this decade," a statement by Micron in its Q3 FY2024 financial results report reads. "This Idaho fab will not contribute to meaningful bit supply until fiscal 2027 and the New York construction capex is not expected to contribute to bit supply growth until fiscal 2028 or later. The timing of future [wafer fab equipment] spend in these fabs will be managed to align supply growth with expected demand growth."

Micron's fiscal year 2027 starts in September 2026, so the new fab near Boise, Idaho, is set to start operations between September 2026 and September 2027. The company's fiscal 2028 starts in September 2027, so the fab will likely begin operations in calendar 2028 or later, probably depending on the demand for DRAM memory in the coming years. That said, Micron's U.S. memory fabs will begin operations between late 2026 and 2029, which aligns with the company's original plans. 

Construction of the fab in Idaho is well underway. In contrast, construction of the New York facility has yet to begin as the company is working on regulatory and permitting processes in the state. 

Micron's capital expenditure (CaPex) plan for FY2024 is approximately $8.0 billion, with a decrease in year-over-year spending on wafer fabrication equipment (WFE). In Q4 FY2024, the company will spend around $3 billion on fab construction, new wafer fab tools, and various expansions/upgrades.

Looking ahead to FY2025, the company plans a substantial increase in capex, targeting a mid-30s percentage of revenue to support various technological and facility advancements. In particular, it expects its quarterly CapEx to average above the $3 billion level seen in the fourth quarter of FY2024, which means that it plans to spend about $12 billion in its fiscal 2025, which begins in late September.

Half or more of the total CapEx increase in FY2025 (i.e., over $2 billion) will be allocated to constructing new fabs in Idaho and New York. Meanwhile, the FY2025 CapEx will significantly rise to fund high-bandwidth memory (HBM) assembly and testing and the construction of fabrication and back-end facilities. This increase also includes investments in technology transitions to meet growing demand. 

"Fab construction in Idaho is underway, and we are working diligently to complete the regulatory and permitting processes in New York," said Sanjay Mehrotra, chief executive officer of Micron, at the company's conference call with investors and financial analysts (via SeekingAlpha). "This additional leading-edge greenfield capacity, along with continued technology transition investments in our Asia facilities, is required to meet long-term demand in the second half of this decade and beyond. These investments support our objective to maintain our current bit share over time and to grow our memory bit supply in line with long-term industry bit demand."

Memory

Noctua Launches New Flagship Cooler: NH-D15 G2 with LGA1851 CPUs Support

On Tuesday, Noctua introduced its second-generation NH-D15 cooler, which offers refined performance and formally supports Intel's next-generation Arrow Lake-S processors in LGA1851 packaging. Alongside its NH-D15 G2 CPU cooler, Noctua also introduced its NF-A14x25r G2 140mm fans. 

The Noctua NH-D15 G2 is an enhanced version of the popular NH-D15 cooler with eight heat pipes, two asymmetrical fin-stack and two speed-offset 140-mm PWM fans (to avoid acoustic interaction phenomena such as periodic humming or intermittent vibrations). According to the manufacturer, these key components are tailored to work efficiently together to deliver superior quiet cooling performance, rivalling many all-in-one water cooling systems and pushing the boundaries of air cooling efficiency.

Noctua offers the NH-D15 G2 in three versions to address the specific requirements of modern CPUs. The regular version is versatile and can be used for AMD's AM5 processors and Intel's LGA1700 CPUs with included mounting accessories. The HBC (High Base Convexity) variant is tailored for LGA1700 processors, especially those subjected to full ILM pressure or those that have deformed over time, ensuring excellent contact quality despite the concave shape of the CPU. Finally, the LBC (Low Base Convexity) version is tailored for flat rectangular CPUs, providing optimal contact on AMD's AM5 and other similar processors.

While there are three versions of NH-D15 G2 aimed at different processors, they are all said to be compatible with a wide range of motherboards and other hardware. The new coolers' offset construction ensures clearance for the top PCIe x16 slot on most current motherboards. Additionally, they feature the upgraded Torx-based SecuFirm2+ multi-socket mounting system and come with Noctua's NT-H2 thermal compound.

For those looking to upgrade existing coolers like the NH-D15, NH-D15S, or NH-U14S series, Noctua is also releasing the NF-A14x25r G2 fans separately. These round-frame fans are fine-tuned in single and dual fan packages to minimize noise levels while offering decent cooling performance.

Finally, Noctua is also prepping a square-frame version of the NF-A14x25 G2 fan for release in September. This variant targets water-cooling radiators and case-cooling applications and promises to extend the versatility of Noctua's cooling solutions further.

All versions of Noctua's NH-D15 G2 coolers cost $149.90/€149.90. One NF-A14x25 G2 fan costs $39.90/€39.90, whereas a package of two fans costs $79.80/€79.80. The cooler is backed with a six-year warranty.

Cases/Cooling/PSUs

Best CPUs for Gaming: July 2024

As the second quarter of 2024 is soon set to unfold, there are many things to be excited about, especially as Computex 2024 has been and gone. We now know that AMD's upcoming Ryzen 9000 series desktop processors using the new Zen 5 cores will be hitting shelves at the end of the month (31st July), and on top of this, AMD also recently slashed pricing on their Zen 4 (Ryzen 8000) processors. Intel still needs to follow suit with their 14th or 13th Gen Core series processors, but right now from a cost standpoint, AMD is in a much better position.

Since the publication of our last guide, the only notable CPU to be launched was Intel's special binned Core i9-14900KS, which not only pushes clock speeds up to 6.2 GHz but is the last processor to feature Intel's iconic Core I series nomenclature. The other big news in the CPU world was from Intel, with a statement issued about pushing users to use the Intel Default Specification on Intel's 14th and 13th Gen processors, which ultimately limits the performance compared to published data. We're still in the process of 

While the CPU market has been relatively quiet so far this year, and things are set to pick up once AMD's Zen 5 and Intel's Arrow Lake desktop chips are all launched onto the market, it means today we are working for the same hymn sheet as our previous guide. With AMD's price drops on Ryzen 7000 series processors, much of the guide reflects this as AMD and Intel's performance is neck and neck in many use cases, but cost certainly plays a big factor in selecting a new CPU. As we move into the rest of 2024, the CPU market looks set to see the rise of the 'AI PC,' which is looking set to be something that many companies will focus on by the end of 2024, both on mobile and desktop platforms.

Guides

NVIDIA's AD102 GPU Pops Up in MSI GeForce RTX 4070 Ti Super Cards

As GPU families enter the later part of their lifecycles, we often see chip manufacturers start to offload stockpiles of salvaged chips that, for one reason or another, didn't make the grade for the tier of cards they normally are used in. These recovered chips are fairly unremarkable overall, but they are unsold silicon that still works and has economic value, leading to them being used in lower-tier cards so that they can be sold. And, judging by the appearance of a new video card design from MSI, it looks like NVIDIA's Ada Lovelace generation of chips has reached that stage, as the Taiwanese video card maker has put out a new GeForce RTX 4070 Ti Super card based on a salvaged AD102 GPU.

Typically based on NVIDIA's AD103 GPU, NVIDIA's GeForce RTX 4070 Ti Super series sits a step below the company's flagship RTX 4080/4090 cards, both of which are based on the bigger and badder AD102 chip. But with some number of AD102 chips inevitably failing to live up to RTX 4080 specifications, rather than being thrown out, these chips can instead be used to make RTX 4070 cards. Which is exactly what MSI has done with their new GeForce RTX 4070 Ti Super Ventus 3X Black OC graphics card.

The card itself is relatively unremarkable – using a binned AD102 chip doesn't come with any advantages, and it should perform just like regular AD103 cards – and for that reason, video card vendors rarely publicly note when they're doing a run of cards with a binned-down version of a bigger chip. However, these larger chips have a tell-tale PCB footprint that usually makes it obvious what's going on. Which, as first noticed by @wxnod, is exactly what's going on with MSI's card.

