AMD Execution Thread [2023]

Status
Not open for further replies.
TSMC N6 is supposed to be a "cost effective" replacement of N7, which should have been well under $10k in 2022, I think we were throwing around the ~$7-$8k number last year.
Do you have a quote from Ian about TSMC N6 cost? Even after all of TSMC's price hikes, I can't believe they moved a cost effective node up close to 2018 introductory prices...
Also, Nvidia's custom N4 is likely on the high side of the 5nm wafer cost range(I think ~$16-$18k was mentioned last time), even with their volume discount.

I previously estimated N31 at ~$120-$130 for pure die costs. ~300mm2 GCD- ~$90-$95 each (<$100), MCD- ~$4 each (<$5).
Using similar numbers, I would guess N32 is around $80-$85. ~200mm2 GCD- ~$65 each, MCD ~$4 each.
Edit- AMD might be selling a N32 package for ~$110-$120.

Apparently, AMD's whole focus for RDNA3 was cost as they made quite a few comments about PPA during their presentation last year.
Here you go:

But actually it seems I somehow miscalculated the MCD cost, so it should be ~$91 with those wafer prices. So maybe a bit below the full AD104 when packaging costs are included. The 7800 XT does have an extra 4GB of memory though when compared to the 4070/4070 Ti.
 

You need these MODDED AMD Drivers! HYPR-RX UNLOCKED for ALL GPUS!


Careful if anyone tries it, it appears to make some games entirely unplayable on RDNA2 even when not enabling HyperRX. Appears to also screw with Netflix video playback for a lot of people.

They only way they were able to get things to work again in some games was to use DDU to remove the modded drivers and revert back to an official driver.

Regards,
SB
 
The RDNA version contains:
  • New AMD Radeon™ Anti-Lag+ Game Support
    • AMD Software: Adrenalin Edition 23.9.2 introduces Anti-Lag+ support for Starfield, Witcher 3, ELDEN RING and Immortal of Aveum™.
      • Up to 45% decrease in latency across select titles when AMD Radeon™ Anti-Lag+ is on, using AMD Software: Adrenalin Edition™️ 23.9.2 on the Radeon™ RX 7900 XTX GPU in select titles, versus when AMD Radeon Anti-Lag+ is off.RS-597
    • AMD Radeon™ Anti-Lag+ features an onscreen overlay that can be used to display the system latency of supported games. When Anti-Lag+ is enabled, the onscreen overlay can be toggled on using ALT+SHIFT+L hotkey.
      • Toggling the hotkey will first enable the status indicator of Anti-Lag+ (a white triangle), and then display latency in ms or number of frames.
      • To compare the difference between Anti-Lag+ and Anti-Lag, hold the DEL key. To compare the difference between Anti-Lag+ On and Off, hold the right CTRL key.
      • Use the ALT+SHIFT+F hotkey to monitor FPS when Anti-Lag or Anti-Lag+ is enabled in the game.
 
23.9.2's are out, they've split them into separate RDNA and GCN drivers, with latter only getting new game support
Modern gfx APIs must have been a godsend to them since if they feel confident behind their driver implementation to make this move ...

Even if they do drop official support, many apps in the future released with D3D12 will very likely continue to run fine for years to come as long they don't require any unsupported features ...
 
September 25, 2023
AMD’s software stacks for each class of product are separate: ROCm (short for Radeon Open Compute platform) targets its Instinct data center GPU lines (and, soon, its Radeon consumer GPUs), Vitis AI targets its FPGAs, and ZenDNN targets its client devices.

How far along is AMD with unifying these stacks?

“We have enormous customer pull coming, and that is dictating quite a bit of our near-term plans,” Boppana told EE Times in an interview after his talk here. “The plane is flying right now, so we cannot disassemble the engine. However, we are absolutely doing things at the foundational level to make more unification happen in our stack.”
 

“I will miss every single one of you, fighting shoulder to shoulder in the trenches together, the excitement we shared during new product launches, and the joy of being in the arena for this wonderful, vibrant industry.

May you continue to punch above your weight class and one day… beat the final boss.”


Is Jensen the final boss?
 
Nah, AMD didn't gut client gfx BU back then.
Either they're gutting it or Scott was let go for the recent massive messaging fuckups.
Maybe Scott and Ryan switch places :ROFLMAO: (Shrout just announced he quit at Intel and will let know his next plans in couple weeks)
 

Lamini, led by machine learning expert Sharon Zhou (CEO) and ex-NVIDIA CUDA software architect Greg Diamos (CTO), announced that they have been building LLMs in production for over a year on AMD GPUs and ROCm with customers. They claim ROCm has reached software parity with CUDA for LLMs
 
Status
Not open for further replies.
Back
Top