SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 FPrate: LTO native -Ofast - Released GCCs vs Trunk

-Ofast -flto -march=native -mtune=native

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
503.bwaves_r 100.00% 100.00% 100.00% 100.00% 100.00%
507.cactuBSSN_r 100.00% 100.00% 100.00% 100.00% 100.00%
508.namd_r 100.00% 100.00% 99.00% 100.00% 99.00%
510.parest_r 100.00% 93.90% 93.90% 100.00% 100.00%
511.povray_r 100.00% 103.58% 102.30% 100.00% 98.77%
519.lbm_r 100.00% 100.00% 100.00% 100.00% 100.00%
521.wrf_r 100.00% 99.13% 100.00% 100.00% 100.87%
526.blender_r 100.00% 97.60% 100.72% 100.00% 103.19%
527.cam4_r 100.00% 98.90% 98.90% 100.00% 100.00%
538.imagick_r 100.00% 96.49% 95.32% 100.00% 98.79%
544.nab_r 100.00% 100.00% 87.17% 100.00% 87.17%
549.fotonik3d_r 100.00% 99.56% 99.56% 100.00% 100.00%
554.roms_r 100.00% 99.00% 99.00% 100.00% 100.00%
GCC 15 GCC 15.3.0 (revision 4db0e8df15bef836558857c291c323add11d035c)
GCC 16 GCC 16.2.0 (revision 78d4ac73dd391005b895a6148cd9831e28e1208b)
Trunk GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.
See also continuous gcc branch benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/branch.

Non-debug executable sizes

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
503.bwaves_r 100.00% 107.85% 108.72% 100.00% 100.81%
507.cactuBSSN_r 100.00% 100.61% 98.17% 100.00% 97.57%
508.namd_r 100.00% 103.67% 104.76% 100.00% 101.05%
510.parest_r 100.00% 108.57% 107.76% 100.00% 99.26%
511.povray_r 100.00% 103.16% 100.72% 100.00% 97.63%
519.lbm_r 100.00% 107.17% 104.01% 100.00% 97.05%
521.wrf_r 100.00% 97.23% 97.24% 100.00% 100.01%
526.blender_r 100.00% 96.64% 95.83% 100.00% 99.17%
527.cam4_r 100.00% 99.77% 99.73% 100.00% 99.96%
538.imagick_r 100.00% 102.01% 96.43% 100.00% 94.53%
544.nab_r 100.00% 100.26% 100.15% 100.00% 99.88%
549.fotonik3d_r 100.00% 101.03% 96.80% 100.00% 95.81%
554.roms_r 100.00% 99.49% 92.93% 100.00% 93.40%
Smaller binary sizes are always better.

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

Run-times

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
503.bwaves_r 100.00% 99.25% 99.25% 100.00% 100.00%
507.cactuBSSN_r 100.00% 98.90% 100.55% 100.00% 101.66%
508.namd_r 100.00% 100.69% 98.17% 100.00% 97.50%
510.parest_r 100.00% 95.74% 96.90% 100.00% 101.21%
511.povray_r 100.00% 104.39% 102.08% 100.00% 97.79%
519.lbm_r 100.00% 95.26% 92.77% 100.00% 97.38%
521.wrf_r 100.00% 97.68% 98.55% 100.00% 100.89%
526.blender_r 100.00% 100.00% 106.45% 100.00% 106.45%
527.cam4_r 100.00% 100.00% 100.00% 100.00% 100.00%
538.imagick_r 100.00% 92.80% 91.53% 100.00% 98.63%
544.nab_r 100.00% 99.29% 88.84% 100.00% 89.47%
549.fotonik3d_r 100.00% 100.00% 100.47% 100.00% 100.47%
554.roms_r 100.00% 97.89% 97.89% 100.00% 100.00%
GCC 15 GCC 15.3.0 (revision 4db0e8df15bef836558857c291c323add11d035c)
GCC 16 GCC 16.2.0 (revision 78d4ac73dd391005b895a6148cd9831e28e1208b)
Trunk GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.
See also continuous gcc branch benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/branch.

Non-debug executable sizes

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
503.bwaves_r 100.00% 116.98% 119.41% 100.00% 102.07%
507.cactuBSSN_r 100.00% 101.30% 98.77% 100.00% 97.50%
508.namd_r 100.00% 105.34% 104.56% 100.00% 99.26%
510.parest_r 100.00% 108.69% 107.55% 100.00% 98.95%
511.povray_r 100.00% 104.10% 101.53% 100.00% 97.53%
519.lbm_r 100.00% 109.33% 105.21% 100.00% 96.23%
521.wrf_r 100.00% 100.19% 100.27% 100.00% 100.09%
526.blender_r 100.00% 99.48% 98.89% 100.00% 99.41%
527.cam4_r 100.00% 100.18% 100.17% 100.00% 99.99%
538.imagick_r 100.00% 102.36% 101.24% 100.00% 98.90%
544.nab_r 100.00% 100.12% 100.24% 100.00% 100.12%
549.fotonik3d_r 100.00% 102.78% 101.98% 100.00% 99.22%
554.roms_r 100.00% 103.50% 98.55% 100.00% 95.22%
Smaller binary sizes are always better.