I'm just hoping that its fully implemented not the we will just combine two AVX 128 piplines ala how they are doing AVX 256 on Zen
One would hope they have had the foresight to do this but I'm not holding out for it. AVX512 is a very, very limited use case, the RTX of processor world at this point in time. Don't get me wrong it has uses though but just not very widespread at all. IIRC there were only two common apps that could make use of AVX-512 in past.
 
One would hope they have had the foresight to do this but I'm not holding out for it. AVX512 is a very, very limited use case, the RTX of processor world at this point in time. Don't get me wrong it has uses though but just not very widespread at all. IIRC there were only two common apps that could make use of AVX-512 in past.

I'm watching from a HPC perspective and AVX512 has been taking off fairly quickly. I will probably be buying some cluster nodes in the near future and the CPU density of the Epyc 2 at hopefully a lower cost than intel has me intrigued.

It might not be a big deal for these nodes but I know we have a lot of researchers optimizing code for some of these things. I also know I'm very much in a niche case compared to most.
 
  • Like
Reactions: N4CR
like this
the only unofficial clocks for the 64/128 cpu i've seen was 1.4Ghz base 2.0Ghz max boost on an engineering sample that was posted on sisoftware.. we won't know til it's officially launched what the actual numbers end up being but it's probably some where around there.
That would imply some great IPC gains. Like I said lets wait for the official 3rd party analysis. Fansites (ermmmm) and marketing tend to release misleading information. If true, awesome. True competition! On server side we aren't concerned with OC/Unlocked/base speed, it's how well it performs to cost and generates a much greater profit for the provider. AMD hasn't been able to compete in that sector for many years. This would change a lot if true.

Intel hire your CEO, you are likely losing in the biggest market lol.
 
I'm just hoping that its fully implemented not the we will just combine two AVX 128 piplines ala how they are doing AVX 256 on Zen

Zen2 increases units/datapaths from 128bit to 256bit

fpu.png
 
Back
Top