Ada Lovelace Lineup: MSI GeForce RTX 4070 TiS (AD103), RTX 4070 TiS (AD102), & RTX 4090 (AD102)

The tell, in this case, is the rear board shot provided by MSI. The larger AD102 GPU uses an equally larger mounting bracket, and is paired with a slightly more complex array of filtering capacitors on the back side of the board PCB. Ultimately, since these are visible in MSI's photos of their GeForce RTX 4070 Ti Super Ventus 3X Black OC, it's easy to compare it to other video cards and see that it has exactly the same capacitor layout as MSI's GeForce RTX 4090, thus confirming the use of an AD102 GPU.

Chip curiosities aside, all of NVIDIA GeForce RTX 4070 Ti Super graphics cards – no matter whether they are based on the AD102 or AD103 GPU – come with a GPU with 8,448 active CUDA cores and 16 GB of GDDR6X memory, so it doesn't (typically) matter which chip they carry. Otherwise, compared to a fully-enabled AD102 chip, the RTX 4070 Ti Super specifications are relatively modest, with fewer than half as many CUDA cores, underscoring how the AD102 chip being used in MSI's card is a pretty heavy salvage bin.

As for the rest of the card, MSI GeForce RTX 4070 Ti Super Ventus 3X Black OC is a relatively hefty card overall, with a cooling system to match. Being overclocked, the Ventus also has a slightly higher TDP than normal GeForce RTX 4070 Ti Super cards, weighing in at 295 Watts, or 10 Watts above baseline cards.

Meanwhile, MSI is apparently not the only video card manufacturer using salvaged AD102 chips for GeForce RTX 4070 Ti Super, either. @wxnod has also posted a screenshot obtained on an Inno3D GeForce RTX 4070 Ti Super based on an AD102 GPU.

Sources: MSI, @wxnod

GPUs

SK hynix to Enter 60 TB SSD Club Next Quarter

SK hynix this week reported its financial results for the second quarter, as well as offering a glimpse at its plans for the coming quarters. Notably among the company's plans for the year is the release of a SK hynix-branded 60 TB SSD, which will mark the firm's entry into the ultra-premium enterprise SSD league.

"SK hynix plans to expand sales of high-capacity eSSD and lead the market in the second half with 60TB products, expecting eSSD sales to be more than quadrupled compared to last year," a statement by SK hynix reads.

Currently there are only two standard form-factor 61.44 TB SSDs on the market: the Solidigm D5-P5336 (U.2/15mm and E1.L), and the Samsung BM1743 (U.2/15mm and E3.S). Both are built from a proprietary controller (Solidigm's controller still carries an Intel logotype) with a PCIe 4.0 x4 interface and use QLC NAND for storage.

SK hynix's brief mention of the drive means that tere aren't any formal specifications or capabilities to discuss just yet. But it is reasonable to assume that the company will use its own QLC memory for their ultra-high-capacity drives. What's more intriguing are which controller the company plans to use and how it is going to position its 60 TB-class SSD.

Internally, SK hynix has access to multiple controller teams, both of which have the expertise to develop an enterprise-grade controller suitable for a 60 TB drive. SK hynix technically owns Solidigm, the former Intel SSD and NAND unit, giving SK hynix the option of using Solidigm's controller, or even reselling a rebadged D5-P5336 outright. Alternatively, SK hynix has their own (original) internal SSD team, who is responsible for building their well-received Aries SSD controller, among other works.

Ultra-high-capacity SSDs for performance demanding read-intensive storage applications, such as AI inference on the edge or content delivery networks, is a promising premium market. So SK hynix is finding itself highly incentivized to enter it with a compelling offering. 

SSDs

SK hynix Wraps up Dev Work on High-End PCB01 PCIe 5.0 SSD for OEMs, Launching Later This Year

SK hynix early in Friday announced that the company has finished the development of it's PCB01 PCIe Gen5 SSD, the company's forthcoming high-end SSD for OEMs. Based on the company's new Alistar platform, the PCB01 is designed to deliver chart-topping performance for client machines. And, as a sign of the times, SK hynix is positioning the PCB01 for AI PCs, looking to synergize with the overall industry interest in anything and everything AI.

The bare, OEM-focused drives have previously been shown off by SK hynix, and make no attempt to hide what's under the hood. The PCB01 relies on SK hynix's Alistar controller, which features a PCIe Gen5 x4 host interface on the front end and eight NAND channels on the back end, placing it solidly in the realm of high-end SSDs. Paired with the Alistar controller is the company's latest 238-layer TLC NAND (H25T1TD48C & H25T2TD88C), which offers a maximum transfer speed of 2400 MT/second. Being that this is a high-end client SSD, there's also a DRAM chip on board, though the company isn't disclosing its capacity.

As with other high-end PCIe 5.0 client SSDs, SK hynix is planning on hitting peak read speeds of up to 14GB/second on the drive, while peak sequential write speeds should top 12GB/second (with pSLC caching, of course) – performance figures well within the realm of possibility for an 8 channel drive. As for random performance, at Computex the company was telling attendees that the drives should be able to sustain 4K random read and write rates of 2 million IOPS, which is very high as well. The SSDs are also said to consume up to 30% less power than 'predecessors,' according to SK hynix, though the company didn't elaborate on that figure. Typically in the storage industry, energy figures are based on iso-performance (rather than peak performance) – essentially measuring energy efficiency per bit rather than toal power consumption – and that is likely the case here as well.

At least initially, SK Hynix plans to release its PCB01 in three capacities – 512 GB, 1 TB, and 2 TB. The company has previously disclosed that their 238L TLC NAND has a capacity of 512Gbit, so these are typical capacity figures for single-sided drives. And while the focus of the company's press release this week was on OEM drives, this is the same controller and NAND that is also going into the company's previously-teased retail Platinum P51 SSD, so this week's reveal offers a bit more detail into what to expect from that drive family as well.

Specs aside, Ahn Hyun, the Head of the N-S Committee at SK hynix, said that multiple global CPU providers for on-device AI PCs are seeking collaboration for the compatibility validation process, which is underway, so expect PCB01 drives inside PCs in this back-to-school and holiday seasons.

"We will work towards enhancing our leadership as the global top AI memory provider also in the NAND solution space by successfully completing the customer validation and mass production of PCB01, which will be in the limelight," Ahn Hyun said.

SSDs

Two Is Better Than One: LG Starts Production of 13-inch Tandem OLED Display for Laptops

OLED panels have a number of advantages, including deep blacks, fast response times, and energy efficiency; most of these stemming from the fact that they do not need backlighting. However they also have drawbacks, as well, as trying to drive them to be as bright as a high-tier LCD will quickly wear out the organic material used. Researchers have been spending the past couple of decades developing ways to prolong the lifespans of OLED materials, and recently LG has put together a novel (if brute force) solution: halve the work by doubling the number of pixels. This is the basis of the company's new tandem OLED technology, which has recently gone into mass production.

The Tandem OLED technology introduced by LG Display uses two stacks of red, green, and blue (RGB) organic light-emitting layers, which are layered on top fo each other, essentially reducing how bright each layer needs to individually be in order to hit a specific cumulative brightness. By combining multiple OLED pixels running at a lower brightness, tandem OLED displays are intended to offer higher brightness and durability than traditional single panel OLED displays, reducing the wear on the organic materials in normal situations – and by extension, making it possible to crank up the brightness of the panels well beyond what a single panel could sustain without cooking itself. Overall, LG claims that tandem panels can hit over three-times the brightness of standard OLED panels.

The switch to tandem panels also comes with energy efficiency benefits, as the power consumption of OLED pixels is not linear with the output brightness.  According to LG, their tandem panels consume up to 40% less power. More interesting from the manufacturing side of matters, LG's tandem panel stack is 40% thinner (and 28%) lighter than existing OLED laptop screens, despite having to get a whole second layer of pixels in there.

