Much has been said regarding AMD's Threadripper CPUs, particularly when it comes to how they are manufactured. At first, we thought Threadripper was actually EPYC in disguise, due it having what appeared to be four full-fledged 8-core modules - the same design as AMD's server-bound 32-core EPYC chips. The presence of gold-plating under all four dies seemed to confirm that these were in fact four full Threadripper dies, instead of two dies and two spacers (as AMD's statements led us to believe) for even IHS pressure on the four dies, instead of the uneven pressure that would result from the chip only having two physical dies present.

In a Forbes interview conducted by Anthony Leather, AMD officials Senior Vice President and General Manager Jim Anderson, Corporate Vice President of Worldwide Marketing John Taylor, Sarah Youngbauer of AMD's communications team, and James Prior, AMD's Senior Product Manager, have shed some light on the development process for AMD's equivalent of a flash hit - their HEDT, HCC Threadripper chips. Threadripper, which leverages AMD's Zen architecture used in their Ryzen and EPYC processors, makes use of an MCM design with up to 16 cores and 32 threads, with AMD's svelte Zen, 8-core base units linked through the company's high speed interconnect Infinity Fabric.

This has allowed the company to scale designs from four core processors with Ryzen 3, all the way towards the current cream of the crop Threadripper 1950X. It's an extremely scalable design, which brings with it improved yields and some pretty significant cost savings for AMD due to smaller dies. This, in turn, means the company is able to more agressively price their Ryzen and Threadripper processors compared to the competition, at least when it comes to high core and thread counts for the same price bracket - and the success of that business decision is showing.For our forum lurkers, this article is marked as an editorial.

At its pre-launch media conference call for the Ryzen Threadripper 1900X, AMD mentioned that the chip has been carved out from the common 4-die EPYC MCM using a "4-0-4-0 diagonal configuration," which led to some confusion as to which cores/dies AMD disabled to carve out the $549 8-core HEDT processor. The company shed some light on this matter, responding to questions from TechPowerUp.

It turns out, that the Threadripper 1900X features an entire CCX (quad-core CPU complex) disabled per active die on the multi-chip module, so the CCX that's enabled has 8 MB of L3 cache; and access to the die's entire uncore resources, such as the dual-channel memory controller, PCIe root complex, etc. With two such active "Zeppelin" dies, the Threadripper 1900X ends up with 8 cores, 16 MB of L3 cache, a quad-channel memory interface, and 64 PCIe lanes.

In a response to AMD's current uptake in the consumer, HEDT and server markets with its vertical slice of the Zen architecture, Intel has started rebranding their products and image, changing product names and placement in a bid to increase the "freshness" factor of its offerings. E5 and E7 Xeons are gone, with the introduction of a metallic naming scheme: Bronze, Silver, Gold and Platinum are now Intel's Xeon products, and Xeon-W takes the spot as Intel's workstation-oriented product stack. They do this by being - essentially - a conversion of Intel's Core i9 X299 family of processors towards the professional market with inclusion of professional-geared features. And as is usual with Intel, a new chipset - C422 - is needed in order for these to properly function.

The new Xeon-W product family will still make use of the LGA 2066 socket, bringing with them ECC and vPro support. The Xeon-W CPU family will feature 4 to 18 cores, support up to 512GB of ECC RDIMM/LRDIMM memory, support dual 512-bit FMAs, and peak clocks of 4 GHz base and 4.5 GHz Turbo. All the parts will support 48 PCIe 3.0 lanes from the processor,and CPUs in the Xeon-W stack are rated at 140W TDP: with exception of the quad cores, which come in at at 120W. Xeon-W processors only support Turbo Boost 2.0, instead of their Core i9 counterparts' Turbo 3.0.

AMD, presenting at a HotChips talk, packed with VLSI gurus, defended its decision to make AMD EPYC enterprise processors in the socket SP3r2 package a multi-chip module (MCM) of four 8-core "Summit Ridge" dies, rather than building a monolithic 32-core die. For starters, it stressed on the benefits of a single kind of silicon, which it can use across its performance-desktop, high-end desktop, and enterprise product-stacks; which translates into higher yields. "Summit Ridge" is the only CPU silicon based on the "Zen" micro-architecture, with the company giving finishing touches to the second silicon, codenamed "Raven Ridge," which will power mobile and desktop APUs. AMD has to pick the best-performing dies out of a common bin. The top 5% dies go into powering the company's Ryzen Threadripper HEDT processors, and a higher percentile go into making EPYC.

