Category: Hardware

Details About 3rd Generation Ryzen Threadripper Appear: 24 and 32 Cores, Up to 280 W

AMD has already officially confirmed plans to launch its 3rd Generation Ryzen Threadripper processors this November, but naturally kept all the details to itself. At the time, it stated that the first products to market will be 24-core products. According to the company’s Product Master document that was published by AMD, it looks like some preliminary information about the CPUs is now in.

Based on the document (as well as previously released claims by the company), AMD’s Ryzen Threadripper 3000-series family will include 16-core, 24-core, and 32-core CPUs with a TDP up to 280 W for its most extreme parts. This is up from a 250 W TDP featured by AMD’s Ryzen Threadripper 2000 processors with up to 32 cores. The new process technology along with a higher thermal design power will likely enable the new CPUs to run faster and offer a significant performance increase when compared to predecessors (plus, keep in mind performance boost from the Zen 2 microarchitecture).

The new CPUs are also said to use a new variant of the SP3 socket called SP3r3, which questions if they will be drop-in compatible with previous-gen motherboards. Considering the fact that AMD is preparing new chipsets for its next-generation Ryzen Threadripper processors, but has been stated with providing socket compatibility, it will be interesting when AMD defines the compatibility table. This is likely a differentiator for PCIe 4.0 support.

AMD’s Product Master list is seemingly targeted at the company’s sales teams and contains OPN codes along with US ECCN, HTS, and CCATS codes that are required by the US export regulators. Although the document mentions a variety of products, at least some of them are potential or partner-specific devices not exactly located in the company’s roadmap for mass products. Therefore, the information from the list should be taken with a grain of salt, as some of these parts may not be hitting the market.

More Navi 14 GPUs Might Be Coming Soon

A new driver holds up to 14 different device IDs for Navi 14 graphics cards. Starting with the obvious, AMD has announced the Radeon RX 5500 and RX5500M, which are identified with the 7340:C7 and 7340:C1 device IDs, respectively. That leaves us with 12 unused device IDs. There’s no guarantee that they will all come out of the woodwork though.

Some of the other device IDs, such as the 7341:00 and 7340:CF have already appeared in public benchmark databases. The 7341:00 device seemingly has 8GB of GDDR6 memory while the 7340:CF device is rolling with 3GB of GDDR6 memory. AMD has already revealed that the Radeon RX 5500 will be available with 4GB and 8GB of GDDR6 memory. So, the 7340:CF device could be the 8GB variant.

Back in August, AMD partner Sapphire filed a bunch of Navi models with the Eurasian Economic Commission (EEC). In its submission, there were mentions of other Radeon RX 5000-series models, including the RX 5550 XT, RX 5550 and RX 5500 XT. The unclaimed device IDs could be for those models. Additionally, AMD produces custom graphics cards for its customers like Apple. Then there’s also the possibility that some of the device IDs are for prototypes that will probably never make it to the market.

Although AMD has officially announced the Radeon RX 5500, the chipmaker didn’t put a price tag or release date on the graphics card. We hope to see it land in the last quarter of the year.

Intel to Discontinue Nearly All Desktop Kaby Lake CPUs

Intel has announced End-of-Life plan for most of its desktop Kaby Lake and remaining Skylake processors. The boxed and tray versions of the chips will be available for interested parties for one more year and then will become history. The move will enable Intel to cut the number of product SKUs it offers to partners and reduce pressure on its factory network, which will help to increase supply of newer products made using various versions of Intel’s 14 nm process technology.

Introduced early in 2017, Intel’s desktop 7th Generation Core processors (Kaby Lake) have been around for nearly three years now. The CPUs certainly served their purpose, but it is time for them to go and Intel recommends its partners to place their final orders on these products by April 24, 2020. The final shipments will be made by October 9, 2020. Some of Intel’s Kaby Lake and Skylake products will be moved to Internet of Things (IoT) status and will be available for a little longer to IoT customers and probably some PC makers as there are still previous-generation motherboards on the market that need to be sold.