In terms of specifications, the 13-inch tandem OLED panel feature a WQXGA+ (2880×1800) resolution and can cover 100% of the DCI-P3 color gamut. The panel is also certified to meet VESA's Display HDR True Black 500 requirements, which among other things, requires that it can hit 500 nits of brightness. And given that this tech is meant to go into tablets and laptops, it shouldn't come as any surprise that the display panel is also touch sensitive, as well.

"We will continue to strengthen the competitiveness of OLED products for IT applications and offer differentiated customer value based on distinctive strengths of Tandem OLED, such as long life, high brightness, and low power consumption," said Jae-Won Jang, Vice President and Head of the Medium Display Product Planning Division at LG Display.

Without any doubts, LG's Tandem OLED display panel looks impressive. The company is banking on it doing well in the high-end laptop and tablet markets, where manufacturers have been somewhat hesitant to embrace OLED displays due to power concerns. The technology has already been adopted by Apple for their most recent iPad Pro tablets, and now LG is making it available to a wider group of OEMs.

What remains to be seen is the technology's cost. Computer-grade OLED panels are already a more expensive  option, and this one ups the ante with two layers of OLED pixels. So it isn't a question of whether it will be reserved for premium, high-margin devices, but a matter of just how much it will add to the final price tag.

For now, LG Display does not disclose which PC OEMs are set to use its 13-inch Tandem OLED panel, though as the company is a supplier to virtually all of the PC OEMs, there's little doubt it should crop up in multiple laptops soon enough.

Displays

SK hynix to Enter 60 TB SSD Club Next Quarter

SK hynix this week reported its financial results for the second quarter, as well as offering a glimpse at its plans for the coming quarters. Notably among the company's plans for the year is the release of a SK hynix-branded 60 TB SSD, which will mark the firm's entry into the ultra-premium enterprise SSD league.

"SK hynix plans to expand sales of high-capacity eSSD and lead the market in the second half with 60TB products, expecting eSSD sales to be more than quadrupled compared to last year," a statement by SK hynix reads.

Currently there are only two standard form-factor 61.44 TB SSDs on the market: the Solidigm D5-P5336 (U.2/15mm and E1.L), and the Samsung BM1743 (U.2/15mm and E3.S). Both are built from a proprietary controller (Solidigm's controller still carries an Intel logotype) with a PCIe 4.0 x4 interface and use QLC NAND for storage.

SK hynix's brief mention of the drive means that tere aren't any formal specifications or capabilities to discuss just yet. But it is reasonable to assume that the company will use its own QLC memory for their ultra-high-capacity drives. What's more intriguing are which controller the company plans to use and how it is going to position its 60 TB-class SSD.

Internally, SK hynix has access to multiple controller teams, both of which have the expertise to develop an enterprise-grade controller suitable for a 60 TB drive. SK hynix technically owns Solidigm, the former Intel SSD and NAND unit, giving SK hynix the option of using Solidigm's controller, or even reselling a rebadged D5-P5336 outright. Alternatively, SK hynix has their own (original) internal SSD team, who is responsible for building their well-received Aries SSD controller, among other works.

Ultra-high-capacity SSDs for performance demanding read-intensive storage applications, such as AI inference on the edge or content delivery networks, is a promising premium market. So SK hynix is finding itself highly incentivized to enter it with a compelling offering. 

SSDs

Samsung Validates LPDDR5X Running at 10.7 GT/sec with MediaTek's Dimensity 9400 SoC

Samsung has successfully validated its new LPDDR5X-10700 memory with MediaTek's upcoming Dimensity platform. At present, 10.7 GT/s is the highest performing speed grade of LPDDR5X DRAM slated to be released this year, so the upcoming Dimensity 9400 system-on-chip will get the highest memory bandwidth available for a mobile application processor.

The verification process involved Samsung's 16 GB LPDDR5X package and MediaTek's soon-to-be-announced Dimensity 9400 SoC for high-end 5G smartphones. Usage of LPDDR5X-10700 provides a memory bandwidth of 85.6 GB/second over a 64-bit interface, which will be available for bandwidth-hungry applications like graphics and generative AI.

"Working together with Samsung Electronics has made it possible for MediaTek's next-generation Dimensity chipset to become the world's first to be validated at LPDDR5X operating speeds up to 10.7Gbps, enabling upcoming devices to deliver AI functionality and mobile performance at a level we have never seen before," said JC Hsu, Corporate Senior Vice President at MediaTek. "This updated architecture will make it easier for developers and users to leverage more AI capabilities and take advantage of more features with less impact on battery life."

Samsung's LPDDR5X 10.7 GT/s memory in made on the company's 12nm-class DRAM process technology and is said to provide a more than 25% improvement in power efficiency over previous-generation LPDDR5X, in addition to extra performance. This will positively affect improved user experience, including enhanced on-device AI capabilities, such as faster voice-to-text conversion, and better quality graphics.

Overall, the two companies completed this process in just three months. Though it remains to be seen when smartphones based on the Dimensity 9400 application processor and LPDDR5X memory are set to be available on the market, as MediaTek has not yet even formally announced the SoC itself.

"Through our strategic cooperation with MediaTek, Samsung has verified the industry's fastest LPDDR5X DRAM that is poised to lead the AI smartphone market," said YongCheol Bae, Executive Vice President of Memory Product Planning at Samsung Electronics. "Samsung will continue to innovate through active collaboration with customers and provide optimum solutions for the on-device AI era."

Memory

Intel Addresses Desktop Raptor Lake Instability Issues: Faults Excessive Voltage from Microcode, Fix Coming in August

In what started last year as a handful of reports about instability with Intel's Raptor Lake desktop chips has, over the last several months, grown into a much larger saga. Facing their biggest client chip instability impediment in decades, Intel has been under increasing pressure to figure out the root cause of the issue and fix it, as claims of damaged chips have stacked up and rumors have swirled amidst the silence from Intel. But, at long last, it looks like Intel's latest saga is about to reach its end, as today the company has announced that they've found the cause of the issue, and will be rolling out a microcode fix next month to resolve it.

Officially, Intel has been working to identify the cause of desktop Raptor Lake’s instability issues since at least February of this year, if not sooner. In the interim they have discovered a couple of correlating factors – telling motherboard vendors to stop using ridiculous power settings for their out-of-the-box configurations, and finding a voltage-related bug in Enhanced Thermal Velocity Boost (eTVB) – but neither factor was the smoking gun that set all of this into motion. All of which had left Intel to continue searching for the root cause in private, and lots of awkward silence to fill the gaps in the public.

But it looks like Intel’s search has finally come to an end – even if Intel isn’t putting the smoking gun on public display quite yet. According to a fresh update posted to the company’s community website, Intel has determined the root cause at last, and has a fix in the works.

Per the company’s announcement, Intel has tracked down the cause of the instability issue to “elevated operating voltages”, that at its heart, stems from a flawed algorithm in Intel’s microcode that requested the wrong voltage. Consequently, Intel will be able to resolve the issue through a new microcode update, which pending validation, is expected to be released in the middle of August.

Based on extensive analysis of Intel Core 13th/14th Gen desktop processors returned to us due to instability issues, we have determined that elevated operating voltage is causing instability issues in some 13th/14th Gen desktop processors. Our analysis of returned processors confirms that the elevated operating voltage is stemming from a microcode algorithm resulting in incorrect voltage requests to the processor.

Intel is delivering a microcode patch which addresses the root cause of exposure to elevated voltages. We are continuing validation to ensure that scenarios of instability reported to Intel regarding its Core 13th/14th Gen desktop processors are addressed. Intel is currently targeting mid-August for patch release to partners following full validation.