The relatively smaller 8-core common die has an inherently higher yield than a larger chip due to the rule of inverse-exponential reduction in yield with increasing die-size. This, coupled with the R&D costs that would have gone into developing the hypothetical monolithic 32-core "Zen" based silicon, works out to a significant cost saving for the company. A 4-die/32-core MCM is 0.59X the cost of a hypothetical monolithic 32-core die, according to the company, which is a cost-saving that enables the company to aggressively price its products. The slide AMD used in its presentation also confirms that each 8-core "Summit Ridge" die features four external Infinity Fabric links, besides the one that connects the two CCX units with each other. On a 4-die EPYC MCM, three out of four of those external IF links wire out to the neighboring dies, and one link per die probably wires out to a neighboring socket on 2P machines.

An issue on AMD's Ryzen performance under certain Linux workloads, which caused segmentation faults in very heavy, continuous workloads on the Ryzen silicon (parallel compilation workloads in particular) has been confirmed by AMD. Tests like Phoronix's Test Suite's stress run quickly bring the Ryzen processors to their knees with multiple segmentation faults. While this problem is easy to cause under very heavy workloads, the issue is virtually absent under normal Linux desktop workloads and benchmarking,

AMD also confirmed this issue is not present in EPYC or Threadripper processors, but are isolated to early Ryzen samples under Linux (AMD's testing under Windows has found no such behavior.) AMD's analysis has also found that these Ryzen segmentation faults aren't isolated to a particular motherboard vendor, but are problems with the processors themselves. AMD encourages Ryzen customers who believe to be affected by the problem to contact AMD Customer Care. Some of those who have contacted customer care about the segmentation faults have in turn been affected by thermal, power, or other problems, but AMD says they are committed to working with those encountering this performance marginality issue under Linux. AMD will also be stepping up their Linux testing/QA for future consumer products.

Noctua today presented three high-end heatsinks for the TR4 and SP3 sockets of AMD's latest Ryzen Threadripper (X399) and Epyc platforms. The new TR4-SP3 versions of the award-winning NH-U14S, NH-U12S and NH-U9 coolers are tailored custom models that feature a larger contact surface as well as a SecuFirm2 mounting system for TR4/SP3. Supplied with Noctua's renowned NF-A15, NF-F12 and NF-A9 premium fans, the new coolers combine outstanding cooling performance with exceptional quietness of operation.

"A high-end platform usually calls for high-end cooling solutions and that's precisely the case with AMD's new Epyc and upcoming Ryzen Threadripper ecosystems", explains Roland Mossig (Noctua CEO). "We have therefore customized our award-winning NH-U14S, NH-U12S and NH-U9 coolers so that users can enjoy the performance and quietness of these proven premium heatsinks on the new AMD platforms."

It has been a huge weekend of product announcements and launches from AMD, which expanded not just its client computing CPU lineup on both ends, but also expanded its Radeon graphics cards family with both client- and professional-segment graphics cards. This article provides a brief summary of everything AMD launched or announced today, with their possible market-availability dates.

AMD today officially announced some more details on its brain-child and market-stormer Ryzen Threadripper HEDT line of CPUs. Ryzen is a true new stand-alone architecture for AMD, the result of more than four years of careful planning and silicon design towards reaching a truly scalable, highly-flexible, non-glued together MCM design that could power all experiences and workloads through a single architecture design. The Ryzen architecture is already powering desktops with Ryzen 3, 5 and 7 desktop CPUs; has extended to server-side deployments through its EPYC line-up and will begin shipping for professionals with Ryzen PRO starting in Q3 2017. Also announced was that it will find its way to mobile APUs around Q4, paired with the new Vega graphics microarchitecture; and will even power professional-geared mobile solutions in 1H18. But more immediately, it's coming to the HEDT market. And AMD is putting that fight in the hands of Threadripper.

