The following story was originally posted on ShroutResearch.com.
AMD announced today that it has hired two new executives to run its graphics division after the departure of Radeon Technologies Group’s previous lead. Raja Koduri left AMD in November to join Intel and launch its new Core and Visual Computing group, creating a hole in the leadership of this critical division at AMD. CEO Lisa Su filled in during Koduri’s sabbatical and subsequent exit, but the company had been searching for the right replacements since late last year.
Appointed as the senior vice president and GM of the Radeon Technologies Group, Mike Rayfield comes to AMD from previous stints at both Micron and NVIDIA. Rayfield will cover all aspects of the business management of AMD’s graphics division, including consumer, professional, game consoles, and the semi-custom division that recently announced a partnership with Intel. At Micron he served as the senior vice president of the Mobile Business Unit, responsible for company’s direction in working with wireless technology providers (smart phones, tablets, etc.) across various memory categories. While at NVIDIA, Rayfield was the general manager of the Mobile Business Unit helping to create the Tegra brand and products. Though in a different division at the time, Rayfield’s knowledge and experience in the NVIDIA organization may help AMD better address the graphics markets.
David Wang is now the senior vice president of engineering for the AMD Radeon Technologies Group and is responsible for the development of new graphics architectures, the hardware and software that integrate them, and the future strategy of where AMD will invest in graphics R&D. Wang is an alumni of AMD, working as corporate vice president for graphics IP and chip development before leaving in 2012 for Synaptics. David has more than 25 years of graphics and silicon experience, starting at LSI Logic, through ArtX, then ATI, before being acquired by AMD.
The hires come at a critical time for AMD. Though the processor division responsible for the Zen architecture and Ryzen/EPYC processors continues to make strong movement against the Intel dominated space, NVIDIA’s stranglehold on the graphics markets for gaming, machine learning, and autonomous driving are expanding the gap between the graphics chip vendors. The Vega architecture was meant to close it (at least somewhat) but NVIDIA remains the leader in the space by a not insignificant margin. Changing that is and should be AMD’s primary goal for the next few years.
AMD is hoping that by creating this two-headed spear of leadership for its Radeon graphics division it can get the group back on track. Rayfield will be taking over all business aspects of the graphics portion of AMD and that includes the addition of the semi-custom segment, previously a part of the EESC (Enterprise, Embedded, and Semi-Custom) group under senior vice president Forrest Norrod. AMD believes that with the growth and expansion of the enterprise segment with its EPYC processor family, and because the emphasis on the semi-custom group continues to be the advantage AMD holds in its graphics portfolio, the long-term strategy can be better executed with that group under the Radeon Technologies umbrella.
The return of Wang as the technical lead for the graphics division could bring significant positive momentum to the group that has struggled in the weeks leading up to the release of its Vega architecture. The product family based on that tech underwhelmed and had concerns over availability, pricing, and timing. Wang has a strong history in the graphics field, with experience as far back as any high-level graphics executive in the business. While at ATI and AMD, Wang worked on architectures from 2002 through 2012, with several periods of graphics leadership under his belt. Competing against the giant that NVIDIA has become will be a challenge that requires significant technical knowledge and risk-taking and Wang has the acumen to get it done.
AMD CEO Lisa Su expressed excitement and trust in the new graphics executives. “Mike and David are industry leaders who bring proven track records of delivering profitable business growth and leadership product roadmaps,” she says. “We enter 2018 with incredible momentum for our graphics business based on the full set of GPU products we introduced last year for the consumer, professional, and machine learning markets. Under Mike and David’s leadership, I am confident we will continue to grow the footprint of Radeon across the gaming, immersive, and GPU compute markets.”
I wonder how many cards AMD
I wonder how many cards AMD is selling? Right now the market for Vega (and Polaris too) is almost entirely disconnected from its gaming performance.
As of right now AMD is nearly
As of right now AMD is nearly out of stock on all their GPU SKUs and continue to sell out within a day or 2 when new inventory comes in. (this info is based on nowinstock.net) This is mainly because of Crypto mining and it is honestly the saving grace for AMD on the GPU side. If they can manage to get a new GPU that can compete quick enough they may not miss a beat.