Intel is committed to making this right with our customers, and we continue asking any customers currently experiencing instability issues on their Intel Core 13th/14th Gen desktop processors reach out to Intel Customer Support for further assistance.
-Intel Community Post

And while there’s nothing good for Intel about Raptor Lake’s instability issues or the need to fix them, that the problem can be ascribed to (or at least fixed by) microcode is about the best possible outcome the company could hope for. Across the full spectrum of potential causes, microcode is the easiest to fix at scale – microcode updates are already distributed through OS updates, and all chips of a given stepping (millions in all) run the same microcode. Even a motherboard BIOS-related issue would be much harder to fix given the vast number of different boards out there, never mind a true hardware flaw that would require Intel to replace even more chips than they already have.

Still, we’d also be remiss if we didn’t note that microcode is regularly used to paper over issues further down in the processor, as we’ve most famously seen with the Meltdown/Spectre fixes several years ago. So while Intel is publicly attributing the issue to microcode bugs, there are several more layers to the onion that is modern CPUs that could be playing a part. In that respect, a microcode fix grants the least amoun... CPUs

The Lian Li Hydroshift LCD 360S AIO Cooler Review: Sleek, Stylish, and Lively

Among the packed field of PC hardware manufacturers, Lian Li is a company that arguably shouldn't even need an introduction. The quirky company has developed a devout following thanks to its focus on premium-quality aluminum computer cases that, more often that not, and come with in some rather unique designs. Over the years, the company has developed a solid reputation for its meticulous craftsmanship, durability, and elegant aesthetics. And consequently, when the company made the decision to expand beyond aluminum cases and in to other PC peripherals, that development attracted quite a bit of attention to see what kind of a touch Lian Li could bring to the rest of the PC ecosystem.

Lian Li's focus on premium products means that the company doesn't really make much in the way of products that are merely basic, and that kind of mentality has extended beyond cases and into the rest of their peripherals. Case in point is the subject of today's review: Lian Li's new all-in-one CPU cooler, the HydroShift LCD 360S AIO. Not content just to make a powerful 360 mm cooler, Lian Li has gone a step above by integrating recesses and other features to help hide the tubing around the cooler, and then for the coup de grace, added a high-quality 2.88-inch IPS display to the pump block.

Overall, this new product marks a significant milestone for Lian Li, as it combines advanced cooling technology with the company's signature aesthetic appeal, making for a cooler that's aimed at both enthusiasts and professional users seeking high-end thermal performance and visual customization. Overall, the HydroShift LCD 360S is undeniably designed first and foremost with aesthetics in mind, but the shiny pump block is backed up with one of the most powerful 360 mm cooler designs on the market today. So Lian Li is throwing everything they have at the new HydroShift coolers.

Overall, the 360S is part of a trio of HydroShift 360 mm coolers the company is launching this summer. All three share a similar design, although with some pump changes and the addition of RGB fan lighting, depending on the specific model, with the 360S effectively serving as the base model.

Cases/Cooling/PSUs

The Lian Li EDGE EG1000 1000W ATX 3.1 PSU Review: Power On The Edge

Lian Li Industrial Co., Ltd., established in 1983, is a Taiwanese company specializing in the manufacture of computer cases, power supplies, and accessories. They are one of the oldest players in the PC market and are known for their focus on aluminum-based designs. Lian Li produces a range of products aimed at both consumer and industrial markets, with the company's offerings including mid-tower and full-tower cases and more compact cases for smaller builds. Amongst consumers and PC enthusiasts, Lian Li's products are recognized for their build quality, modularity, and innovative features, catering to a diverse set of needs in the PC building community.

This review focuses on the latest addition to Lian Li's PSU lineup: the EG1000 Platinum ATX 3.1 PSU. This power supply unit partially complies with the ATX 3.1 design guide (the paragraphs related to electrical quality and performance). It is designed to meet the demanding requirements of modern gaming PCs, with its specifications indicating good efficiency and robust power delivery. Featuring fully modular cables with individually sleeved wires, dynamic fan control for optimal cooling, and advanced internal topologies, the EG1000 Platinum aims to provide both reliability and performance. However, behind its long list of features, the highlight of the EG1000 Platinum is the shape of the chassis itself, which forgoes the ATX cuboid shape and standard length.

Cases/Cooling/PSUs

Tenstorrent Launches Wormhole AI Processors: 466 FP8 TFLOPS at 300W

Tenstorrent has unveiled its next-generation Wormhole processor for AI workloads that promises to offer decent performance at a low price. The company currently offers two add-on PCIe cards carrying one or two Wormhole processors as well as TT-LoudBox, and TT-QuietBox workstations aimed at software developers. The whole of today's release is aimed at developers rather than those who will deploy the Wormhole boards for their commercial workloads.

“It is always rewarding to get more of our products into developer hands. Releasing development systems with our Wormhole™ card helps developers scale up and work on multi-chip AI software.” said Jim Keller, CEO of Tenstorrent. “In addition to this launch, we are excited that the tape-out and power-on for our second generation, Blackhole, is going very well.”

Each Wormhole processor packs 72 Tensix cores (featuring five RISC-V cores supporting various data formats) with 108 MB of SRAM to deliver 262 FP8 TFLOPS at 1 GHz at 160W thermal design power. A single-chip Wormhole n150 card carries 12 GB of GDDR6 memory featuring a 288 GB/s bandwidth.

Wormhole processors offer flexible scalability to meet the varying needs of workloads. In a standard workstation setup with four Wormhole n300 cards, the processors can merge to function as a single unit, appearing as a unified, extensive network of Tensix cores to the software. This configuration allows the accelerators to either work on the same workload, be divided among four developers or run up to eight distinct AI models simultaneously. A crucial feature of this scalability is that it operates natively without the need for virtualization. In data center environments, Wormhole processors will scale both inside one machine using PCIe or outside of a single machine using Ethernet. 

From performance standpoint, Tenstorrent's single-chip Wormhole n150 card (72 Tensix cores at 1 GHz, 108 MB SRAM, 12 GB GDDR6 at 288 GB/s) is capable of 262 FP8 TFLOPS at 160W, whereas the dual-chip Wormhole n300 board (128 Tensix cores at 1 GHz, 192 MB SRAM, aggregated 24 GB GDDR6 at 576 GB/s) can offer up to 466 FP8 TFLOPS at 300W (according to Tom's Hardware).

To put that 466 FP8 TFLOPS at 300W number into context, let's compare it to what AI market leader Nvidia has to offer at this thermal design power. Nvidia's A100 does not support FP8, but it does support INT8 and its peak performance is 624 TOPS (1,248 TOPS with sparsity). By contrast, Nvidia's H100 supports FP8 and its peak performance is massive 1,670 TFLOPS (3,341 TFLOPS with sparsity) at 300W, which is a big difference from Tenstorrent's Wormhole n300. 

There is a big catch though. Tenstorrent's Wormhole n150 is offered for $999, whereas n300 is available for $1,399. By contrast, one Nvidia H100 card can retail for $30,000, depending on quantities. Of course, we do not know whether four or eight Wormhole processors can indeed deliver the performance of a single H300, though they will do so at 600W or 1200W TDP, respectively.

In addition to cards, Tenstorrent offers developers pre-built workstations with four n300 cards inside the less expensive Xeon-based TT-LoudBox with active cooling and a premium EPYC-powered TT-QuietBox with liquid cooling.

Sources: Tenstorrent, Tom's Hardware

AI

Kioxia's High-Performance 3D QLC NAND Enables High-End High-Capacity SSDs

This week, Kioxia introduced its new 3D QLC NAND devices aimed at high-performance, high-capacity drives that could redefine what we typically expect from QLC-based SSDs. The components are 1 Tb and 2 Tb 3D QLC NAND ICs with a 3600 MT/s interface speed that could enable M.2-2230 SSDs with a 4 TB capacity and decent performance.