Intel’s desktop 6th Generation Core CPUs were launched in 2016 and most of them have been in EOL status for a while. This week, Intel said it would stop taking orders on the remaining desktop Skylake products on April 24, 2020, and will cease their shipments by October 9, 2020.

Winding down production of desktop Skylake and Kaby Lake processors in the next few months will free manufacturing capacities for newer Intel products and will enable the company to increase shipments of newer CPUs, such as 8th and 9th Generation Coffee Lake, that are also made using Intel’s 14 nm fabrication technology.

TRX40 Chipset For Upcoming AMD Ryzen Threadripper Listed

In what is being put down to an inadvertent leak by MSI, it could be that the currently unannounced chipset for AMD’s new 3rd Gen Ryzen Threadripper series processors has been officially mentioned (not withstanding the unconfirmed leaks from various sources). With the new AMD Ryzen Threadripper third generation set to hit the shelves in November, a lot of discussion is circulating about potential motherboard chipsets names and monikers. On one end we’re expecting compatibility with current X399 motherboards, and yet on the other new chipsets are expected to hit the market.

On MSI’s promotional pages, in which users can redeem rewards for purchasing specific and eligible MSI products, an unknown motherboard was revealed which piqued interest among users. Under MSI’s $25 Steam Gift Card promotion, an unreleased model named Creator TRX40 was listed which at present, hasn’t been announced by MSI or AMD. We know that the Creator series is MSI’s new content creator focused range with our announcement that MSI is refreshing some X299 models, notably the MSI Creator X299.

Little is known about what AMD intends to call the chipset for the new HEDT 3rd generation Threadripper processors, but it has been speculated that it might be called TRX80. With the accidental leak on the MSI website which has since been removed with much haste, it adds further speculation to the fire that AMD might release two chipsets instead of one. We have no confirmation as to what each of these chipset names mean, and some users have theorised that it could relate to consumer/commercial platforms or the memory channel count.

We know that AMD’s new 3rd Gen Ryzen Threadripper series will start with an entry-level model with 24 cores, with a 64-core part speculated to be in the pipeline. AMD has kept its cards closely to its chest as we’ve seen from previous product launches, but it remains an exciting time for users looking to build a single-socket system that features high core and thread count based on Ryzen’s Zen2 7nm architecture. We expect more information to become available closer to the next-gen Ryzen Threadripper series is set to launch in November.

MSI Creator TRX40 Motherboard For AMD’s 3rd Gen Ryzen Threadripper CPUs Leaks Out

AMD will be launching their 3rd Gen Ryzen Threadripper processors premiering with 24 cores next month. While there have been various rumors about the platform and socket support, it looks like MSI has leaked out one of their own motherboards, confirming a few theories for AMD’s 3rd HEDT lineup.

MSI Readies Creator TRX40 Motherboard For 3rd Gen ‘HEDT’ AMD Ryzen Threadripper CPUs

Spotted by Videocardz, the MSI Creator TRX40 was listed on a promo page by MSI which lets users redeem a $25 Steam gift card if they purchase an eligible product. One of those products as you might have guessed is the new Threadripper series motherboard and the naming convention more or less confirms that earlier reports were true. Previously, the MSI TRX40 PRO 10G motherboard was also registered at EEC which confirms that there will be not one but several models displayed by MSI for the new Threadripper lineup.

Motherboard makers are currently busy preparing a range of HEDT products for both Intel & AMD processors. We have already seen the Creator X299 which was announced yesterday and puts a lot of focus towards the creator market. Features such as more I/O, more capacity and higher power stability through the use of best-in-class VRMs (90A Power Stages) are just a few highlights of MSI’s Creator lineup.