Well that’s good for
Well that’s good for AMD(Selling out of GPU stock) and Nvidia appears to be touting Volta’s compute/AI ability also so that’s where the market is leadimg towards GPUs used for compute and AI even more than graphics.
If you look at The Vega GPU micro-arch on it’s own and not tied to any graphics only usage then Raja did what was asked of of him by AMD(nCUs, HBCC/HBC, etc.) Vega’s gaming problems are mostly tied to that one Vega 10 base die design and not really the Vega GPU micro-arch as the Vega 10 base die design is more shader compute heavy with only enough maximum ROPs(64) offered under that Vega 10 base die design variant(Single variant) to compete with the GP104(64 ROPs max) based GTX 1080. And the Vega 10 base die’s design was frozen at a time before there was that GP102 based GTX 1080Ti(88 ROPs out of a maximum of 96 ROPs offered by the GP102) base die design.
So Nvidia tapes out 4 different base die designs(GP100, GP102, GP104, GP106, GP108) that span the performance envelope top to bottom better than AMD’s only one Vega 10 base die design can. This gives Nvidia an advantage that AMD curretly can not match, with Nvidia able to throw more ROPs out there using a different base die design(GP102 at 96 ROPs) instead of GP104(64 ROPs max) and to make the GTX 1080Ti(88 ROPs) the real winner for Pixel fill rates against Vega 56/64(both based on the Vega 10 die with 64 ROPs max) and even the GP104 based GTX 1080(64 ROPs max).
What AMD lacks is more specilized GPU Die tape-outs compared to Nvidia namely tapeouts with more ROPs available for more powerful Raster back ends where the pixel fill rates can fill the frame buffer faster and produce those FPS metrics that everybody uses to judge a GPUs gaming performance with. Vega’s GPU micro-arch is with that HBCC/HBC IP and HBM2 able to be utilized as a last level cache and even explicit primitive shaders and FP16 packed math made use of by new games and other such improvments will still be useful in Vega discrete mobile variants. This is especially true for any Vega discrete mobile GPU varants with 4GB or less of HBM2 that can make use of Vega’s HBCC/HBC-HBM2(Cache) for games that may benfiit from more virtual VRAM paged in and out of HBM2(HBC) by Vega’s HBCC.
So all AMD has to do even before Navi is ready is respin a new Vega base die variant that’s has more ROPs available than the 64 maximum offered by the Vega 10 base die design and AMD could compete better for gaming. It all depends on how far along Navi is and Navi appears to be little more than the Vega GPU micro-arch with tweaks with the real revolutionary part of Navi’s design the part that will be those modular Navi dies that will allow AMD to do similar to Zen/Zeppelin sorts of die scaling with GPUs in addition to CPUs as is done with Zen/Zeppelin modular dies.
So David Wang’s task is to possibly respin a Vega based die Variant with more ROPs or if Navi is far enough along then maybe that will be better. AMD needs to proparly fund RTG to tape out multiple Vega base die variants if there is time before Navi is ready or just move on to getting Navi’s modular die based designs ready. And Navi should allow AMD to do with GPUs what AMD does with Zen/Zeppelin die based CPUs. So a modular Navi die and AMD can scale up at will just by adding more Navi dies with no need for as many base GPU die tapeouts.
sounds good either way,
sounds good either way, thanks for the analysis
amd: go su
intel: go sue
Logorrhea much?
Logorrhea much?
Not a happy camper are we
Not a happy camper are we WhyMe! You are always quick to register you dislike but always lack the skills to properly debate technology.
What the problem there Bubba, is it that technology discussion should be treated like some sporting match where those with the smallest of brains can easily comment with grunts and hisses and still be considered acceptable.
Poor little WHyMe that little child(Webbed Hands And Feet) inside get so insulted by even the slightest wording that goes beyond any grunts and hisses. You poor little peckerwood down in that rusting doublewide by the abandoned coal mine! Maybe your Swife(sister wife) Brandeen will bake you some rhubarb pie.
no he said you’ve never seen
no he said you’ve never seen a vagina. He didn’t say you couldn’t debate technology
While you did get a laugh out
While you did get a laugh out of me, this thread is coming close to being deleted.
you hold all the threads, or
you hold all the threads, or is it strings
My Vega 64 Liquid was a good
My Vega 64 Liquid was a good buy. I can sell it on ebay for $1500 and it plays games.
