AMD

AMD Launches Radeon RX Vega 64 and Vega 56, Taking On GeForce GTX 1080 and 1070 (hothardware.com) 84

MojoKid writes: AMD has finally launched its Radeon RX Vega series of graphics cards today, based on the company's next generation Vega 10 GPU architecture. There are three base card specs announced, though there are four cards total, with a Limited Edition air-cooled card as well. Three of the cards have 64 NGCs (Next Generation Compute Units) with 4096 stream processors, while Radeon RX Vega 56 is comprised of 56 NCGs with 3584 SPs. Base clocks range from roughly 1150 to 1400MHz, with boost clocks from 1470MHz to 1670MHz or so. All cards come with 8GB of HBM2 and sport 484GB/sec of memory bandwidth, except for Vega 56, which has a bit less, at 410GB/s. They are power-hungry as well, ranging from the 345 Watt liquid-cooled Radeon RX Vega 64, to the 295 Watt air-cooled RX Vega 64 and 210 Watt Radeon RX Vega 56. Performance-wise, Radeon RX Vega 64 is neck-and-neck with NVIDIA's GeForce GTX 1080, winning some and losing some, with flashes of strength in DirectX 12-based games and benchmarks. Vega 64 also maintains generally better minimum frame rates versus GTX 1080. Radeon RX Vega 56 is a more credible midrange threat that handily out-performs a GeForce GTX 1070 across the board. In DX12 gaming, Radeon RX Vega 56 stretches its lead over the similarly-priced GTX 1070. Both cards, however, are more power-hungry, louder and run hotter than NVIDIA's high-end GeForce GTX 1080. Radeon RX Vega 64 cards will retail for $499 (Liquid Cooled cards at $699), while Radeon RX Vega 56 drops in at $399. All cards should be available at retail starting today.
AMD

AMD Ryzen Threadripper Launched: Performance Benchmarks Vs Intel Skylake-X (hothardware.com) 122

Reader MojoKid writes: AMD continues its attack on the desktop CPU market versus Intel today, with the official launch of the company's Ryzen Threadripper processors. Threadripper is AMD's high-end, many-core desktop processor, that leverages the same Zen microarchitecture that debuted with Ryzen 7. The top-end Ryzen Threadripper 1950X is a multi-chip module featuring 16 processor cores (two discrete die), with support for 32 threads. The base frequency for the 1950X is 3.4GHz, with all-core boost clocks of up to 3.7GHz. Four of the cores will regularly boost up to 4GHz, however, and power and temperature permitting, those four cores will reach 4.2GHz when XFR kicks in. The 12-core Threadripper 1920X has very similar clocks and its boost and XFR frequencies are exactly the same. The Threadripper 1920X's base-clock, however, is 100MHz higher than its big brother, at 3.5GHz. In a litany of benchmarks with multi-threaded workloads, Threadripper 1950X and 1920X high core-counts, in addition to strong SMT scaling, result in the best multi-threaded scores seen from any single CPU to date. Threadripper also offers massive amounts of memory bandwidth and more IO than other Intel processors. Though absolute power consumption is somewhat high, Threadrippers are significantly more efficient than AMD's previous-generation processors. In lightly-threaded workloads, Threadripper trails Intel's latest Skylake-X CPUs, however, which translates to lower performance in applications and games that can't leverage all of Threadripper's additional compute resources. Threadripper 1950X and 1920X processors are available starting today at $999 and $799, respectively. On a per-core basis, they're less expensive than Intel Skylake-X and very competitively priced.
Intel

Intel's 8th-Gen 'Coffee Lake' Core CPUs Will Be Revealed During the Great American Eclipse (pcworld.com) 98

