Re:Iguana: fast SIMD-optimized decompression
by powturbo on Thursday June 01, 2023 @04:05AM
(#36148620)
> and even the performance cores have to be downgraded for it.
Old rumours, it's no longer true.
Re:Iguana: fast SIMD-optimized decompression
by skavi on Thursday June 01, 2023 @04:15AM
(#36148667)
not really rumors, but also not necessarily true. the performance cores on consumer intel parts no longer have AVX-512 enabled. This is because it’s tough to implement on the smaller efficiency cores, and scheduling is more annoying with heterogeneous ISAs.
Re:Iguana: fast SIMD-optimized decompression
by powturbo on Thursday June 01, 2023 @04:35AM
(#36148765)
I understand downgrading as thermal throtlling.
Re:Iguana: fast SIMD-optimized decompression
by Aardwolf on Thursday June 01, 2023 @04:46AM
(#36148799)
I did mean have the feature removed, I edited the comment (replaced "downgraded" with "have this feature removed". I actually originally typed "crippled" but wanted to phrase it more mildly). What I mean is to avoid the heterogeneous cores they chose to disable it in the performance cores rather than e.g. emulate or double pump it in the efficiency cores.
And they chose to give you efficiency cores + performance cores without avx512 on all consumer CPU's, rather than, what would be my preference, sell some with only performance cores with avx512 like their previous gen CPUs. I'd be ok with efficiency cores on a laptop, sure, but not on a desktop machine.
But AMD already gives plenty fast cores with avx512 now so I'm satisfied, I'm simply not an intel user anymore. Of course intel's choice still affects me anyway since the less consumers have avx512, the less software will make use of it.
Re:Iguana: fast SIMD-optimized decompression
by pjc50 on Thursday June 01, 2023 @06:25AM
(#36149343)
So .. what's the vector set that people should be targeting for wide compatibility these days?
Re:Iguana: fast SIMD-optimized decompression
by powturbo on Thursday June 01, 2023 @06:32AM
(#36149389)
Ideally, dynamic detecting cpu isa (sse/avx2/avx512) and respectively dispatching, used in most of my projects for ex. [1] and [2]
Re:Iguana: fast SIMD-optimized decompression
by adgjlsfhk1 on Thursday June 01, 2023 @12:18PM
(#36153362)
honestly you can probably stop supporting pre avx2 for a lot of use cases. avx2 is 2014 so those machines are pretty much at end of life. this also has the advantage that it lets you assume native fma hardware
Re:Iguana: fast SIMD-optimized decompression
by duped on Thursday June 01, 2023 @10:28AM
(#36151514)
You can use function multiversioning (1) to provide multiple implementations of the same function for different targets that is selected at runtime.
(1) https://releases.llvm.org/7.0.0/tools/clang/docs/AttributeRe...
The modern AMD Zen 4 CPU's for consumers do support avx512 (including avx512_vbmi2).