News: 1655742605

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Former AMD chip architect says it was wrong to can Arm project

(2022/06/20)


Amid the renewed interest in Arm-based servers, it is easy to forget that one company with experience in building server platforms actually brought to market its own Arm-based processor before apparently losing interest: AMD.

Now it has emerged that Jim Keller, a key architect who worked on Arm development at AMD, reckons the chipmaker was wrong to halt the project after he left the company in 2016.

Keller was speaking at an event in April, and gave a talk on the "Future of Compute", but the remarks were unreported until picked up by [1]WCCF TECH .

[2]

In the talk, available [3]online via YouTube, Keller discusses how when planning the Zen 3 core – now at the heart of AMD's "Milan" Epyc processor chips – he and other engineers realized that much of the architecture was very similar for Arm and X86 "because all modern computers are actually RISC machines inside," and hence according to Keller, "the only blocks you have to change are the [instruction] decoders, so we were looking to build a computer that could do either, although they stupidly cancelled that project."

[4]

[5]

That project was apparently the K12, which was planned to be AMD's first custom microarchitecture based on the 64-bit ARMv8-A instruction set, and would have led to chips that would follow on after the Opteron A1100 series chips, which were based on Arm's Cortex-A57 core designs.

Keller's involvement at AMD goes back a long way, as he worked on the Athlon (K7) processor and the K8 microarchitecture, which introduced a 64-bit instruction set for X86 processors and was thus the foundation of the first generation of Opteron chips, which saw AMD claim a sizable fraction of the X86 server market in the first decade of this century.

[6]

After working at SiByte (later acquired by Broadcom), PA Semiconductor, and then Apple, Keller rejoined AMD in 2012 to lead development of the first Zen microarchitecture and on Arm-based chips. He [7]left again , to join Tesla in 2016, before the first Zen-based Epyc chips were launched.

Keller went to Intel in 2018 and [8]ejected in 2020. He then rocked up at [9]Tenstorrent .

[10]Nvidia wants to lure you to the Arm side with fresh server bait

[11]AMD beds Arm in 'Project SkyBridge' and home-grown 'K12' ARM architecture chips

[12]AMD bests Intel in cloud CPU performance study

[13]US to help Japan make leading-edge 2nm chips, possibly by 2025

[14]Who's growing faster than Nvidia and AMD? Rising datacenter star Marvell

The Opteron A1100, codenamed [15]Seattle , was [16]launched in early 2016 with up to eight cores and clock speeds of 1.7–2.0GHz. AMD had originally planned to follow it up with a sixteen core version, but by the end of 2016 the company had apparently changed its mind and decided to focus all its efforts on the X86 architecture and the Zen core designs.

Since then, things have changed, with some of the cloud providers introducing Arm-based servers, such as AWS operating [17]EC2 instances based on its own Graviton chips and Microsoft offering Azure virtual machines powered by [18]Ampere 's Altra Arm-based processors.

Who knows, perhaps AMD might be tempted to revive its K12 project and bring it up to date if interest in Arm servers continues to blossom? ®

Get our [19]Tech Resources



[1] https://wccftech.com/legendary-chip-architect-jim-keller-says-amd-stupidly-cancelled-k12-arm-cpu-project-after-he-left-the-company/

[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YrDuCSFqTz8jnnNmfhUY5gAAAIQ&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[3] https://www.youtube.com/watch?v=OvgdU5FkG-0&t=1525s

[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YrDuCSFqTz8jnnNmfhUY5gAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YrDuCSFqTz8jnnNmfhUY5gAAAIQ&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YrDuCSFqTz8jnnNmfhUY5gAAAIQ&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[7] https://www.theregister.com/2015/09/20/jim_keller_quits_amd/

[8] https://www.theregister.com/2020/06/11/jim_keller_intel/

[9] https://www.theregister.com/2021/04/22/tenstorrent_sifive_intelligence_x280_licensed/

[10] https://www.theregister.com/2022/06/17/nvidia_arm_servers/

[11] https://www.theregister.com/2014/05/05/amd_announces_project_sky_bridge_and_k12_demos_seattle/

[12] https://www.theregister.com/2022/06/16/amd_beats_intel_in_cloud/

[13] https://www.theregister.com/2022/06/15/us_japan_chips/

[14] https://www.theregister.com/2022/06/14/fabless_marvell_nvidia_amd/

[15] https://www.theregister.com/2018/11/27/amazon_aws_graviton_specs/

[16] https://www.theregister.com/2016/01/14/amd_arm_seattle_launch/

[17] https://www.theregister.com/2022/03/29/aws_arm_servers_datacenters/

[18] https://www.theregister.com/2022/05/17/arm_datacenter_certification/

[19] https://whitepapers.theregister.com/



Low bus factor

Gene Cash

So it looks like this had a low "bus factor" - I have a feeling Keller was the major driver and major repository of ARM knowledge in the company, so when he left, they didn't have much choice other than to cancel it.

I wonder if his leaving is a underlying cause of why AMD hasn't been such a CPU powerhouse recently... just a SWAG.

Re: Low bus factor

cornetman

> I wonder if his leaving is a underlying cause of why AMD hasn't been such a CPU powerhouse recently... just a SWAG.

Not sure where you've been for the last 3-4 years. AMD are destroying Intel in various spaces.

My suspicion is that the secret to Lisa Su's success with AMD have been due to the following factors:

1. Laser focus on important tech areas that advance their portfolio in meaningful ways. The dropping of ARM might be related to avoid stretching their R&D resources too thin. AMD are a big company but they are still dwarfed by the likes of Intel. They surely cannot be as diverse as Intel. I also think this is the reason in part why their Radeon group has been so slow initially. I know that lots of people were really disappointed with early offerings on the graphical side.

2. Slow yet merciless execution with a clear 5-10 year plan.

3. Key strategic partnerships with key players. Keeping their foot in the door with Microsoft and Sony must have been a great money spinner for AMD, getting them through the lean periods during key development phases.

AMD have a lot of loyal fans due to their underdog status. I try to choose AMD over Intel because I hate the shady practices of Intel in the past and I think that honest, steady, hard-working face resonates with a lot of consumers. It certainly does with me.

Re: Low bus factor

martinusher

ARM also has a big "Me Too" factor -- everyone and their dog is developing it.

I figure if they're going ARM-like it will probably be with RISC-V. Even that's got a bit of me too in it.

AMD / XILINX

andrewmm

And now , AMD have bought Xilinx

that a have a good few chips with combinations of AMD CPUs, Memory and programable logic in them

Re: AMD / XILINX

elsergiovolador

Xilinx makes quite a few CPLD and FPGA chips. Hope AMD won't force them to shelve them.

But why bother?

talk_is_cheap

If AMD can advance their x86 based core design at the same speed as the ARM core design, there is little value in splitting their R&D budget and paying third-party additional IP fees. We already know that AMD is planning x86 cores with different levels of functionality and the only thing that seems unclear is how (or if) AMD plans to mix core types on a single CPU.

And AMD Moves On....

William Higinbotham

https://www.tomshardware.com/news/from-opteron-to-milan-crusher-supercomputer-comes-online-with-amd-cpus-and-gpus

I would have you imagine, then, that there exists in the mind of man a block
of wax... and that we remember and know what is imprinted as long as the
image lasts; but when the image is effaced, or cannot be taken, then we
forget or do not know.
-- Plato, Dialogs, Theateus 191

[Quoted in "VMS Internals and Data Structures", V4.4, when
referring to image activation and termination.]