Brad Chacos, writing for PCWorld: Intel's response to AMD's disruptive Ryzen processors is soon to get its time in the sun. Well, sort of. On Tuesday, Intel announced plans to livestream the launch of its 8th-generation "Coffee Lake" processors on August 21 -- the same day that the great American solar eclipse casts its shadow across the United States. Intel's throwing shade. Eighth-gen Coffee Lake chips will be built using a revised version of Intel's 14nm process technology for an unprecedented fourth time, following in the footsteps of Broadwell, Skylake, and Kaby Lake architectures. You'll probably also need a new motherboard to use them. But most notably, Intel claims 8th-gen Core chips will be up to 30 percent faster than today's Kaby Lake processors in some applications. Intel chips haven't seen a performance leap like that in years. Beyond that, little is officially known about Coffee Lake, though the churning internet rumor mill thinks that Intel will up the core counts this time around to combat the threat of Ryzen.
AMD

Preview of AMD Ryzen Threadripper Shows Chip Handily Out-Pacing Intel Core i9 (hothardware.com) 180

MojoKid writes: AMD is still days away from the formal launch of their Ryzen Threadripper family of 12 and 16-core processors but OEM system builder Dell and its Alienware gaming PC division had an inside track on first silicon in the channel. The Alienware Area-51 Threadripper Edition sports a 16-core Ryzen Threadripper 1950X processor that boosts to 4GHz with a base clock of 3.4GHz and an all-core boost at 3.6GHz. From a price standpoint, the 16-core Threadripper chip goes head-to-head with Intel's 10-core Core i9-7900X at a $999 MSRP. In early benchmark runs of the Alienware system, AMD's Ryzen Threadripper is showing as much as a 37% percent performance advantage over the Intel Core i9 Skylake-X chip, in highly threaded general compute workload benchmarks like Cinebench and Blender. In gaming, Threadripper is showing roughly performance parity with the Core i9 chip in some tests, but trailing by as much as 20% in lower resolution 1080p gaming, as is characteristic for many Ryzen CPUs currently, in certain games. Regardless, when you consider the general performance upside with Ryzen Threadripper versus Intel's current fastest desktop chip, along with its more aggressive per-core pricing (12-core Threadripper at $799), AMD's new flagship enthusiast/performance workstation desktop chips are lining up pretty well versus Intel's.
AMD

AMD Confirms Linux 'Performance Marginality Problem' On Ryzen (phoronix.com) 120

An anonymous reader writes: Ryzen customers experiencing segmentation faults under Linux when firing off many compilation processes have now had their problem officially acknowledged by AMD. The company describes it as a "performance marginality problem" affecting some Ryzen customers and only on Linux. AMD confirmed Threadripper and Epyc processors are unaffected; they will be dealing with the issue on a customer-by-customer basis, and their future consumer products will see better Linux testing/validation. Ryzen customers believed to be affected by the problem can contact AMD Customer Care. Michael Larabel writes via Phoronix: "With the Ryzen segmentation faults on Linux they are found to occur with many, parallel compilation workloads in particular -- certainly not the workloads most Linux users will be firing off on a frequent basis unless intentionally running scripts like ryzen-test/kill-ryzen. As I've previously written, my Ryzen Linux boxes have been working out great except in cases of intentional torture testing with these heavy parallel compilation tasks. [AMD's] analysis has also found that these Ryzen segmentation faults aren't isolated to a particular motherboard vendor or the like, contrary to rumors/noise online due to the complexity of the problem."
Intel

Intel Releases Final Core i9 Specs and Release Dates -- And Threadripper Is Faster (Sometimes) (pcworld.com) 91