Kioxia's 1 Tb (128 MB) and 2 Tb (256 TB) 3D QLC NAND devices are made on the company's BICS 8 process technology and feature 238 active layers as well as CMOS directly Bonded to Array (CBA) design, which implies that CMOS (including interface and buffers circuitry) is built on a specialized node and bonded to the memory array. Such a manufacturing process enabled Kioxia (and its manufacturing partner Western Digital) to achieve a particularly high interface speed of 3600 MT/s.

In addition to being one of the industry's first 2 Tb QLC NAND devices, the component features a 70% higher write power efficiency compared to Kioxia's BICS 5 3D QLC NAND devices, which is a bit vague statement as the new ICs have higher capacity and performance in general. This feature will be valuable for data centre applications, though I do not expect someone to use 3D QLC memory for write-intensive applications in general. Yet, these devices will be just what the doctor ordered for AI: read-intensive, content distribution, and backup storage.

It is interesting to note that Kioxia's 1 Tb 3D QLC NAND, optimized for performance, has a 30% faster sequential write performance and a 15% lower read latency than the 2 Tb 3D QLC component. These qualities (alongside a 3600 MT/s interface) promise to make Kioxia's 1 Tb 3D QLC competitive even for higher-end PCIe Gen5 x4 SSDs, which currently exclusively use 3D TLC memory.

The remarkable storage density of Kioxia's 2Tb 3D QLC NAND devices will allow customers to create high-capacity SSDs in compact form factors. For instance, a 16-Hi stacked package (measuring 11.5 mm × 13.5 mm × 1.5 mm) can be used to build a 4TB M.2-2230 drive or a 16TB M.2-2280 drive. Even a single 16-Hi package could be enough to build a particularly fast client SSD.

Kioxia is now sampling its 2 Tb 3D QLC NAND BiCS 8 memory with customers, such as Pure Storage.

"We have a long-standing relationship with Kioxia and are delighted to incorporate their eighth-generation BiCS Flash 2Tb QLC flash memory products to enhance the performance and efficiency of our all-flash storage solutions," said Charles Giancarlo, CEO of Pure Storage. "Pure's unified all-flash data storage platform is able to meet the demanding needs of artificial intelligence as well as the aggressive costs of backup storage. Backed by Kioxia technology, Pure Storage will continue to offer unmatched performance, power efficiency, and reliability, delivering exceptional value to our customers."

"We are pleased to be shipping samples of our new 2Tb QLC with the new eighth-generation BiCS flash technology," said Hideshi Miyajima, CTO of Kioxia. "With its industry-leading high bit density, high speed data transfer, and superior power efficiency, the 2Tb QLC product will offer new value for rapidly emerging AI applications and large storage applications demanding power and space savings."

There is no word on when the 1 Tb 3D QLC BiCS 8 memory will be sampled or released to the market.

SSDs

Frore Unveils Waterproof AirJet Mini Sport for Smartphones

Over the past couple of years, Frore Systems has demonstrated several ways that its AirJet solid-state active cooling systems can be used to improve cooling in fanless devices like laptops, tablets, SSDs, and edge computing devices. But there are a subset of those applications that need their cooling options to also be waterproof, and Frore is looking to address those as well. To that end, this week Frore introduced its AirJet Mini Sport, a waterproof, IP68-rated solid-state cooling device that is aimed at use in smartphones and action cameras.

Introduced at MWC Shanghai to attract attention of China-based handset vendors, edge and industrial computing devices, and action cameras, the AirJet Mini Sport is an enhanced version of Frore's AirJet Mini Slim. This version has been fully waterproofed, offering IP68-level protection that allows it to work while being submerged in over 1.5 meters of water for up to 30 minutes. Internally, the AirJet Mini Sport can effectively dissipate 5.25 Watts of heat by generating 1750 Pascals of back pressure, while consuming 1 Watt of energy itself.

Elsewhere, Frore claims that the AirJet Mini Sport can be used to provide 2.5 Watts of cooling capacity to smartphones. Which, although not enough to cover the complete power consumption/heat dissipation of a high-end SoC, would have a significant impact on both burst and steady-state performance by allowing those chips to run at peak clocks and power for longer periods of time.

To ensure consistent performance of Frore's AirJet Mini Sport in diverse environments, the cooling device includes features such as dust resistance and self-cleaning. In addition, just like AirJet Mini Slim, the Sport-badged version its own thermal sensor to control its own operation and maintain optimal performance. As a result, Frore claims that smartphones and action cameras with the AirJet Mini Sport can achieve up to 80% better performance.

"We are excited to announce the waterproof AirJet Mini Sport," said Dr. Seshu Madhavapeddy, founder and CEO of Frore Systems. "Consumers demand increased performance in compact devices they can use anywhere, on land or in water. AirJet unleashes device performance, now enabling users to do more with their IP68 dustproof and waterproof devices."

Air Cooling

Micron Expands Datacenter DRAM Portfolio with MR-DIMMs

The compute market has always been hungry for memory bandwidth, particularly for high-performance applications in servers and datacenters. In recent years, the explosion in core counts per socket has further accentuated this need. Despite progress in DDR speeds, the available bandwidth per core has unfortunately not seen a corresponding scaling.

The stakeholders in the industry have been attempting to address this by building additional technology on top of existing widely-adopted memory standards. With DDR5, there are currently two technologies attempting to increase the peak bandwidth beyond the official speeds. In late 2022, SK hynix introduced MCR-DIMMs meant for operating with specific Intel server platforms. On the other hand, JEDEC - the standards-setting body - also developed specifications for MR-DIMMs with a similar approach. Both of them build upon existing DDR5 technologies by attempting to combine multiple ranks to improve peak bandwidth and latency.

How MR-DIMMs Work

The MR-DIMM standard is conceptually simple - there are multiple ranks of memory modules operating at standard DDR5 speeds with a data buffer in front. The buffer operates at 2x the speed on the host interface side, allowing for essentially double the transfer rates. The challenges obviously lie in being able to operate the logic in the host memory controller at the higher speed and keeping the power consumption / thermals in check.

The first version of the JEDEC MR-DIMM standard specifies speeds of 8800 MT/s, with the next generation at 12800 MT/s. JEDEC also has a clear roadmap for this technology, keeping it in sync with the the improvements in the DDR5 standard.

Micron MR-DIMMs - Bandwidth and Capacity Plays

Micron and Intel have been working closely in the last few quarters to bring their former's first-generation MR-DIMM lineup to the market. Intel's Xeon 6 Family with P-Cores (Granite Rapids) is the first platform to bring MR-DIMM support at 8800 MT/s on the host side. Micron's standard-sized MR-DIMMs (suitable for 1U servers) and TFF (tall form-factor) MR-DIMMs (for 2U+ servers) have been qualified for use with the same.

The benefits offered by MR-DIMMs are evident from the JEDEC specifications, allowing for increased data rates and system bandwidth, with improvements in latency. On the capacity side, allowing for additional ranks on the modules has enabled Micron to offer a 256 GB capacity point. It must be noted that some vendors are also using TSV (through-silicon vias) technology to to increase the per-package capacity at standard DDR5 speeds, but this adds additional cost and complexity that are largely absent in the MR-DIMM manufacturing process.

The tall form-factor (TFF) MR-DIMMs have a larger surface area compared to the standard-sized ones. For the same airflow configuration, this allows the DIMM to have a better thermal profile. This provides benefits for energy efficiency as well by reducing the possibility of thermal throttling.

Micron is launching a comprehensive lineup of MR-DIMMs in both standard and tall form-factors today, with multiple DRAM densities and speed options as noted above.