AMD pits its HEDT line-up to developers, researchers, prosumers, creators, and even multi-tasking gamers. Increased compute capabilities with up to 16 cores and 32 threads; larger memory footprint, increased I/O and storage, and support for many more GPUs and PCIe lanes ensure a stable, impressive platform for today's large data sets and tomorrow's exponentially more resource-intensive workloads. AMD will execute this with a three-pronged approach. There will be three processor models on offer for their HEDT platform. The $999 TR 1950X and $799 TR 1920X are known quantities already, with their respective 16 cores (32 threads) and 12 cores (24 threads). The new addition, however, comes in the form of the $549 TR 1900X, which offers not only 8 cores (16 threads) and 3.8 GHz base, 4.0 GHz boost clocks, but a clear upgrade path within AMD's new platform. Say what you will about AMD's offerings and execution, one thing is for sure: Zen and all the silicon it powers have prompted a reshuffle of the CPU landscape as we hadn't seen in years. Coincidence? AMD doesn't think so.

Overclocker extraordinaire der8auer, who was one of the most vocal enthusiasts calling out for better VRM designs on Intel's X299 platform (and who worked with ASUS on redesigning the VRM cooling in its motherboards) has gone and done it: he delidded a Ryzen Threadripper CPU. And this delidding went on to deliver the goods: Ryzen Threadripper delidded is EPYC (pun intended.)

Instead of the expected MCM composed of two dies (with two CCXs of four cores per die, delivering the 16 cores we were expecting), Ryzen Threadripper is actually a much more interesting chip: it seems to be a full fledged EPYC chip, with four dies of eight cores. According to der8auer, when questioned, AMD confirmed that 16-core Threadripper 1950X CPUs are configured with two working eight core dies (four CCXs of four cores each), while the other eight-core dies are disabled by AMD.

So, yes, Intel, I think the AMD engineers who have developed the Zen architecture from the ground-up would take issue with that. Especially when AMD's "Glued-together" dies actually wipe the proverbial floor with the blue company's chips in power-performance ratios, and deliver much better multi-threaded performance than Intel's offerings. Not bad for a "Glued-together" solution, I'd say.

Our resident W1zzard had this to say regarding AMD's latest CPUs: "The SenseMi power-management system seems to be working well in idle, with the 8-core machine drawing the same amount of power as Intel's quad-core "Kaby Lake" machine." And "At stock speeds, the energy-efficiency of Ryzen is truly phenomenal. Prime95 loads all cores and threads on the chip, and the Ryzen ends up with as much power draw as the quad-core Intel i7-7700K. The high power draw result of the overclocked chip is due to the increased voltage needed to achieve stable operation." And let's not forget this: This is epic. We're assuming you've sifted through our game-test results before seeing this page, and so you'll find that the gaming power draw of the 8-core Ryzen makes Intel's quad-core i7-7700K look bad. Power draw is as much as 30W lesser! Ryzen is hands down the most energy-efficient performance CPU AMD ever made, and easily outclasses Intel's 14 nm "leadership." Good show."

We have included the full AMD presentation deck at the end, so be sure to load the entire story.

Following the global excitement generated by the launch of its new EPYC family of server processors, AMD (NASDAQ: AMD) today added another tier to its enterprise CPU portfolio with the introduction of AMD Ryzen PRO desktop processors. Designed to meet the demands of today's compute-intensive workplace, Ryzen PRO desktop processors will bring reliability, security, and performance to enterprise desktops worldwide.

"Today marks another important step in our journey to bring innovation and excitement back to the PC industry: the launch of our Ryzen PRO desktop CPUs that will bring disruptive levels of performance to the premium commercial market," said Jim Anderson, senior vice president and general manager, Computing and Graphics Group, AMD. "Offering a significant leap in generational performance, leadership multi-threaded performance, and the first-ever 8-core,16-thread CPU for commercial-grade PCs, Ryzen PRO provides a portfolio of technology choices that meet the evolving needs of businesses today and tomorrow." Ryzen PRO Lineup Delivering breakthrough responsiveness for the most demanding enterprise-class applications and multi-tasking workflows, the 'Zen' core in every Ryzen PRO processor provides up to 52 percent improvement in compute capability over the previous generation, and the Ryzen 7 PRO 1700 offers up to 62 percent more multi-threaded performance than select competing solutions.