On Monday, Intel took the wraps of final details of its Core i9 microprocessors. From a report: Remember that Intel's Core X-series family (also called the Core i9) was announced with several key omissions: namely the clock speeds of the 12-core Core i9-7920X and above, as well as the thermal design power, or TDP. On Monday, Intel filled those in. The 12-core Core i9-7920X launches Aug. 28 while the 14-, 16-, and 18-core Core i9 chips ship on Sept. 25. Perhaps most important, though, is that we now know how fast Intel's Core i9s will run. When Intel inadvertently revealed that its 12-core Core i9-7920X was 2.9-GHz -- slower than the comparable AMD Threadripper -- a subset of the internet had a small freakout. We now know that that will be true for the remaining Core i9s as well, but with a big caveat. Here are the remaining speeds and feeds for the high-end Core i9 chips:
Core i9-7980XE (18 cores, 36 threads): 2.6GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7960X (16 cores, 32 threads): 2.8GHz; Boost, 4.2GHz to 4.4 GHz.
Core i9-7940X 14 cores, 28 threads: 3.1GHz; Boost: 4.3GHz to 4.4GHz.
Core i9-7920X (12 cores, 24 threads): 2.9-GHz; Boost: 4.3-GHz to 4.4GHz.

Note that the boost speeds refer to both Intel's Turbo Boost Technology 2.0 and 3.0. [...] Essentially, both Intel and AMD can claim the title of fastest processor. Threadripper's base clock speeds are faster, but Intel's boost speeds climb higher than Threadripper can. It's also important to note that while Threadripper consumes 180 watts, even the fastest Core i9 chips Intel has announced have a lower TDP of 165 watts.

Intel

Intel's Upcoming Coffee Lake CPUs Won't Work With Today's Motherboards (pcworld.com) 240

Intel's upcoming Coffee Lake CPUs won't work with existing 200-series motherboards that support Kaby Lake, a manufacturer confirmed on Wednesday. In a Twitter post by Asrock last Saturday, the company confirmed the news when asked if "the Z270 Supercarrier [will] get support for the upcoming @intel Coffee Lake CPUs." Their response: "No, Coffee Lake CPU is not compatible with 200-series motherboards." PCWorld reports: According to at least one reliable source outside of Intel, the new Coffee Lake CPU will indeed not be compatible with Z270 boards, even though the chipsets with the upcoming Z370 appear to be the same, PCWorld was told. The source added that there are hopes in the industry that Intel will change its mind on compatibility. Tomshardware.com said it had independently confirmed the news with Asrock officials as well.

Why this matters: The vast majority of new CPU sales are in new systems, and they likely won't be impacted by the incompatibility. However, there's also a very large and very vocal crowd of builders and upgraders who still swap out older, slower CPUs for newer, faster CPUs to maximize their investment. An upgrade-in-place doesn't sell an Intel chipset, but it at least keeps them on the Intel platform. If consumers are forced to dump an existing Z270 motherboard for a newer Z370 to get a six-core Coffee Lake CPU, Intel risks driving them into the arms of AMD and its Ryzen CPUs.

Businesses

NVIDIA Announces Quadro And TITAN xP External GPU Solutions, OptiX 5.0 SDK (hothardware.com) 36

Brandon Hill, writing for HotHardware: AMD isn't the only hardware company making waves this week at SIGGRAPH 2017. NVIDIA is looking to bolster its position in the professional graphics arena with a few new breakthroughs. The first of which is the addition of two new external graphics solutions that are targeted at professional artists and designers who primarily work with notebooks. NVIDIA is making it possible for these professionals to use either Pascal-based TITAN xP or Quadro graphics cards within an external GPU (eGPU) enclosure. NVIDIA will be partnering with a number of hardware partners including Bizon, Magma, and Sonnet, who will make compatible solutions available in September. NVIDIA is also playing up two of its strengths in artifice intelligence (AI) by launching the OptiX 5.0 SDK. With version 5.0, the OptiX is gaining ray tracing support to help speed up processing with regards to visual designs. This new release also adds GPU-accelerated motion blur along with AI-enhanced denoising capabilities.
AMD

AMD Unveils Radeon RX Vega Series Consumer Graphics Cards Starting At $399 (hothardware.com) 91

