Intel ARC GPUs, Xe Architecture for dGPUs

Discussion in 'Architecture and Products' started by DavidGraham, Dec 12, 2018.

Tags:
Page 12 of 42
  1. Bondrewd

    Bondrewd
    Veteran

    Joined:
    Sep 16, 2017
    Messages:
    1,682
    Likes Received:
    846
    ODI is very not soon(TM).
    Also Co-EMIB is not ODI.
     
    #221 Bondrewd, May 19, 2020
  2. del42sa

    del42sa
    Newcomer

    Joined:
    Jun 29, 2017
    Messages:
    208
    Likes Received:
    137
    #222 del42sa, May 20, 2020
    tinokun, Lightman, sonen and 1 other person like this.
  3. Deleted member 13524

    Deleted member 13524 Guest

    Those are excellent results, and they should give Renoir a hard time.

    I wonder how the 48 EU Tiger Lake is being so much better than the 64 EU Ice Lake. Maybe they Xe EUs are wider?
     
    #223 Deleted member 13524, May 20, 2020
  4. CarstenS

    CarstenS
    Legend Subscriber

    Joined:
    May 31, 2002
    Messages:
    5,800
    Likes Received:
    3,920
    Location:
    Germany
    #224 CarstenS, May 20, 2020
    Silent_Buddha, tinokun, Lightman and 1 other person like this.
  5. Dayman1225

    Dayman1225
    Newcomer

    Joined:
    Sep 9, 2017
    Messages:
    77
    Likes Received:
    169
    Raja has posted photos of what appears to be 3 Seperate Xe -HP based GPUs.


    We have seen the one on the left hand side before but the smaller one and larger ones are new.

    Many are speculating that the larger one is a 4 Tile Arctic Sound GPU
    [​IMG]
    And this picture here that Raja posted seems to prove that:
    upload_2020-6-26_22-4-47.png
    ATS = Arctic Sound
    4T = 4 Tiles.

    Raja also gave a vague hint on performance


    Almost 1 PetaOps, many assume this is INT8.
     
    #225 Dayman1225, Jun 26, 2020
    tinokun, Lightman and digitalwanderer like this.
  6. Deleted member 13524

    Deleted member 13524 Guest

    256 TOPS per Arctic Sound chip.
    If that's INT8, then 128 TFLOPs FP16, 64 TFLOPs FP32?

    Naah, way too much. Unless it's using some dedicated tensor units and those are matrix operations like nvidia's hardware.
     
    #226 Deleted member 13524, Jun 26, 2020
  7. Kaotik

    Kaotik Drunk Member
    Legend

    Joined:
    Apr 16, 2003
    Messages:
    10,244
    Likes Received:
    4,462
    Location:
    Finland
    They could very well have some tensor units in there, or some other means to run low precisions at much higher rate.
    These chips fit nicely with the old leak too
    upload_2020-6-27_5-11-3.png
     
    #227 Kaotik, Jun 27, 2020
    Dayman1225 and Lightman like this.
  8. Kaotik

    Kaotik Drunk Member
    Legend

    Joined:
    Apr 16, 2003
    Messages:
    10,244
    Likes Received:
    4,462
    Location:
    Finland
    https://newsroom.intel.com/press-kits/architecture-day-2020

    All kinds of details, also gaming GPUs will be "Xe-HPG" and they'll be made at external foundry (read: TSMC)
    In Xe-LP they've gone from Gen11's 4 FP/Int + 4 FP/ExtMath to 8 FP/Int + 2 ExtMath pipes and two EUs now share Thread controller. 6 texturing units capable of 48 texels/clock and 24 ROPs for 96 EUs

    edit:
    Also, Xe-HP FP32 FLOPS: 1 tile ~10.6 TFLOPS, 2 tiles ~21.2 TFLOPS (1.999x) and 4 tiles ~42,3 TFLOPS (3.993x). 512 EUs running at 1,3 GHz
     
    #228 Kaotik, Aug 13, 2020
    Lightman, tinokun, Deleted member 13524 and 5 others like this.
  9. Deleted member 13524

    Deleted member 13524 Guest

    Could be Samsung. TSMC will be for their CPU but OTOH they seem pretty full with orders.
     
    #229 Deleted member 13524, Aug 13, 2020
  10. digitalwanderer

    digitalwanderer Dangerously Mirthful
    Legend

    Joined:
    Feb 19, 2002
    Messages:
    18,987
    Likes Received:
    3,529
    Location:
    Winfield, IN USA
    Well Raja sure seems excited, so you know it's gonna bomb! :D
     
    #230 digitalwanderer, Aug 13, 2020
    Kej, Lightman, Cuthalu and 6 others like this.
  11. DavidGraham

    DavidGraham
    Veteran

    Joined:
    Dec 22, 2009
    Messages:
    3,976
    Likes Received:
    5,210
    I am not expecting the gaming chips (Xe-HPG) to provide any stellar or competitive performance with AMD or NVIDIA, since it relies on the same scalability scheme as Xe-HPC, ie: relies on racking up several tiles of graphics to scale up core count, this will be a mess for drivers and games in general.
     
    #231 DavidGraham, Aug 13, 2020
    Lightman, digitalwanderer and PSman1700 like this.
  12. Bondrewd

    Bondrewd
    Veteran

    Joined:
    Sep 16, 2017
    Messages:
    1,682
    Likes Received:
    846
    No, those are single dies packed in organic carriers.
    The actual IP is just subpar.
     
    #232 Bondrewd, Aug 14, 2020
  13. DavidGraham

    DavidGraham
    Veteran

    Joined:
    Dec 22, 2009
    Messages:
    3,976
    Likes Received:
    5,210
    The architecture relies purely on software scoreboarding (software schedulers), which means Intel will have it's hands full writing good drivers for it to achieve good utilization (VLIW5 days anyone?), then on top of that they are scaling it up through tiling (multi core/die approach), which means it's going to be a nightmare to write drivers for, and to extract good performance from.