MRDIMM Benefits - Intel Granite Rapids Gets a Performance Boost

Micron and Intel hosted a media / analyst briefing recently to demonstrate the benefits of MR-DIMMs for Xeon 6 with P-Cores (Granite Rapids). Using a 2P configuration with 96-core Xeon 6 processors, benchmarks for different ... Memory

The AMD Ryzen AI 9 HX 370 Review: Unleashing Zen 5 and RDNA 3.5 Into Notebooks

During the opening keynote delivered by AMD CEO Dr. Lisa Su at Computex 2024, AMD finally lifted the lid on their highly-anticipated Zen 5 microarchitecture. The backbone for the next couple of years of everything CPU at AMD, the company unveiled their plans to bring Zen 5 in the consumer market, announcing both their next-generation mobile and desktop products at the same time. With a tight schedule that will see both platforms launch within weeks of each other, today AMD is taking their first step with the launch of the Ryzen AI 300 series – codenamed Strix Point – their new Zen 5-powered mobile SoC.

The latest and greatest from AMD, the Strix Point brings significant architectural improvements across AMD's entire IP portfolio. Headlining the chip, of course, is the company's new Zen 5 CPU microarchitecture, which is taking multiple steps to improve on CPU performance without the benefits of big clockspeed gains. And reflecting the industry's current heavy emphasis on AI performance, Strix Point also includes the latest XDNA 2-based NPU, which boasts up to 50 TOPS of performance. Other improvements include an upgraded integrated graphics processor, with AMD moving to the RDNA 3.5 graphics architecture.

The architectural updates in Strix Point are also seeing AMD opt for a heterogenous CPU design from the very start, incorporating both performance and efficiency cores as a means of offering better overall performance in power-constrained devices. AMD first introduced their compact Zen cores in the middle of the Zen 4 generation, and while they made it into products such as AMD's small-die Phoenix 2 platform, this is the first time AMD's flagship mobile silicon has included them as well. And while this change is going to be transparent from a user perspective, under the hood it represents an important improvement in CPU design. As a result, all Ryzen AI 300 chips are going to include a mix of not only AMD's (mostly) full-fat Zen 5 CPU cores, but also their compact Zen 5c cores, boosting the chips' total CPU core counts and performance in multi-threaded situations.

For today's launch, the AMD Ryzen AI 300 series will consist of just three SKUs: the flagship Ryzen AI 9 HX 375, with 12 CPU cores, as well as the Ryzen AI 9 HX 370 and Ryzen 9 365, with 12 and 10 cores respectively. All three SoCs combine both the regular Zen 5 core with the more compact Zen 5c cores to make up the CPU cluster, and are paired with a powerful Raden 890M/880M GPU, and a XDNA 2-based NPU.

As the successor to the Zen 4-based Phoenix/Hawk Point, the AMD Ryzen AI 300 series is targeting a diverse and active notebook market that has become the largest segment of the PC industry overall. And it is telling that, for the first time in the Zen era, AMD is launching their mobile chips first – if only by days – rather than their typical desktop-first launch. It's both a reflection on how the PC industry has changed over the years, and how AMD has continued to iterate and improve upon its mobile chips; this is as close to mobile-first as the company has ever been.

Getting down to business, for our review of the Ryzen AI 300 series, we are taking a look at ASUS's Zenbook S 16 (2024), a 16-inch laptop that's equipped with AMD's Ryzen AI 9 HX 370. The sightly more modest Ryzen features four Zen 5 CPU cores and 8 Zen 5c CPU cores, as well as AMD's latest RDNA 3.5 Radeon 890M integrated graphics. Overall, the HX 370 has a configurable TDP of between 15 and 54 W, depending on the desired notebook configuration.

Fleshing out the rest of the Zenbook S 16, ASUS has equipped the laptop with a bevy of features and technologies fitting for a flagship Ryzen notebook. The centerpiece of the laptop is a Lumina OLED 16-inch display, with a resolution of up to 2880 x 1800 and a variable 120 Hz refresh rate. Meanwhile, inside the Zenbook S 16 is 32 GB of LPDDR5 memory and a 1 TB PCIe 4.0 NVMe SSD. And while this is a 16-inch class notebook, ASUS has still designed it with an emphasis on portability, leading to the Zenbook S 16 coming in at 1.1 cm thick, and weighting 1.5 kg. That petite design also means ASUS has configured the Ryzen AI 9 HX 370 chip inside rather conservatively: out of the box, the chip runs at a TDP of just 17 Watts.

CPUs

The Enermax PlatiGemini 1200W ATX 3.1 + ATX12VO PSU Review: The Swiss Army Knife

In the retail PC PSU space, most of the focus on new standards and their capabilities in the past couple of years has been on ATX 3.0 and it's quick follow-up successor, ATX 3.1. And while the revised ATX standard is certainly the most important new standard for the rank-and-file PC builder, it's not the only standard that has been released as of late. Intel and its partners have also developed a standard, that in some respects, goes even farther out by dropping some of the legacy aspects of ATX and its increasingly esoteric secondary voltages: ATX12VO.

Short for "ATX 12 Volts Only", ATX12VO is a standard that's been slower to take off as it makes a pretty hard break with backwards compatibility. But with so many motherboard functions running off of 12V (CPUs and GPUs, for a start), the need for a PSU to provide secondary voltages like 3.3V and 5V just aren't what they once were 20 years ago - or even 10. So we've slowly seen PC manufacturers and motherboard makers test the waters, with a handful of designs using the more petite ATX standard.

Meanwhile on the power supply side of things, the outcome has been a bit more interesting, if messy. While ATX12VO motherboards need matching PSUs, there's nothing to say that such a PSU can only be ATX12VO. To reference an ancient meme, the thought at some PSU manufacturers has been "why not both?", leading to high-end PSUs that can bridge the compatibility gap by offering both ATX 3.1 and ATX12VO compatibility.

The first example of such a PSU to make it in our labs is Enermax's new PlatiGemini 1200W PSU. Designed to be the Swiss knife of modern top-tier PCs, Enermax's PSU offers support for both ATX 3.1 and ATX12VO - ensuring it can power virtually any PC - while driving both modes with a sizeable 1200W design that can pretty much power virtually any desktop PC one can hope to build today. Plus, with features like fully modular cables with per-wire sleeving, a dynamic hybrid fan control for optimal cooling, and advanced power topologies, the PlatiGemini 1200W aims to deliver both reliability and performance on top of its multi-mode compatibility. The end result is a very interesting (if premium) product that can do it all.

Cases/Cooling/PSUs

NVIDIA's AD102 GPU Pops Up in MSI GeForce RTX 4070 Ti Super Cards

As GPU families enter the later part of their lifecycles, we often see chip manufacturers start to offload stockpiles of salvaged chips that, for one reason or another, didn't make the grade for the tier of cards they normally are used in. These recovered chips are fairly unremarkable overall, but they are unsold silicon that still works and has economic value, leading to them being used in lower-tier cards so that they can be sold. And, judging by the appearance of a new video card design from MSI, it looks like NVIDIA's Ada Lovelace generation of chips has reached that stage, as the Taiwanese video card maker has put out a new GeForce RTX 4070 Ti Super card based on a salvaged AD102 GPU.

Typically based on NVIDIA's AD103 GPU, NVIDIA's GeForce RTX 4070 Ti Super series sits a step below the company's flagship RTX 4080/4090 cards, both of which are based on the bigger and badder AD102 chip. But with some number of AD102 chips inevitably failing to live up to RTX 4080 specifications, rather than being thrown out, these chips can instead be used to make RTX 4070 cards. Which is exactly what MSI has done with their new GeForce RTX 4070 Ti Super Ventus 3X Black OC graphics card.

The card itself is relatively unremarkable – using a binned AD102 chip doesn't come with any advantages, and it should perform just like regular AD103 cards – and for that reason, video card vendors rarely publicly note when they're doing a run of cards with a binned-down version of a bigger chip. However, these larger chips have a tell-tale PCB footprint that usually makes it obvious what's going on. Which, as first noticed by @wxnod, is exactly what's going on with MSI's card.