MojoKid writes: AMD has officially lifted the veil on its new Radeon RX consumer graphics line-up, featuring the company's next-generation Vega GPU architecture. Initially, there are four cards in the Radeon RX Vega line-up, the standard air-cooled Radeon RX Vega 64, a Radeon RX Vega 64 Limited Edition with stylized metal fan shroud, the liquid-cooled Radeon RX Vega 64 Liquid, and the lower-cost Radeon RX Vega 56. At the heart of all Radeon RX Vega series cards is the Vega 10 GPU which is comprised of roughly 12.5 billion transistors and is manufactured using a 14nm FinFET LPP process. Vega 10 can reliably reach the 1.7GHz range, whereas AMD's previous gen Fiji hovered around 1GHz. The base GPU clock speed of the air-cooled Vega 64 is 1,247MHz with a boost clock of 1,546MHz. There is 8GB of HBM2 memory on-board that offers up peak bandwidth of 484GB/s. All told, the Radeon RX Vega 64 is capable of 25.3 TFLOPs (half-precision) of compute performance. The Radeon RX Vega 64 Liquid-Cooled Edition has the same GPU configuration, but with higher base and boost clocks -- 1,406MHz and 1,677MHz, respectively. The lower cost Radeon RX Vega 56 features the same Vega 10 GPU, but 8 of its CUs have been disabled and its clocks are somewhat lower. Although AMD touts a number of efficiency improvements, the Vega RX series requires some serious power. Vega 56 board power is in the 210 Watt range, while the top-end liquid-cooled card hits 345 Watts. AMD claims top-end Vega cards will be competitive with NVIDIA's GeForce GTX 1080 series of cards. AMD Radeon RX Vega graphics cards are expected to ship on August 14th.
AMD

AMD Launches Ryzen 3 Series Low Cost Processors Starting At $109 (hothardware.com) 80

Reader MojoKid writes: AMD is launching a new series of Ryzen processors today, the affordably priced Ryzen 3. Ryzen 3 will complement the previously launched Ryzen 7 and Ryzen 5 series of desktop processors, but will target entry-level price points. Ryzen 3 features the same die as its higher-end Ryzen 7 and 5 siblings, but has fewer active cores and symmetrical multi-threading (SMT) has been disabled. Ryzen 3 processors feature quad-core configurations, leverage the same socket and chip packaging and are also fully unlocked for easy overclocking. The Ryzen 3 1300X has a base clock of 3.5GHz, with a 3.6GHz all-core boost clock, a 3.7GHz two-core boost, and a max XFR boost clock of 3.9GHz. The Ryzen 3 1200's default clocks are decidedly lower. Its base and all-boost clocks are both only 3.1GHz, and its two-core boost tops out at 3.4GHz. XFR pushes its max single-core clock up to 3.45GHz. In the benchmarks, with multi-threaded workloads, the Ryzen 3's quad-core configuration generally gives it an edge over the dual-core / quad-thread Intel Core i3 and in some cases allows it to compete with more expensive Intel Core i5 chips. With single or lightly threaded workloads, however, Kaby Lake-based Core i3s are likely to pull ahead due to their increased IPC and typically higher clocks. Ryzen 3 1300X will retail for $129, while Ryzen 3 1200 will list for $109. Retail chips should be available today in the channel.
AMD

AMD Has No Plans To Release PSP Code (twitch.tv) 125

AMD has faced calls from Edward Snowden, Libreboot and the Reddit community to release the source code to the AMD Secure Processor (PSP), a network-capable co-processor which some believe has the capacity to act as a backdoor. But despite some signs earlier that it might consider opening the PSP code at some point, the chip-maker has now confirmed that there hasn't been a change of heart yet. "We have no plans on releasing it to the public," the company executives said in a tech talk (video).
AMD

