Other Linux Distributions Begin Analyzing Clear Linux's Performance Optimizations

[peppercats]

How so? I mean I know intel are idiots in various ways (I mean did they really give money to the feminist frequency? One of the most toxic things in the history of the internet...?), but how do they cheat benches? I know there are many unfair benchmarking softwares out there more tailored to one vendors products than anothers, is that what you mean? That's not exactly cheating.

benchmarks with AMDcripple
I will never buy an Intel product in my life if I can help it, the company is very scummy.

[SaucyJack]

How so? I mean I know intel are idiots in various ways (I mean did they really give money to the feminist frequency? One of the most toxic things in the history of the internet...?), but how do they cheat benches? I know there are many unfair benchmarking softwares out there more tailored to one vendors products than anothers, is that what you mean? That's not exactly cheating.

...Are you trolling for Intel here or are you actually that clueless? In fact, the lawsuit over the big one that killed AMD financially recently ended (Pentium 4). As I'm sure you've seen what a big difference optimizations can make just from phoronix here, benchmarks were being compiled with an AMD hostile compiler scamming people into thinking Intel had a vastly superior CPU when in fact it was inferior in every way.

[duby229]

...Are you trolling for Intel here or are you actually that clueless? In fact, the lawsuit over the big one that killed AMD financially recently ended (Pentium 4). As I'm sure you've seen what a big difference optimizations can make just from phoronix here, benchmarks were being compiled with an AMD hostile compiler scamming people into thinking Intel had a vastly superior CPU when in fact it was inferior in every way.

But you know that's not even half of it. Intel's CPU's are pretty good. I wouldn't exactly call them great though. Everybody thinks that because they have the performance crown, that must mean their architecture is great, but it's just not true. Believe me their architecture has some serious scalability issues.

EDIT: If Intel was AMD's size there is no way in hell their architecture could achieve the performance they've managed to get from it.

EDIT: I think GF was the worst thing AMD ever did and I think it's Intel's fault partly. (Dirk Meyers, great CPU designer, horrible CEO)

[jrch2k8]

For the conspirationist theorists this happenend on Windows only for software compiled for certain versions of ICC(not VS or GCC just ICC) aka Intel C/C++ Compiler and some benchmarks for Windows recently on discussions like PcMark(I think) that nobody on its serious mind will take seriously this days, since its been ages since everybody knows they fix the tests for money.

On Linux this doesn't happen with anyone since the toolchain is Open aka GCC/Clang and this has nothing to do with Clear linux speedups either, if you ever used Gentoo you will know this by default and get why is faster on certain task when optimized.

The support problem is simple, there are very specific architecture specific optimization (Like AVX2.0 outside steamroller or FMA4 that exist only on AMD, to name few) that force processor family specific binaries, that is why no general use distro use them and compile for generic targets only with generic safe only flags (like -O2 and SSE2), so it works acceptably fast on any CPU

[arjan_intel]

The support problem is simple, there are very specific architecture specific optimization (Like AVX2.0 outside steamroller or FMA4 that exist only on AMD, to name few) that force processor family specific binaries, that is why no general use distro use them and compile for generic targets only with generic safe only flags (like -O2 and SSE2), so it works acceptably fast on any CPU

This is EXACTLY why in Clear Linux we've worked on FMV (Function Multi Versioning, where you can mark specific hot functions for the compiler to create multiple copies for different instruction sets), on making sure /usr/lib64/avx2 works (where we put specific avx2 optimized libraries that are used on AVX2 systems, while the /usr/lib64 lib is used on all other systems) etc etc...

Yes you can just give up and throw your hands in the air (and conspiracy theories are fun) but we tackled the problem (and not by just picking AVX2 as the minimum CPU generation, that would be great for performance but not practical to other distros etc)

[rabcor]

...Are you trolling for Intel here or are you actually that clueless? In fact, the lawsuit over the big one that killed AMD financially recently ended (Pentium 4). As I'm sure you've seen what a big difference optimizations can make just from phoronix here, benchmarks were being compiled with an AMD hostile compiler scamming people into thinking Intel had a vastly superior CPU when in fact it was inferior in every way.

Actually clueless. I don't care much for intel, mostly cus unfair prices. I don't use AMD cus their cpus are weaker (even if intel has that amdcripple thing in their compiler), I'm one of the people waiting for zen in hopes of fair pricing unlike that intel overpriced shit.

[carewolf]

For the conspirationist theorists this happenend on Windows only for software compiled for certain versions of ICC(not VS or GCC just ICC) aka Intel C/C++ Compiler and some benchmarks for Windows recently on discussions like PcMark(I think) that nobody on its serious mind will take seriously this days, since its been ages since everybody knows they fix the tests for money.

On Linux this doesn't happen with anyone since the toolchain is Open aka GCC/Clang and this has nothing to do with Clear linux speedups either, if you ever used Gentoo you will know this by default and get why is faster on certain task when optimized.

It also happens on Linux if you use ICC. I used to contribute patches to open source projects to make the build with ICC, but stopped when Intel 10 years ago added the AMD cripple code (I happened to have an AMD at the time).

[haplo602]

This is EXACTLY why in Clear Linux we've worked on FMV (Function Multi Versioning, where you can mark specific hot functions for the compiler to create multiple copies for different instruction sets), on making sure /usr/lib64/avx2 works (where we put specific avx2 optimized libraries that are used on AVX2 systems, while the /usr/lib64 lib is used on all other systems) etc etc...

Yes you can just give up and throw your hands in the air (and conspiracy theories are fun) but we tackled the problem (and not by just picking AVX2 as the minimum CPU generation, that would be great for performance but not practical to other distros etc)

Sorry ... WHAT ? You mean bloat for binary distros ??? Why bother when this can be done properly on the package level with compiler/flag profiles ? Yes it takes up space for the distro servers to keep all the package versions, but it is still the better solution. Yours is a half of a solution ... Arguably better than nothing, but still sub-optimal.

[jrch2k8]

This is EXACTLY why in Clear Linux we've worked on FMV (Function Multi Versioning, where you can mark specific hot functions for the compiler to create multiple copies for different instruction sets), on making sure /usr/lib64/avx2 works (where we put specific avx2 optimized libraries that are used on AVX2 systems, while the /usr/lib64 lib is used on all other systems) etc etc...

Yes you can just give up and throw your hands in the air (and conspiracy theories are fun) but we tackled the problem (and not by just picking AVX2 as the minimum CPU generation, that would be great for performance but not practical to other distros etc)

I would love that FAT elf project get resucitated, it would be a great addition for FMV

[fabiohl]

This. And the guy that analyzed the code confirmed that they did comment out code rather than fix things. GG Intel.

I think comment unused code is a also a validade optimization, among others of couse.