By: Adrian (a.delete@this.acm.org), May 24, 2022 2:39 pm
Room: Moderated Discussions
Michael S (already5chosen.delete@this.yahoo.com) on May 24, 2022 3:48 am wrote:
>
> Did you try to measure efficiency gain yourself?
> For example, AVX-512 vs AVX2 on your own JPEG-XL decode on widely available Tiger Lake or Rocket lake CPU?
> I never did it myself, but would be very surprised if for the whole decode process, from
> getting compressed images from SSD to display you will be able to see measurable difference
> in energy consumption at all. Even for batch-mode from-memory-to-memory decode (but whole
> process, not just DIMS-friendly parts) I would expect reduction of 10-15% at best.
>
I have not tried this on more recent Intel CPUs, but in a measurement on Skylake Server CPUs (with 2 512-bit FMA units) done a few years ago, the ratio between the energies needed to compute some LINPACK benchmark in AVX-512 and in AVX2 (i.e. with 256-bit FMA/LD/ST) modes was around 5/6, so a little more than your maximum estimation, but not much more.
>
> Did you try to measure efficiency gain yourself?
> For example, AVX-512 vs AVX2 on your own JPEG-XL decode on widely available Tiger Lake or Rocket lake CPU?
> I never did it myself, but would be very surprised if for the whole decode process, from
> getting compressed images from SSD to display you will be able to see measurable difference
> in energy consumption at all. Even for batch-mode from-memory-to-memory decode (but whole
> process, not just DIMS-friendly parts) I would expect reduction of 10-15% at best.
>
I have not tried this on more recent Intel CPUs, but in a measurement on Skylake Server CPUs (with 2 512-bit FMA units) done a few years ago, the ratio between the energies needed to compute some LINPACK benchmark in AVX-512 and in AVX2 (i.e. with 256-bit FMA/LD/ST) modes was around 5/6, so a little more than your maximum estimation, but not much more.