GIGABYTE introduced the MZ30-AR0 motherboard for single-socket AMD EPYC processor-powered servers and workstations. The motherboard is built in the E-ATX form-factor, and features a single SP3r2 socket, for AMD EPYC 7000-series processors. Given that EPYC is a full-fledged SoC, the board has no chipset. An ASpeed AST2500 remote-management chip puts out basic display and IPMI features. The board draws power from two 8-pin EPS connectors, besides the 24-pin ATX. The CPU socket is flanked by 16 DDR4 DIMM slots, which support up to 512 GB of octa-channel DDR4 memory.

Expansion slots include four PCI-Express 3.0 x16 with full-time x16 wiring, one x16 with x8 wiring, and two x8 slots. Storage connectivity includes four slimSAS 12 Gb/s ports, which put out sixteen SATA 6 Gb/s ports, and a 32 Gb/s M.2 slot. Networking is care of two 10 GbE ports driven by a Broadcom BCM57810S processor, and a single GbE port. There's no onboard audio or even USB 3.1 ports.

AMD has unveiled the pricing scheme for its latest EPYC line of datacenter processors. In a series of graphs and tables at it's EPYC launch presentation, it outlined a comprehensive platform that it claims beats Intel on a performance per dollar basis across the entire 64 thread spectrum. What is known up to now in regards to pricing can be summarized in this nice table we have made for you below:

Today, AMD, along with its global ecosystem of server customers and partners, launched the EPYC 7000 series of high-performance datacenter processors. With up to 32 high-performance "Zen" cores and an unparalleled feature set, the record-setting AMD EPYC design delivers greater performance across a full range of integer, floating point, memory bandwidth, and I/O benchmarks and workloads.

A set of details on AMD's upcoming Naples platform's family of EPYC CPUs has leaked on the site videocardz.com, claiming that the top product, the EPYC 7601, will feature a turbo clock of no less than 3.2 GHz (base clocks are more moderate at around 2.2 GHz for the EPYC 7601), with a core count of 32 Zen units, and 64 threads. These clocks are pretty high for a 32-core CPU, and will probably be only for a handful of cores at a time. The EPYC 7601 is also a hot chip according to this leak, in both thermals and price. It costs over $4000 USD, and has a TDP of 180 Watts (not bad at all for a 32-core part, though!)

The rest of the lineup is detailed as well, but as this is a leak that is blatantly admitting to violating an NDA, it goes without saying this could be nothing more than a fabrication. Take it with your usual dose of healthy skepticism.

For the full details of the leak (including the rest of the expected lineup), you can view the source link below.

At its Computex 2017 event, AMD announced the market availability of EPYC enterprise processors based on the "Zen" architecture. The chips will be available from June 20, 2017. The company didn't get into the specifics, however we know that EPYC consists of chips with up to 32 cores, up to 64 threads, up to 32 MB of L3 cache, and a vast 112-lane PCIe root complex. There could be 1P and 2P models.

On the back of impressive performance, yield, and cost metric for AMD's market-warping Ryzen and server-shaking EPYC processors, securities firm Rosenblatt Securities' Hans Mosesmann has affirmed a "Buy" rating for AMD's stock, while saddling Intel with a seldom-seen "Sell". All in all, there have been a number of changes in Intel's market ratings; there seems to be a downgrade trend towards either "Hold" or "Sell" scenarios compared to the usual "Buy" ratings given by hedge funds and financial analysts - ratings which are undoubtedly affected (at least in part) by AMD's Ryzen and EPYC execution.

Today on their Financial Analyst Day 2017, AMD has taken the lid off their "Naples" Zen implementation. The balanced Zen core in its unrestrained, server-grade level has become EPYC, with AMD CEO Lisa Su holding the silicon in her bare hands. The new EPYC platform with its I/O performance improvements allows more GPUs to be connected to a CPU than any other platform, with up to 128 PCIe lanes being expected on these server-grade chips.