There are no details mentioned for the Creator TRX40 but considering that it’s now listed by a manufacturer and the new 3rd Gen AMD Ryzen Threadripper CPUs launch next month, an official introduction of these boards may not be that far away. Other TRX40 motherboards from Gigabyte and ASUS have also leaked out prior to this which include:

ASUS PRIME TRX40-PRO

ASUS ROG STRIX TRX40-E GAMING

TRX40 AORUS Xtreme Waterforce

TRX40 AORUS Xtreme

TRX40 AORUS Master

TRX40 AORUS Pro WIFI

TRX40 DESIGNARE

AMD Ryzen Threadripper 3000 Series CPUs – Here’s What To Expect In Terms of Price, Specs, and Performance

The AMD Ryzen Threadripper 3000 series family is internally known as “Castle Peak” and is stated to bring dominant leadership in the HEDT market. The family will prove to be a new watermark in performance and overall efficiency while new platform features will be introduced on the new and enhanced motherboards to take them to the next level.

Currently, reports state that there will be two separate platforms for enthusiasts and workstation chips, the TRX40 and WRX80. The details of these platforms are listed below.

AMD TRX40 ‘Enthusiast’ Ryzen Threadripper 3000 Processors

The TRX4 HEDT platform would feature quad-channel memory, UDIMM memory support with 2 DIMMs per channel and up to 256 GB of capacity per channel. This means that the platform will support up to 1 TB of memory. There would also be support for 64 Gen 4 PCIe lanes with 16 lanes switchable with the SATA interface. There also seems to be info regarding TDPs and we can also notice the segmentation here too. The Group ‘A’ series processors, which are the HEDT lineup for the TRX40 platform with 280W TDP, Tcase Max temperature of 60C and Tctl Max of 100C.

AMD WRX80 ‘Workstation’ Ryzen Threadripper 3000 Processors

Now coming to the WRX80 series, we are looking at a pure workstation lineup. Even the processors built around this platform are spec’d similar to the EPYC 7002 variants, featuring 8-channel DDR4-3200 support in UDIMM, RDIMM, LRDIMM flavors. The platform would support 1 DIMM/channel featuring support for up to 2 TB of memory. There wouldn’t be any OC support like the TRX40 series but you get 96-128 Gen4 PCIe lanes with 32 switchable lanes to SATA. The Group ‘B’ series processors which are the workstation lineup for the WRX80 platform will also feature a 280W TDP but different temperature range of Tcase Max temperature of 81C and Tctl Max of 100C.

Considering that AMD would want to remain in a dominant position with the Threadripper 3000 series, we will be looking at some spectacular amounts of multi-threaded performance numbers which will only get better with the added clock speeds thanks to the 7nm process node. The CPUs will also be getting major core bumps, but AMD would like to keep prices close to current levels.

AMD CEO Dr. Lisa Su:

“You know. it’s very interesting, some of the things that circulate on the Internet—I don’t think we ever said that Threadripper was not going to continue—it somehow took on a life of its own on the Internet,” Su said, speaking to a small group of reporters following her keynote. “You will see more [Threadripper] from us; you will definitely see more.

If mainstream is moving up, then Threadripper will have to move up, up—and that’s what we’re working on.”

If we look at the trend with AMD’s jump from Ryzen Threadripper 1000 to Ryzen Threadripper 2000, we saw that the new processors with core parity of the previous generation were priced around the same with a $200-$300 shaved off from their previous price tag. The 1950X became 2950X and cost $200 US less. The higher core count parts were at a different market tier entirely, costing north of $1200 US but at the same time, much cheaper than their Core-X competitors.

In terms of raw performance output, the new die layout remains to be tested, but since it is more refined over the previous two generations with a stronger interconnect between them, the cache and latency performance may end up giving a bigger boost to total system responsiveness. AMD will definitely be aiming for both LGA 2066 and LGA 3647 lines with their new chips. Intel has said that their upcoming Core-X series will offer a much better value proposition with 2x better perf per dollar compared to Skylake-X, but that remains to be seen in real-world benchmarks and reviews.