Chipmakers Nvidia, AMD Ride Cryptocurrency Wave -- For Now (bloomberg.com) 57

During California's Gold Rush, it was often the sellers of pickaxes and shovels who made the most money. In the frenzy to get rich quick from cryptocurrencies, some investors are calling computer chipmakers the modern-day equivalent. From a report: Shares of Nvidia and Advanced Micro Devices have gained at least 14 percent since the beginning of June, spurred in part by about a 10-fold boom from April to June in a market, known as ethereum, for a currency that can be used to buy computing power over the internet. What's the link between ethereum and these Silicon Valley chipmakers? It lies in the really powerful graphics processors, designed to make computer games more realistic, that are also needed to gain access to encrypted digital currencies. Nvidia and AMD have rallied in the last month and a half even as investors have ignored chip stocks leaving the benchmark Philadelphia Stock Exchange Semiconductor Index up about 1 percent. Nvidia has gained 14 percent and AMD rallied 27 percent. While some of that has come from optimism around new products for other markets, analysts are projecting that sales related to cryptocurrencies will result in a spike in revenue for both companies. Even so, investors shouldn't bank on a lasting impact from the cryptocurrency boom, said Stacy Rasgon, an analyst at Sanford C. Bernstein & Co. "This has happened before," Rasgon said. "It lasted about a quarter." [...] Like bitcoin, ethereum is an attempt by an online community to create an economy that doesn't rely on government-backed currencies. Unlike bitcoin, it's focused solely on offering decentralized computing and storage services. Those seeking to use these services -- and speculators looking for a quick profit by creating and then selling ether -- have seized on graphics cards, which excel at performing multiple simple calculations in parallel, as a faster way to claim the blocks of code that act as the currency of the ethereum market. Demand from ethereum miners has created temporary shortages of some of the graphics cards, according to analysts, who cite sold-out products at online retailers. Estimates of additional sales from this demand run as high as $875 million, according to RBC Capital Markets analyst Mitch Steves. That would roughly equal AMD's total sales from graphics chips last year, or half of Nvidia's quarterly sales of those components. But Steves and other analysts are also quick to warn that the market opportunity could fizzle out.
AMD

AMD Threadripper 1950X Trounces Core I9-7900X In Multithreading Benchmark (pcper.com) 114

dryriver writes: The Cinebench R15 benchmark is a popular tool for measuring how well CPUs cope with multithreaded compute loads. AMD's Threadripper 1950X 16 core CPU, priced at $999 according to AMD, benchmarks 41% faster in Cinebench R15 than Intel's also $999 10 core Core i9-7900X CPU. While Intel's Core i9-7900X scores 2186 points on Cinebench, AMD's Threadripper 1950X scores 3046 points. Even the cheaper 12 core $799 Threadripper 1920X is over 200 points faster in Cinebench R15 than Intel's Core i9-7900X. Intel has its own 16 core Core i9-7960X in the works, performance yet unknown, priced at $1,699, but AMD's 16 core part currently appears to be a full $700 cheaper than Intel's MSRP. It remaines to be seen who is faster in single-threaded performance -- Intel may take that crown --and what the power consumption of a fully loaded Threadripper looks like compared to its Core i9 counterpart.
AMD

AMD Unveils Ryzen Threadripper 1950X 16-Core and 1920X 12-Core Specs and Pricing (hothardware.com) 85

MojoKid writes: AMD first teased its Ryzen Threadripper series of high-end desktop (HEDT) processors back in mid-May, but is now sharing additional details on the first two products in the family. Both processors are based on the 14nm Zen core, make use of AMD's new Socket TR4 interface, support quad-channel DDR memory, and feature a total of 64 PCIe lanes. In addition, both processors will come from the factory unlocked. Ryzen Threadripper 1920X will have 12 Cores, 24 Threads, and 3.5/4.0 GHz (Base Clock/Precision Boost) clock speeds. Ryzen Threadripper 1950X will have 16 Cores, 32 Threads, and 3.4/4.0 GHz (Base Clock/Precision Boost) clock speeds. Pricing is set at $799 and $999, respectively, with availability in early August, though Dell's Alienware gaming PC division will have systems shipping with the new chip starting this month. AMD also put the new chips up against Intel's Core i7-7900X 10-core CPU in a Cinebench benchmark run in a video demo, and the 12-core Threadripper chip beats Intel's currently available Skylake-X chip handily, while the 16-core Threadripper outpaces it even further.
AMD

