SPEC CPU result comparison

15th October 2026 - Normalized report

SPECFP 2006: LTO -Ofast native - LLVM, AOCC and GCC

-Ofast -march=native -flto

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark LLVM AOCC GCC AOCC GCC
410.bwaves 100.00% 97.98% 83.06% 100.00% 84.77%
416.gamess 100.00% 94.13% 108.48% 100.00% 115.24%
433.milc 100.00% 86.64% 84.93% 100.00% 98.02%
434.zeusmp 100.00% 121.09% 112.50% 100.00% 92.90%
435.gromacs 100.00% CE 87.99% CE
436.cactusADM 100.00% CE 89.73% CE
437.leslie3d 100.00% 106.38% 103.36% 100.00% 97.16%
444.namd 100.00% 108.42% 112.50% 100.00% 103.76%
447.dealII 100.00% 96.77% 97.85% 100.00% 101.11%
450.soplex 100.00% 97.70% 103.32% 100.00% 105.74%
453.povray RE RE 100.00% RE
454.calculix 100.00% CE 70.46% CE
459.GemsFDTD 100.00% 99.22% 102.86% 100.00% 103.67%
465.tonto 100.00% 142.69% 65.04% 100.00% 45.58%
470.lbm 100.00% 50.27% 54.08% 100.00% 107.57%
481.wrf 100.00% 82.49% 64.69% 100.00% 78.42%
482.sphinx3 100.00% 93.87% 101.61% 100.00% 108.25%
LLVM clang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc)
AOCC AMD clang version 17.0.6 (CLANG: AOCC_5.2.0-Build#2035 2026_04_10)
GCC GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.

Non-debug executable sizes

Benchmark LLVM AOCC GCC AOCC GCC
410.bwaves 100.00% 5.50% 4.30% 100.00% 78.18%
416.gamess 100.00% 100.55% 99.85% 100.00% 99.30%
433.milc 100.00% 100.06% 80.07% 100.00% 80.02%
434.zeusmp 100.00% 124.74% 124.73% 100.00% 99.99%
435.gromacs 100.00% CE 87.69% CE
436.cactusADM 100.00% CE 41.01% CE
437.leslie3d 100.00% 23.07% 15.64% 100.00% 67.78%
444.namd 100.00% 73.17% 142.47% 100.00% 194.71%
447.dealII 100.00% 89.99% 88.07% 100.00% 97.86%
450.soplex 100.00% 82.58% 90.39% 100.00% 109.46%
453.povray RE RE 100.00% RE
454.calculix 100.00% CE 60.89% CE
459.GemsFDTD 100.00% 35.98% 31.61% 100.00% 87.86%
465.tonto 100.00% 65.74% 44.34% 100.00% 67.45%
470.lbm 100.00% 135.06% 52.29% 100.00% 38.72%
481.wrf 100.00% 81.49% 84.48% 100.00% 103.67%
482.sphinx3 100.00% 103.92% 86.42% 100.00% 83.16%
Smaller binary sizes are always better.