AMD CDNA Discussion Thread

As expected, it's not looking good...

 

42 is still much bigger than 15.

 

only 2.8x the competition ,

 

Yeah, only 2,8-times faster than competition. What a fail!

 

Why does it matter that a single purpose product delivers more performance when the market is limited? Outside of goverments MI200 will be not used in any meaningfull numbers. Nextplatform said that China broke the sustained exaflops performance barrier within 35 MW. I dont see how Frontier should be on par with it when one OAM card has only 44% practical performance. That puts Frontier in the range of ~700 Petaflops at <=33MW.

 

Bingo.
It's made for this exact purpose.

I'm about to die of laughter.
(Let's do the math, ~9216 * 42 * 4 plus whatever peanut Trento gives).

Also it's 29MW you dumdum.
~33W is El Capitan.

 

Hopper and MI300 will be direct competitors launching at about the same time and on the same process node? Both being MCM?

 

Can everyone please take a moment to calm down and return to civil discussions. Leave the petty bickering and name calling out.

 

Sorta, no (MI300 GA is 3Q behind Hopper, same Q as Grace + Hopper), yes.

Kinda.

 

MI300 is going to El Cap. So it should be ready by Q4 2022.

 

Yes but gotta talk GA to GA.
Prod to prod they're ~3Q apart too.

 

Only on paper. Actual numbers are likely different, especially when considering AI workloads.

 

I believe those were measured results from the HPL benchmark so not just paper. Now we can argue that HPL is not representative of anything but it's still the industry standard benchmark.

 

AMD Instinct MI200: Dual-GPU Chiplets and 96 TFLOPS FP64 | Tom's Hardware (tomshardware.com)

 

The 42 and 15 numbers are not theoretical. They're HPL results. The theoretical numbers are 95 and 19.

 

No, according to ORLN's lead engineer David Grant the system is dimensed for 26-28 MW.

https://www.exascaleproject.org/fro...n-and-construction-of-the-mechanical-systems/

Why are some forum members persistently fabricating inferior numbers (like 50 % lower bandwidth than the actual one, ~25 % higher power consumption than the actual one, purposely confusing theoretical and measured performance etc.) and making disastrous speculations based on the biased data? Some posts here sound more like a content created by competitor's marketing team than technological discussion. Everyone can make a mistake, but these "mistakes" are repeatedly CDNA-related and come from people, who are never mistaken when quoting Ampere numbers.

 

It's only standard because it goes back 40 years. HPCG is considered a much more representative benchmark for modern workloads - ones that care much more about memory and cache than simple DPFP throughput. I guess AMD must've forgotten it existed though since they didn't release any figures for it. And obviously they have no reason to hide its HPCG performance as certain forum members who deride recent Chinese efforts as linpack machines are assuring us here it's good for real science™.

 

It's true. HPCG benchmarks on MI200 would be interesting.

 

Yesssssss.
An amazing real science™ stick and no amount of seething ever gonna change that!

Whenever the thing's assembled.

 

if MI200 is so bad, why they buy it? MI100 as a POC has been assessed in lab, right ?