Lexar NM600 M.2 NVMe SSD Review: Affordable and Efficient

Lexar, which was bought by Chinese SSD maker Longsys back in 2017, has been busy churning out new internal and external SSDs as of late. Here we’re looking at the company’s entry-level NVMe drive, the NM600 and seeing how it stacks up against the competition. The NM600 is poised to displace its SATA competition with aggressive pricing and speedy performance figures. But, does it have what it takes to keep up with M.2 alternatives?

At the core of Lexar’s NM600 is Silicon Motion’s SM2263XT, a 4-channel PCIe Gen3x4 NVMe controller. It’s similar to the controller powering the Intel SSD 660p and Crucial P1, but with one catch. The SM2263 utilizes a DRAM cache for the FTL mapping table, while the XT model doesn’t. Rather than incorporate DRAM in its architecture, Silicon Motion took advantage of a rather new NVMe feature called Host Memory Buffer (HMB).

VERDICT

Lexar’s NM600 isn’t a record-breaker, but it delivers speedy performance and is one of the most efficient SSDs we’ve tested. If you’re looking for a new budget M.2 NVMe SSD, it’s well worth considering.

DRAMless SSD controllers have typically offered low performance, especially when taxed with moderate workloads. That’s why HMB came into existence, it tries to speed things up and enable NVMe SSDs to match SATA-level pricing. Instead of using DRAM on the device, HMB technology enables the SM2263XT to utilize a small portion of the system’s RAM as a cache space for the controller’s mapping table. Without the need for DRAM in its design, the NM600 is able to hit the market at low prices and still deliver acceptable performance.

Specifications

Lexar’s NM600 is available in just two capacities, 240GB and 480GB. Based off of these capacity points, we can see that Lexar opted to give them a little bit more over-provisioning to help with performance consistency.

The 240GB model offers rated performance of up to 2.0/1.2GBps read/write, while the 480GB model can hit up to 2.1/1.6GBps read/write under sequential workloads. As this is a DRAMless SSD at lower capacities, their random IOPS performance isn’t the best. The 240GB model is rated for up to 100,000/151,000 IOPS read/write, and the 480GB model is rated for up to 188,000/156,000 IOPS.

These SSDs have a total bytes written (TBW) rating of 120TB at the 240GB capacity and 240TB at the 480GB capacity. And the NM600 is covered by a 3-year warranty, similar to other entry-level SSDs.

HMB isn’t the NM600’s only trick. The drive also features pseudo-SLC caching, which utilizes some of the NAND in an SLC mode for caching bursts of inbound writes. On the flip side, sustained performance isn’t as good as the rated figures. It also supports LDPC (Low-Density Parity Check) error correction code to ensure data integrity and reliability over time. You can also securely erase the device, so if you decide to sell it later on all data will be unrecoverable (or at least not very easily recoverable) by a third party.

A Closer Look

Lexar’s NM600 comes in an M.2 2280 form factor, with a black PCB which helps with aesthetics. But the crooked white sticker on top kind of ruins the look at the same time.

In conjunction with the DRAMless Silicon Motion SM2263XT controller, the NM600 features Micron’s 64L 3D TLC NAND flash. With four packages on the top size of the PCB, the NM600 is also a single-sided form factor design. No other components are on the bottom side, which makes it compatible with the few mobile devices that require the thinner size.

AMD Announces Ryzen 9 3900 and Ryzen 5 3500X

We recently took the then-unannounced Ryzen 9 3900 on a record-breaking spin, but now the chip is officially launched…kind of. AMD announced the Ryzen 9 3900 and Ryzen 5 3500X with little fanfare today, but for OEM and system integrator (SI) customers only. The 3900 is available globally, while the 3500X is only available in China. 

AMD designed the Ryzen 9 3900 to offer most of the 12-core, 24-threaded horsepower of the impressive Ryzen 9 3900X, but within a much lower 65W TDP range than its bigger brother, which sucks down 105W.