Ada Lovelace Lineup: MSI GeForce RTX 4070 TiS (AD103), RTX 4070 TiS (AD102), & RTX 4090 (AD102)

The tell, in this case, is the rear board shot provided by MSI. The larger AD102 GPU uses an equally larger mounting bracket, and is paired with a slightly more complex array of filtering capacitors on the back side of the board PCB. Ultimately, since these are visible in MSI's photos of their GeForce RTX 4070 Ti Super Ventus 3X Black OC, it's easy to compare it to other video cards and see that it has exactly the same capacitor layout as MSI's GeForce RTX 4090, thus confirming the use of an AD102 GPU.

Chip curiosities aside, all of NVIDIA GeForce RTX 4070 Ti Super graphics cards – no matter whether they are based on the AD102 or AD103 GPU – come with a GPU with 8,448 active CUDA cores and 16 GB of GDDR6X memory, so it doesn't (typically) matter which chip they carry. Otherwise, compared to a fully-enabled AD102 chip, the RTX 4070 Ti Super specifications are relatively modest, with fewer than half as many CUDA cores, underscoring how the AD102 chip being used in MSI's card is a pretty heavy salvage bin.

As for the rest of the card, MSI GeForce RTX 4070 Ti Super Ventus 3X Black OC is a relatively hefty card overall, with a cooling system to match. Being overclocked, the Ventus also has a slightly higher TDP than normal GeForce RTX 4070 Ti Super cards, weighing in at 295 Watts, or 10 Watts above baseline cards.

Meanwhile, MSI is apparently not the only video card manufacturer using salvaged AD102 chips for GeForce RTX 4070 Ti Super, either. @wxnod has also posted a screenshot obtained on an Inno3D GeForce RTX 4070 Ti Super based on an AD102 GPU.

Sources: MSI, @wxnod

GPUs

Noctua Launches New Flagship Cooler: NH-D15 G2 with LGA1851 CPUs Support

On Tuesday, Noctua introduced its second-generation NH-D15 cooler, which offers refined performance and formally supports Intel's next-generation Arrow Lake-S processors in LGA1851 packaging. Alongside its NH-D15 G2 CPU cooler, Noctua also introduced its NF-A14x25r G2 140mm fans. 

The Noctua NH-D15 G2 is an enhanced version of the popular NH-D15 cooler with eight heat pipes, two asymmetrical fin-stack and two speed-offset 140-mm PWM fans (to avoid acoustic interaction phenomena such as periodic humming or intermittent vibrations). According to the manufacturer, these key components are tailored to work efficiently together to deliver superior quiet cooling performance, rivalling many all-in-one water cooling systems and pushing the boundaries of air cooling efficiency.

Noctua offers the NH-D15 G2 in three versions to address the specific requirements of modern CPUs. The regular version is versatile and can be used for AMD's AM5 processors and Intel's LGA1700 CPUs with included mounting accessories. The HBC (High Base Convexity) variant is tailored for LGA1700 processors, especially those subjected to full ILM pressure or those that have deformed over time, ensuring excellent contact quality despite the concave shape of the CPU. Finally, the LBC (Low Base Convexity) version is tailored for flat rectangular CPUs, providing optimal contact on AMD's AM5 and other similar processors.

While there are three versions of NH-D15 G2 aimed at different processors, they are all said to be compatible with a wide range of motherboards and other hardware. The new coolers' offset construction ensures clearance for the top PCIe x16 slot on most current motherboards. Additionally, they feature the upgraded Torx-based SecuFirm2+ multi-socket mounting system and come with Noctua's NT-H2 thermal compound.

For those looking to upgrade existing coolers like the NH-D15, NH-D15S, or NH-U14S series, Noctua is also releasing the NF-A14x25r G2 fans separately. These round-frame fans are fine-tuned in single and dual fan packages to minimize noise levels while offering decent cooling performance.

Finally, Noctua is also prepping a square-frame version of the NF-A14x25 G2 fan for release in September. This variant targets water-cooling radiators and case-cooling applications and promises to extend the versatility of Noctua's cooling solutions further.

All versions of Noctua's NH-D15 G2 coolers cost $149.90/€149.90. One NF-A14x25 G2 fan costs $39.90/€39.90, whereas a package of two fans costs $79.80/€79.80. The cooler is backed with a six-year warranty.

Cases/Cooling/PSUs

Kioxia's High-Performance 3D QLC NAND Enables High-End High-Capacity SSDs

This week, Kioxia introduced its new 3D QLC NAND devices aimed at high-performance, high-capacity drives that could redefine what we typically expect from QLC-based SSDs. The components are 1 Tb and 2 Tb 3D QLC NAND ICs with a 3600 MT/s interface speed that could enable M.2-2230 SSDs with a 4 TB capacity and decent performance.

Kioxia's 1 Tb (128 MB) and 2 Tb (256 TB) 3D QLC NAND devices are made on the company's BICS 8 process technology and feature 238 active layers as well as CMOS directly Bonded to Array (CBA) design, which implies that CMOS (including interface and buffers circuitry) is built on a specialized node and bonded to the memory array. Such a manufacturing process enabled Kioxia (and its manufacturing partner Western Digital) to achieve a particularly high interface speed of 3600 MT/s.

In addition to being one of the industry's first 2 Tb QLC NAND devices, the component features a 70% higher write power efficiency compared to Kioxia's BICS 5 3D QLC NAND devices, which is a bit vague statement as the new ICs have higher capacity and performance in general. This feature will be valuable for data centre applications, though I do not expect someone to use 3D QLC memory for write-intensive applications in general. Yet, these devices will be just what the doctor ordered for AI: read-intensive, content distribution, and backup storage.

It is interesting to note that Kioxia's 1 Tb 3D QLC NAND, optimized for performance, has a 30% faster sequential write performance and a 15% lower read latency than the 2 Tb 3D QLC component. These qualities (alongside a 3600 MT/s interface) promise to make Kioxia's 1 Tb 3D QLC competitive even for higher-end PCIe Gen5 x4 SSDs, which currently exclusively use 3D TLC memory.

The remarkable storage density of Kioxia's 2Tb 3D QLC NAND devices will allow customers to create high-capacity SSDs in compact form factors. For instance, a 16-Hi stacked package (measuring 11.5 mm × 13.5 mm × 1.5 mm) can be used to build a 4TB M.2-2230 drive or a 16TB M.2-2280 drive. Even a single 16-Hi package could be enough to build a particularly fast client SSD.

Kioxia is now sampling its 2 Tb 3D QLC NAND BiCS 8 memory with customers, such as Pure Storage.

"We have a long-standing relationship with Kioxia and are delighted to incorporate their eighth-generation BiCS Flash 2Tb QLC flash memory products to enhance the performance and efficiency of our all-flash storage solutions," said Charles Giancarlo, CEO of Pure Storage. "Pure's unified all-flash data storage platform is able to meet the demanding needs of artificial intelligence as well as the aggressive costs of backup storage. Backed by Kioxia technology, Pure Storage will continue to offer unmatched performance, power efficiency, and reliability, delivering exceptional value to our customers."

"We are pleased to be shipping samples of our new 2Tb QLC with the new eighth-generation BiCS flash technology," said Hideshi Miyajima, CTO of Kioxia. "With its industry-leading high bit density, high speed data transfer, and superior power efficiency, the 2Tb QLC product will offer new value for rapidly emerging AI applications and large storage applications demanding power and space savings."

There is no word on when the 1 Tb 3D QLC BiCS 8 memory will be sampled or released to the market.

SSDs

AMD Delays Ryzen 9000 Launch 1 to 2 Weeks Due to Chip Quality Issues

AMD sends word this afternoon that the company is delaying the launch of their Ryzen 9000 series desktop processors. The first Zen 5 architecture-based desktop chips were slated to launch next week, on July 31^st. But citing quality issues that are significant enough that AMD is even pulling back stock already sent to distributors, AMD is delaying the launch by one to two weeks. The Ryzen 9000 launch will now be a staggered launch, with the Ryzen 5 9600X and Ryzen 7 9700X launching on August 8^th, while the Ryzen 9 9900X and flagship Ryzen 9 9950X will launch a week after that, on August 15^th.

