GCC 6.2/7.0 vs. LLVM Clang 3.9/4.0 SVN Compiler Performance

[phoronix]

Phoronix: GCC 6.2/7.0 vs. LLVM Clang 3.9/4.0 SVN Compiler Performance

Earlier this week I published some GCC 5.4 vs. GCC 6.2 vs. GCC 7.0 SVN development benchmarks with a Core i7 6800K Broadwell-E system. For those curious how the LLVM Clang compiler stack is comparing, here are some tests on the same system when running fresh benchmarks of LLVM Clang 3.9 as well as LLVM Clang 4.0 SVN.

http://www.phoronix.com/vr.php?view=23852

[davidbepo]

hope they fix clang 4.0 regressions

[Marc Driftmeyer]

Should test it against the fresh LLVM-3.9.1-rc3 finally with POLARIS OpenCL support recognized in Mesa.

[SaucyJack]

Llvm 4.0 is looking like a big bag of ass

[andre30correia]

now the BSD distro change to llvm they have a worst performance c compiler lol

[slacka]

Llvm 4.0 is looking like a big bag of ass

I see you didn't make it to the second page. From those tests, GCC is a big bag of ass. You can't judge a compiler by a couple of benchmarks and you can't fault the SVN development branch for performance regressions. If developers are too afraid to break things, no progress will be made. If they haven't fixed the regressions by the 4.0 release, then you have cause to bitch. In the mean time. STFU. Your insults on OSS add no value here.

[slacka]

now the BSD distro change to llvm they have a worst performance c compiler lol

What's up with all the tolls today that couldn't make it past the first page to post. GCC looks like shit if you just judged it by the second page of benchmarks .
ebizzy: clang 420K, gcc 186K
MAFFT: clang 4.57, gcc 3.8
SciMark: clang 2193, gcc 1596
SciMark SMM: clang 2865, gcc 2562

Looks like BSD made a fine choice to me.

[starshipeleven]

I see you didn't make it to the second page.

Actually, in second page Clang is still a big bag of ass. It's in third page where you start getting benches where it is equal or even better than gcc. Oh, and in last page it is again a big bag of ass.

Now, it is still under development so I'm not going to state Clang 4 will suck, but please get your facts straight.

Looks like BSD made a fine choice to me.

BSD made a political choice, they didn't choose clang because it was better but because it was the only modern compiler they could use in their GPL-phobic environment.

So yes, for them it is a fine choice regardless of performance. They couldn't stay indefinitely on an ancient gcc.

[labyrinth153]

Which if it is more valuable to them just leaves them to improve its performance? I am seeing the same names and same tired arguments.

[jacob]

Hi Michael, would it be possible for you to run the same comparison on ARMv8?