Here is a question for Shrout
Here is a question for Shrout Research. Samsung has just released a custom ARM core that is twice as wide order superscalar as the usual ARM holdings refrence design cores. The Samsung Exynos M3(1) is very similar in design to the Apple A series cores now that Samsung has started to actually design for extra wide order superscalar on its custom ARMv8A running custom CPU/Exynos M3 micro-arch.
So could Shrout Research go in search of AMD’s custom K12 project and try to get some definitive answers from Lisa Su on that other Jim Keller project(K12) that Keller/His K12 team completed at around the same time that Keller/Zen Team was working up the Zen/x86 CPU micro-arch.
That Samsung Exynos M3 looks to be even more on the back end execution pipeline parts to be even more powerful than Apples A series cores and this new Exynos M3 is startiing to look more like a desktop class CPU core rather than only for phones/tablets. I hate that AMD is being to focused on x86 and trying to become a little version of Intel when AMD shoule be Bring its K12 to market also and not becoming to dependent on any x86/CISC ISA only markets.
(1)
“The Samsung Exynos M3 – 6-wide Decode With 50%+ IPC Increase”
https://www.anandtech.com/show/12361/samsung-exynos-m3-architecture
AMD ran ATI graphics into the
AMD ran ATI graphics into the ground. It’s such a shame, ATI’s last card they engineered was the R9 290 and it was a terrific value and had excellent performance. The 290 also aged incredibly well and was respun repeatedly. As a Canadian I wish ATI never sold out, I miss competition in the GPU space.
No AMD did not, It’s just
No AMD did not, It’s just that the gamers’ little money does not speak as loudly as the Pro Compute/AI market’s money or even the Coin Miners’ Money.
So there you have it there, Canadian, if you want to game it’s not a good time because AMD’s Lisa Su says that AMD can no longer be the low price GPU offering any longer! And it’s the compute/AI market that BOTH Nvidia and AMD will be going after. Gamers who can not pay can eat console for their gaming nutrition going forward.
There is competition in the GPU space mostly focusing on the Compute/AI markets and Nvidia spends more money on gamers because gaming still makes up about 58% of Nvidia’s total revenues. You Just wait for old JHH over at Nvidia to start getting more of Nvidia revenues from non gaming GPU sales. But currently Nvidia has many GPU base die tape-outs(GP100, GP102, GP104, GP106, GP108) compared to AMDs only one current Vega 10 base die tape-out(discret GPU die) that only offers 64 ROPs max compared to Nvidia’s GP102(96 available ROPs) with the GP102 based GTX 1080Ti using 88 of the GP102’s 96 available ROPs.
Vega 20 at 7nm is going to be a Compute/AI monster for AMD and let’s hope that there will be dual core on single PCIe card variants also wired up via the Infinity Fabric.
Nvidia has the gaming market with that GP102 and there is still more ROPs(96) available for Nvidia to make use of if they need to best the GTX 1080Ti’s 88 ROPs.
So Please fill free to demand that AMD tape out a new Vega micro-arch base die variant with more than 64 available ROPs. Nvidia tapes out 5 base die variants all with varying amounts of ROPs/other resources. And ROPs are what fling the frames for gaming mostly and AMD’s shader heavy designs are loved by the miners/others for that compute(hashing).
AMD’s loving those miners and the folks buying those Project 47 petaflops supercomputer in a single cabinet. That Project 47 single cabnet system takes 80 Vega 10 base die based Radeon Instinct MI25 SKUs per cabinet along with 20, 32 core Epyc CPUs and 20 Epyc/SP3 MBs. There is where Vega is going inside that and Vega 10 bsaed Radeon Pro WX 9100’s and WX 9100-SSD variants. Vega is a hit with the folks with the deep pockets and sucess is not measured by gaming sales only anymore for both AMD and Nvidia.
I bought a 1080 Ti a 3 months
I bought a 1080 Ti a 3 months ago for less than MSRP. That’s not what I’m bitching about, I’m complaining about gaming performance on every AMD GPU since the 290. I understand AMD is making money off miners but they completed folded on the gaming market.