Benchmarking Utility Shows AMD Ryzen Rapidly Stealing Market Share From Intel (hothardware.com) 119

According to PassMark, which publishes a benchmarking utility called PerformanceTest, the launch of Ryzen chips has resulted in a surge in AMD's share of its CPUs being tested. From a report: In the first quarter of last year, just 20.1 percent of tests were performed on AMD hardware, versus 79.8 percent on Intel chips. The gap widen by the end of the year, with AMD accounting for 17.8 percent of all tests run through Passmark's software, with Intel jumping up to 82.2 percent. Fast forward to the quarter than just ended and things are looking a bit different. AMD's share has climbed to 26.2 percent, while Intel's has slipped to 73.7 percent. Obviously Intel is still dominating, but what this shows us is that AMD was able to take a nearly 10 percent chunk out what is probably the enthusiast market from Intel. The reason we believe this is largely relegated to the enthusiast market is because AMD's Ryzen architecture is brand new, and that would be the most logical explanation as to why its numbers have suddenly spiked at the expense of Intel.
Open Source

Linux Kernel 4.12 Officially Released (softpedia.com) 55

prisoninmate quotes Softpedia: After seven weeks of announcing release candidate versions, Linus Torvalds today informs the Linux community through a mailing list announcement about the general availability of the Linux 4.12 kernel series. Development on the Linux 4.12 kernel kicked off in mid-May with the first release candidate, and now, seven weeks later we can finally get our hands on the final release... A lot of great improvements, new hardware support, and new security features were added during all this time, which makes it one of the biggest releases, after Linux 4.9...

Prominent features of the Linux 4.12 kernel include initial support for AMD Radeon RX Vega graphics cards, intial Nvidia GeForce GTX 1000 "Pascal" accelerated support, implementation of Budget Fair Queueing (BFQ) and storage-I/O schedulers, more MD RAID enhancements, support for Raspberry Pi's Broadcom BCM2835 thermal driver, a lot of F2FS optimizations, as well as ioctl for the GETFSMAP space mapping ioctl for both XFS and EXT4 filesystems.

Linus said in announcing the release that "I think only 4.9 ends up having had more commits," also noting that 4.9 was a Long Term Support kernel, whereas "4.12 is just plain big."

"There's also nothing particularly odd going on in the tree - it's all just normal development, just more of it than usual."
AMD

AMD Launches Ryzen PRO CPUs: Enhanced Security, Longer Warranty, Better Quality (anandtech.com) 50

Reader harrisonweber shares a report: This morning AMD introduced their Ryzen PRO processors for business and commercial desktop PCs. The new lineup of CPUs includes the Ryzen 3 PRO, Ryzen 5 PRO and Ryzen 7 PRO families with four, six, or eight cores running at various frequencies. A superset to the standard Ryzen chips, the PRO chips have the same feature set as other Ryzen devices, but also offer enhanced security, 24 months availability, a longer warranty and promise to feature better chip quality. The AMD Ryzen PRO lineup of processors consists of six SKUs that belong to the Ryzen 7, Ryzen 5 and Ryzen 3 families targeting different market segments and offering different levels of performance. As one would expect, the Ryzen 7 PRO models are aimed at workstation applications and thus have all eight cores with simultaneous multithreading enabled, the Ryzen 5 PROmodels are designed for advanced mainstream desktops and therefore have four or six cores with SMT, whereas the Ryzen 3 PRO models are aimed at office workloads that work well on quad-core CPUs without SMT. The specifications of the Ryzen 7 PRO and the Ryzen 5 PRO resemble those of regular Ryzen processors. Meanwhile, the Ryzen 3 PRO are the first chips from the Ryzen 3 lineup and thus give us a general idea what to expect from such products: four cores without SMT operating at 3.1-3.5 GHz base frequency along with 2+8 MB of cache.
Graphics