    This is literally the laziest effort for making a new GPU in recent memory.
     
    #233 DavidGraham, Aug 14, 2020
  14. CarstenS

    CarstenS
    Legend Subscriber

    Joined:
    May 31, 2002
    Messages:
    5,800
    Likes Received:
    3,920
    Location:
    Germany
    Removing scoreboarding from hardware enabled the power nightmare that was Fermi to become the somewhat efficient Kepler (among others of course). So, at least for the starting point, which is integrated graphics, that step totally makes sense. And Intel has a ton of software people (idk though if they are necessarily good at gfx driver compilers).
     
    #234 CarstenS, Aug 14, 2020
    Lightman likes this.
  15. Bondrewd

    Bondrewd
    Veteran

    Joined:
    Sep 16, 2017
    Messages:
    1,682
    Likes Received:
    846
    Not really.
    Nah, QC takes the cake.
    Their s/w top talent is very much compiler people so this move isn't unwarranted.
    Unfortunately the IP is still "meh" at best and they clearly lack focus.
    Like dear god, what, 4 flavors of Gen12?
    Why even.
     
    #235 Bondrewd, Aug 14, 2020
  16. DavidGraham

    DavidGraham
    Veteran

    Joined:
    Dec 22, 2009
    Messages:
    3,976
    Likes Received:
    5,210
    It is also one of the reasons why Kepler suck in modern gaming years after the end of life status of it's drivers, that in addition to it's weird FP32 units arrangements, it required very high effort in writing compilers, which didn't really help it on the long run.

    Fermi had troubles with the 40nm process, that was the main reason for it's power hungry status, not hardware schedulers, Tesla and G80 had it before and they were not power hungry chips. Furthermore, Kepler didn't remove them completely, and If I recall correctly, most elements of hardware scheduling came back in Maxwell, Volta and Turing.
     
    #236 DavidGraham, Aug 14, 2020
  17. CarstenS

    CarstenS
    Legend Subscriber

    Joined:
    May 31, 2002
    Messages:
    5,800
    Likes Received:
    3,920
    Location:
    Germany
    Thanks for shortening my quote to fit your narrative. I explicitly said "among others of course)". And yes, Kepler had some failsafe mechanism to keep things in check, in case SW did not work that well.
     
    #237 CarstenS, Aug 14, 2020
  18. Rootax

    Rootax
    Veteran

    Joined:
    Jan 2, 2006
    Messages:
    2,400
    Likes Received:
    1,845
    Location:
    France

    Maybe wait to see real performances before being so affirmative ?
     
    #238 Rootax, Aug 14, 2020
  19. Bondrewd

    Bondrewd
    Veteran

    Joined:
    Sep 16, 2017
    Messages:
    1,682
    Likes Received:
    846
    +20% over Vega8 so I don't even.
     
    #239 Bondrewd, Aug 14, 2020
  20. DavidGraham

    DavidGraham
    Veteran

    Joined:
    Dec 22, 2009
    Messages:
    3,976
    Likes Received:
    5,210
    We have a long history of GPU architectures to judge and forecast performance from, nothing is affirmed of course, but it's worth going through the motions to predict where performance will lie given what we already know from past experiences.

    Furthermore, Xe-LP still retains the abysmal max 1 primitive per clock rate, and worse yet, it lacks all of the features from DX12U, except hardware RT.

    Intel removed hardware scoreboarding from Gen11, which wasn't really that effective there to begin with. Gen11 had one Thread Control unit handling 2 ALUs, each ALU had control over 4 FP32 instructions, so in total each Thread Control unit had access to 8 FP32 instructions, which I would call a pretty weak arrangement to begin with. Intel didn't change this arrangements in Xe-LP, instead it allowed each Thread Control unit to supervise 16 FP32 instructions now, further weakening their already weak position.
     
    #240 DavidGraham, Aug 14, 2020
    Man from Atlantis and digitalwanderer like this.
(You must log in or sign up to post here.)
Page 12 of 42
Loading...
Similar Threads - Intel GPUs Architecture
  1. BRiT

    The Intel Execution in [2022]

    BRiT, Jan 4, 2022, in forum: Graphics and Semiconductor Industry
    Replies:
    10
    Views:
    2,360
    Albuquerque
    May 11, 2022 at 2:51 PM
  2. Jawed

    Intel Xe Ray Tracing

    Jawed, Sep 17, 2021, in forum: Architecture and Products
    Replies:
    4
    Views:
    1,797
    JoeJ
    Sep 18, 2021
  3. green.pixel

    Undervolting older Intel mobile CPU

    green.pixel, Jun 28, 2021, in forum: PC Hardware, Software and Displays
    Replies:
    14
    Views:
    1,705
    green.pixel
    Jul 9, 2021
  4. zamine

    intel array building blocks and GPUs

    zamine, Apr 7, 2013, in forum: GPGPU Technology & Programming
    Replies:
    0
    Views:
    2,313
    zamine
    Apr 7, 2013
  5. Dave Baumann

    NVIDIA Mobile GPUs Support New Intel Centrino

    Dave Baumann, Jan 19, 2005, in forum: Press Releases
    Replies:
    0
    Views:
    1,570
    Dave Baumann
    Jan 19, 2005

Share This Page

Toggle Width
Style
Beyond3D

  • About Us

    Beyond3D has been around for over a decade and prides itself on being the best place on the web for in-depth, technically-driven discussion and analysis of 3D graphics hardware. If you love pixels and transistors, you've come to the right place!

    Beyond3D is proudly published by GPU Tools Ltd.

  • Quick Navigation

Loading...