By using this site, you agree to our Privacy Policy and our Terms of Use. Close
Captain_Yuri said:
Cyran said:

100% intel going to support AVX 512 on Sapphire Rapid.  There no other reason why Golden Core would have AVX-512 on it.  Why make a core with AVX-512 if you not going to support it on anything.  

The question is will AMD support AVX512 on Ryzen or will they disable it on ZEN 4 just like intel disabling it on Golden Cove for Alder Lake.

I'd be surprised if AMD disables it in Zen 4 as that would be a pretty easy selling point even if many consumers may not get much use out of it.

AVX-512 is certainly irrelevant today, but who knows in a couple years?

For Instance, Phenom-based processors couldn't start some games after 2015 anymore since those needed SSE 4.1 and/or 4.2, but those only had the AMD-exclusive 4.a. But in 2009, SSE4 was still even just barely used in HPC.

Either way, it would be funny if in a couple years some game would make use of AVX-512 and then run on Rocket lake, but not it's immediate successors...

Captain_Yuri said:

Intel Core i9-12900K 16 Core Alder Lake CPU Benchmarked on ASUS ROG STRIX Z690-E Gaming WIFI Motherboard, Faster Than Core i9-11900K

https://wccftech.com/intel-core-i9-12900k-16-core-desktop-cpu-benchmarked-asus-rog-strix-z690-e-gaming-wifi-motherboard-faster-than-amd-ryzen-9-5950x/

Now these are early samples but man, if this is true, Alder Lake S is gonna be a flop. The benchmark is clearly not taking advantage of the 5950x's 16 cores 32 threads because of how close it is to an 11900k. So the fact that the i9-12900k is this close means that it's in big trouble! We will see what happens when the CPU actually comes out but if this is what Intel's answer to AMD is... This ain't it chief!

We clearly need some more test samples to see the full picture, but since Alder Lake is supposed to be released in October, I doubt it will change much over what we have right now in the leaks.

So for the high end, I expect Intel to revive it's HEDT platform with Sapphire Rapids and fight both the 5950X and the Threadrippers 5960X and 5970X with that one while Alder Lake will compete with the 5900X at best.

And that's all without the boost AMD is about to give their CPUs with the infinity cache they'll be getting for their refresh...

Captain_Yuri said:

Intel DG2-512 specs compared to nVidia GA104 and AMD Navi 22 (as well DG2-128 & DG2-256)

https://www.reddit.com/r/hardware/comments/p95nyq/intel_dg2512_specs_compared_to_nvidia_ga104_and/

Now I wouldn't compare TF between the GPUs as they are meaningless. But it does give us an interesting idea as to the spec differences. I do think that RDNA 2 is lacking in ML performance but we already knew that with DF. The Ray Tracing performance is something I will be interested in seeing.

Interesting would have been to know the TGP or TBP of Arc to see if it can keep up in that regard or run very hot.

Captain_Yuri said:
Cyran said:

Because if you going to let people run AVX-512 you also going to have to support the power consumption and heat it requires.  Good article by AnandTech on just how much more power is require when running AVX-512 work loads

https://www.anandtech.com/show/16495/intel-rocket-lake-14nm-review-11900k-11700k-11600k/5

That's true. Zen 4s 16 cores is rumoured to have a TDP of 170 watts while being on TSMC's 5nm which is up from 5950x 105 Watts TDP. If they end up disabling AVX512, then they really need to make sure those big cores are able to beat Raptor Lake handly which should be it's direct competition. Especially as Intel is going DDR5 and PCI-E Gen 5 while Zen 4 is rumoured to have DDR5 but only PCI-E Gen 4.

It will be interesting to see how it all plays out.

I'm expecting that 170W chip to be something more special.

Either AMD starts rating TDP like on the servers (where the chips with a TDP of 280W also don't consume more than that; a 5950X at full core utilization is more like 145W, add in the extra memory on top that will come with the refresh and you're close to 170W in real terms already), or AMD comes with an chip that's actually a full 7950X (6950X for the upcoming refresh) plus a relatively powerful mobile GPU chip(-let, as in, 16-32CU), as the big cache they are putting on top of the chips with the upcoming refresh would also do really great for an integrated GPU.

Last edited by Bofferbrauer2 - on 22 August 2021