NVIDIA To Launch Graphics Cards Specifically Designed For Digital Currency Mining (cnbc.com) 105

Digital currency mining is in high demand, causing GPU prices to skyrocket. Nvidia is planning to capitalize on this trend by releasing graphics cards specifically designed for cryptocurrency. From a product listing on ASUS' website: "ASUS Mining P106 is designed for coin mining with high-efficiency components -- delivering maximum hash-rate production at minimum cost. ASUS Mining P106 enhances the megahash rate by up to 36% compared cards in the same segment that are not tailored for mining. The new card is also engineered to be seriously durable, enabling 24/7 operation for uninterrupted coin production." The ASUS Mining P106 uses an Nvidia chip, according to the specifications page on the website. CNBC reports: Nvidia, AMD and ASUS have not officially announced the digital currency mining cards, according to their website press pages. It is not certain when the cards will be available for sale. Nvidia is likely making the cards designed for this use so that the surging digital currency demand doesn't affect its ability to serve the lucrative PC gaming market.
Intel

New HyperThreading Flaw Affects Intel 6th And 7th Generation Skylake and Kaby Lake-Based Processors (hothardware.com) 135

MojoKid writes: A new flaw has been discovered that impacts Intel 6th and 7th Generation Skylake and Kaby Lake-based processors that support HyperThreading. The issue affects all OS types and is detailed by Intel errata documentation and points out that under complex micro-architectural conditions, short loops of less than 64 instructions that use AH, BH, CH or DH registers, as well as their corresponding wider register (e.g. RAX, EAX or AX for AH), may cause unpredictable system behavior, including crashes and potential data loss. The OCaml toolchain community first began investigating processors with these malfunctions back in January and found reports stemming back to at least the first half of 2016.

The OCaml team was able pinpoint the issue to Skylake's HyperThreading implementation and notified Intel. While Intel reportedly did not respond directly, it has issued some microcode fixes since then. That's not the end of the story, however, as the microcode fixes need to be implemented into BIOS/UEFI updates as well and it is not clear at this time if all major vendors have included these changes in their latest revisions.

Intel

AMD Looks To 'Crush' Intel's Xeon With New Epyc Server Chips (extremetech.com) 136

AMD has unveiled the first generation of Epyc, its new range of server processors built around its Zen architecture. Processors will range from the Epyc 7251 -- an eight-core, 16-thread chip running at 2.1 to 2.9GHz in a 120W power envelope -- up to the Epyc 7601: a 32-core, 64-thread monster running at 2.2 to 3.2GHz, with a 180W design power. From a report: These chips are built on the same fundamental architecture as the company's Ryzen CPU cores, and they're aimed at the incredibly powerful data center market. AMD's 32-core / 64-thread Epyc CPUs combine four eight-core dies, each connected to the other via the company's Infinity Fabric. According to AMD, this approach is significantly cheaper than trying to pack 32 cores into a single monolithic die -- that approach would leave the company potentially throwing away huge amounts of silicon during its production ramp. The Infinity Fabric is deliberately over-provisioned to minimize any problems with non-NUMA aware software, according to Anandtech. Each 32-core Epyc CPU will support eight memory channels and two DIMMs per channel, for a total maximum memory capacity of 2TB per socket, or 4TB of RAM in a two-socket system. Each CPU will also offer 128 lanes of PCI Express 3.0 support -- enough to connect up to six GPUs at x16 each with room left over for I/O support. That's in a one-socket system, mind you. In a two-socket system, the total number of available PCI Express 3.0 lanes is unchanged, at 128 (64 PCIe 3.0 lanes are used to handle CPU -- CPU communication). Anandtech has a longer writeup with more details on the CPUs power efficiency and TDP scaling. Further reading: ZDNet, press release.

Slashdot Top Deals