The exceptional announcement, officially coming from AMD’s SVP and GM of Computing and Graphics, Jack Huynh, is short and to the point. Ahead of the launch, AMD found that “the initial production units that were shipped to our channel partners did not meet our full quality expectations.” And, as a result, the company has needed to delay the launch in order to rectify the issue.

Meanwhile, because AMD had already distributed chips to their channel partners – distributors who then filter down to retailers and system builders – this is technically a recall as well, as AMD needs to pull back the first batch of chips and replace them with known good units. That AMD has to essentially take a do-over on initial chip distribution is ultimately what’s driving this delay; it takes the better part of a month to properly seed retailers for a desktop CPU launch with even modest chip volumes, so AMD has to push the launch out to give their supply chain time to catch up.

For the moment, there are no further details on what the quality issue with the first batch of chips is, how many are affected, or what any kind of fix may entail. Whatever the issue is, AMD is simply taking back all stock and replacing it with what they’re calling “fresh units.”

AMD Ryzen 9000 Series Processors Zen 5 Microarchitecture (Granite Ridge)
AnandTech	Cores / Threads	Base Freq	Turbo Freq	L2 Cache	L3 Cache	Memory Support	TDP	Launch Date
Ryzen 9 9950X	16C/32T	4.3GHz	5.7GHz	16 MB	64 MB	DDR5-5600	170W	08/15
Ryzen 9 9900X	12C/24T	4.4GHz	5.6GHz	12 MB	64 MB		120W
Ryzen 7 9700X	8C/16T	3.8GHz	5.5GHz	8 MB	32 MB		65W	08/08
Ryzen 5 9600X	6C/12T	3.9GHz	5.4GHz	6 MB	32 MB		65W

Importantly, however, this announcement is only for the Ryzen 9000 desktop processors, and not the Ryzen AI 300 mobile processors (Strix Point), which are still slated to launch next week. A mobile chip recall would be a much bigger issue (they’re in finished devices that would need significant labor to rework), but also, both the new desktop and mobile Ryzen processors are being made on the same TSMC N4 process node, and have significant overlap due to their shared use of the Zen 5 architecture. To be sure, mobile and desktop are very different dies, but it does strongly imply that whatever the issue is, it’s not a design flaw or a fabrication flaw in the silicon itself.

That AMD is able to re-stage the launch of the desktop Ryzen 9000 chips so quickly – on the order of a few weeks – further points to an issue much farther down the line. If indeed the issue isn’t at the silicon level, then that leaves packaging and testing as the next most likely culprit. Whether that means AMD’s packaging partners had some kind of issue assembling the multi-die chips, or if AMD found some other i... CPUs

SK hynix to Enter 60 TB SSD Club Next Quarter

SK hynix this week reported its financial results for the second quarter, as well as offering a glimpse at its plans for the coming quarters. Notably among the company's plans for the year is the release of a SK hynix-branded 60 TB SSD, which will mark the firm's entry into the ultra-premium enterprise SSD league.

"SK hynix plans to expand sales of high-capacity eSSD and lead the market in the second half with 60TB products, expecting eSSD sales to be more than quadrupled compared to last year," a statement by SK hynix reads.

Currently there are only two standard form-factor 61.44 TB SSDs on the market: the Solidigm D5-P5336 (U.2/15mm and E1.L), and the Samsung BM1743 (U.2/15mm and E3.S). Both are built from a proprietary controller (Solidigm's controller still carries an Intel logotype) with a PCIe 4.0 x4 interface and use QLC NAND for storage.

SK hynix's brief mention of the drive means that tere aren't any formal specifications or capabilities to discuss just yet. But it is reasonable to assume that the company will use its own QLC memory for their ultra-high-capacity drives. What's more intriguing are which controller the company plans to use and how it is going to position its 60 TB-class SSD.

Internally, SK hynix has access to multiple controller teams, both of which have the expertise to develop an enterprise-grade controller suitable for a 60 TB drive. SK hynix technically owns Solidigm, the former Intel SSD and NAND unit, giving SK hynix the option of using Solidigm's controller, or even reselling a rebadged D5-P5336 outright. Alternatively, SK hynix has their own (original) internal SSD team, who is responsible for building their well-received Aries SSD controller, among other works.

Ultra-high-capacity SSDs for performance demanding read-intensive storage applications, such as AI inference on the edge or content delivery networks, is a promising premium market. So SK hynix is finding itself highly incentivized to enter it with a compelling offering. 

SSDs

SK hynix to Enter 60 TB SSD Club Next Quarter

SK hynix this week reported its financial results for the second quarter, as well as offering a glimpse at its plans for the coming quarters. Notably among the company's plans for the year is the release of a SK hynix-branded 60 TB SSD, which will mark the firm's entry into the ultra-premium enterprise SSD league.

"SK hynix plans to expand sales of high-capacity eSSD and lead the market in the second half with 60TB products, expecting eSSD sales to be more than quadrupled compared to last year," a statement by SK hynix reads.

Currently there are only two standard form-factor 61.44 TB SSDs on the market: the Solidigm D5-P5336 (U.2/15mm and E1.L), and the Samsung BM1743 (U.2/15mm and E3.S). Both are built from a proprietary controller (Solidigm's controller still carries an Intel logotype) with a PCIe 4.0 x4 interface and use QLC NAND for storage.

SK hynix's brief mention of the drive means that tere aren't any formal specifications or capabilities to discuss just yet. But it is reasonable to assume that the company will use its own QLC memory for their ultra-high-capacity drives. What's more intriguing are which controller the company plans to use and how it is going to position its 60 TB-class SSD.

Internally, SK hynix has access to multiple controller teams, both of which have the expertise to develop an enterprise-grade controller suitable for a 60 TB drive. SK hynix technically owns Solidigm, the former Intel SSD and NAND unit, giving SK hynix the option of using Solidigm's controller, or even reselling a rebadged D5-P5336 outright. Alternatively, SK hynix has their own (original) internal SSD team, who is responsible for building their well-received Aries SSD controller, among other works.

Ultra-high-capacity SSDs for performance demanding read-intensive storage applications, such as AI inference on the edge or content delivery networks, is a promising premium market. So SK hynix is finding itself highly incentivized to enter it with a compelling offering. 

SSDs

The Corsair RM750e ATX 3.1 Review: Simple And Effective

As mainstream power supplies continue to make their subtle shift to the ATX 3.1 standard, the pace of change is picking up. Already most vendors offer at least one ATX 3.1 unit in their lineups, and thanks to the relatively small set of changes that come with the revised standard, PSU vendors have largely been able to tweak their existing ATX 3.0 designs, allowing for them to quickly roll-out updated power supplies. This means that the inflection point for ATX 3.1 as a whole is quickly approaching, as more and more designs get their update and make their way out to retail shelves.

Today we're looking at our first ATX 3.1-compliant PSU from Corsair, one of the industry's most prolific (and highest profile) power supply vendors. Their revised RMe line of power supplies are aimed at the mainstream gaming market, which is perhaps not too surprising given how important ATX 3.1 support and safety are to video cards. The RM750e model we're looking at today is the smallest capacity for the lineup, which stretches from 750 Watts up to a hefty 1200 Watts.

Overall, the RM750e is built to meet the demands of contemporary gaming systems, and boasts a great balance between features, performance, and cost. It is an 80Plus Gold certified unit with modular cables and PCIe 5.1/ATX 3.1 certified, offering a single 600W 12V-2x6 connector. We will explore its specifications, construction, and performance to determine its standing in today’s market.

Cases/Cooling/PSUs