The 3900’s reduced power consumption manifests as a lower 3.1 GHz base clock than the 3.8 GHz on the 3900X. AMD also pared back the boost clock to 4.3 GHz. Aside from those alterations, the Ryzen 9 3900 is identical to the 3900X, and both support overclocking.

That means the Ryzen 9 3900 could theoretically offer similar performance to the chronically-understocked Ryzen 9 3900X. AMD also recently announced the PRO variant of the 3900, which sports the same specifications.

With AMD already selling every 3900X it can punch out at premium pricing, there’s little impetus to bring the Ryzen 9 3900 to market at a lower price point. It’s really a shame this processor isn’t coming to the retail market as it would present a great value option for budget builders, particularly for small form factor builds. It was plenty impressive in our testing, but as it stands now, this chip will only come in pre-built systems.

Meanwhile, the Ryzen 5 3500X slots in as a six-core model but comes without simultaneous multi-threading (SMT), meaning it only has six threads of performance on offer. AMD tends to avoid de-featuring its retail processors, so the lack of SMT stands out among the Ryzen 3000 product stack.

Surprisingly, the 3500X will only be made available to OEM/SI customers in China. This chip is meant to tackle Intel’s Core i5-9400F in the OEM market, but there are also rumblings of a Ryzen 5 3500 (non-X model) coming to market soon. It might also be destined for the OEM/SI markets, but only time will tell.

Both processors come bearing all the expected goods of the Ryzen 3000 series, like the 7nm process, Zen 2 microarchitecture, and support for PCIe 4.0.

Here’s AMD’s announcement on the matter:

“The Ryzen 9 3900, now available globally, and Ryzen 5 3500X, available in China only, enjoy the features of current Ryzen 3000 series processors, letting OEM and SI partners take full advantage of AMD’s most advanced CPU platform. Both Ryzen processors offer powerful gaming and high-speed productivity performance, support industry leading PCIe 4.0, AMD’s Precision Boost Overdrive and Ryzen Master Utility, and, like all Ryzen 3000 processors, are fully unlocked for performance tuning.”

AMD isn’t sharing pricing for these chips. As OEM models, they are largely bought in volume purchases only. As such, we don’t expect the company to share pricing information any time soon. Both chips are available to OEMs now.

Samsung Develops 12-Layer 3D TSV DRAM: Up to 24 GB HBM2

Samsung on Monday said that it had developed the industry’s first 12-layer 3D packaging for DRAM products. The technology uses through silicon vias (TSVs) to create high-capacity HBM memory devices for applications that benefit from high memory bandwidth and capacities, such as higher-end graphics, FPGAs, and compute cards.

Samsung’s 12-layer DRAM KGSDs (known good stack die) will feature 60,000 TSV holes which is why the manufacturer considers its technology one of the most challenging packaging for mass production. Despite increase of the number of layers from eight to 12, thickness of the package will remain at 720 microns, so Samsung’s partners will not have to change anything on their side to use the new technology. It does mean that we’re seeing DRAM layers getting thinner, with acceptable yields for high-end products.

One of the first products to use Samsung’s 12-layer DRAM packaging technology will be the company’s 24 GB HBM2 KGSDs that will be mass produced shortly. These devices will allow developers of CPUs, GPUs, and FPGAs to install 48 GB or 96 GB of memory in case of 2048 or 4096-bit buses, respectively. It also allows for 12 GB and 6 GB stacks with less dense configurations.

Samsung did not disclose how much will 12-layer 24 GB HBM2 devices cost, but since they will be available exclusively from Samsung, we expect the manufacturer to charge a premium.

Here is what Hong-Joo Baek, executive vice president of TSP (Test & System Package) at Samsung Electronics, had to say:

“Packaging technology that secures all of the intricacies of ultra-performance memory is becoming tremendously important, with the wide variety of new-age applications, such as artificial intelligence (AI) and High Power Computing (HPC). As Moore’s law scaling reaches its limit, the role of 3D-TSV technology is expected to become even more critical. We want to be at the forefront of this state-of-the-art chip packaging technology.”