SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 FPrate: PGO+LTO -O2 generic - LLVM and GCC

-O2 -flto and -fprofile-instr-use vs. -profile-use

Please note that the same "optimization levels" mean different things to different compilers. For example, LLVM and LLVM-based compiler tend to be much more aggressive at -O2, often with the corresponding consequences to size of the generated binaries and compilation time. GCC is sometimes more aggressive at -O3. LLVM is also often by default less strict when it comes to standard compliance by default and performs optimizations GCC would only do with a special compilation option or at -Ofast. Keep this is mind when comparing different compilers.

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark LLVM GCC
503.bwaves_r CE
507.cactuBSSN_r CE
508.namd_r 100.00% 115.46%
510.parest_r 100.00% 96.45%
511.povray_r 100.00% 95.01%
519.lbm_r 100.00% 102.93%
521.wrf_r CE
526.blender_r 100.00% 108.77%
527.cam4_r CE
538.imagick_r 100.00% 181.02%
544.nab_r 100.00% 95.19%
549.fotonik3d_r CE
554.roms_r CE
LLVM clang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc)
GCC GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.

Non-debug executable sizes

Benchmark LLVM GCC
503.bwaves_r CE
507.cactuBSSN_r CE
508.namd_r 100.00% 100.01%
510.parest_r 100.00% 76.47%
511.povray_r 100.00% 88.14%
519.lbm_r 100.00% 95.14%
521.wrf_r CE
526.blender_r 100.00% 74.32%
527.cam4_r CE
538.imagick_r 100.00% 84.30%
544.nab_r 100.00% 99.00%
549.fotonik3d_r CE
554.roms_r CE
Smaller binary sizes are always better.

AMD EPYC 75F3 (zen3 based high-frequency Milan)

Run-times

Benchmark LLVM GCC
503.bwaves_r CE
507.cactuBSSN_r CE
508.namd_r 100.00% 101.23%
510.parest_r 100.00% 100.75%
511.povray_r 100.00% 96.98%
519.lbm_r 100.00% 101.92%
521.wrf_r CE
526.blender_r 100.00% 105.66%
527.cam4_r CE
538.imagick_r 100.00% 173.94%
544.nab_r 100.00% 99.34%
549.fotonik3d_r CE
554.roms_r CE
LLVM clang version 23.1.3 (https://github.com/llvm/llvm-project.git 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (https://github.com/llvm/llvm-project.git 0d261d1ca552c95a8f007e061c787ac7132fbcbc)
GCC GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.

Non-debug executable sizes

Benchmark LLVM GCC
503.bwaves_r CE
507.cactuBSSN_r CE
508.namd_r 100.00% 100.01%
510.parest_r 100.00% 77.12%
511.povray_r 100.00% 88.15%
519.lbm_r 100.00% 95.42%
521.wrf_r CE
526.blender_r 100.00% 74.32%
527.cam4_r CE
538.imagick_r 100.00% 84.30%
544.nab_r 100.00% 99.03%
549.fotonik3d_r CE
554.roms_r CE
Smaller binary sizes are always better.