SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 FPrate: LTO generic -O2 - Released GCCs vs Trunk

-O2 -flto

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
503.bwaves_r 100.00% 100.50% 100.00% 100.00% 99.50%
507.cactuBSSN_r 100.00% 100.00% 100.88% 100.00% 100.88%
508.namd_r 100.00% 99.19% 98.38% 100.00% 99.18%
510.parest_r 100.00% 99.46% 99.46% 100.00% 100.00%
511.povray_r 100.00% 94.64% 96.65% 100.00% 102.12%
519.lbm_r 100.00% 101.08% 101.08% 100.00% 100.00%
521.wrf_r 100.00% 100.60% 102.08% 100.00% 101.48%
526.blender_r 100.00% 98.66% 97.99% 100.00% 99.32%
527.cam4_r 100.00% 100.89% 98.44% 100.00% 97.57%
538.imagick_r 100.00% 100.00% 100.00% 100.00% 100.00%
544.nab_r 100.00% 100.00% 89.27% 100.00% 89.27%
549.fotonik3d_r 100.00% 98.94% 99.58% 100.00% 100.64%
554.roms_r 100.00% 94.78% 94.28% 100.00% 99.48%
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% 108.65% 110.28% 100.00% 101.50%
507.cactuBSSN_r 100.00% 104.22% 101.12% 100.00% 97.03%
508.namd_r 100.00% 101.36% 101.31% 100.00% 99.95%
510.parest_r 100.00% 103.80% 103.03% 100.00% 99.26%
511.povray_r 100.00% 101.59% 101.80% 100.00% 100.21%
519.lbm_r 100.00% 100.00% 99.76% 100.00% 99.76%
521.wrf_r 100.00% 100.74% 99.81% 100.00% 99.08%
526.blender_r 100.00% 101.79% 101.67% 100.00% 99.88%
527.cam4_r 100.00% 100.24% 100.22% 100.00% 99.98%
538.imagick_r 100.00% 104.40% 104.10% 100.00% 99.71%
544.nab_r 100.00% 100.60% 100.84% 100.00% 100.23%
549.fotonik3d_r 100.00% 102.08% 99.47% 100.00% 97.44%
554.roms_r 100.00% 101.86% 101.58% 100.00% 99.73%
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% 99.50% 99.50% 100.00% 100.00%
507.cactuBSSN_r 100.00% 100.00% 101.20% 100.00% 101.20%
508.namd_r 100.00% 100.00% 98.78% 100.00% 98.78%
510.parest_r 100.00% 98.60% 99.30% 100.00% 100.71%
511.povray_r 100.00% 94.90% 97.35% 100.00% 102.58%
519.lbm_r 100.00% 102.17% 102.17% 100.00% 100.00%
521.wrf_r 100.00% 100.83% 102.48% 100.00% 101.64%
526.blender_r 100.00% 97.98% 97.37% 100.00% 99.38%
527.cam4_r 100.00% 101.23% 98.98% 100.00% 97.78%
538.imagick_r 100.00% 100.00% 101.26% 100.00% 101.26%
544.nab_r 100.00% 100.40% 93.94% 100.00% 93.56%
549.fotonik3d_r 100.00% 99.13% 99.13% 100.00% 100.00%
554.roms_r 100.00% 92.53% 92.53% 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% 108.63% 110.26% 100.00% 101.50%
507.cactuBSSN_r 100.00% 104.22% 101.12% 100.00% 97.03%
508.namd_r 100.00% 101.36% 101.32% 100.00% 99.96%
510.parest_r 100.00% 103.80% 103.03% 100.00% 99.26%
511.povray_r 100.00% 101.59% 101.80% 100.00% 100.21%
519.lbm_r 100.00% 100.00% 99.76% 100.00% 99.76%
521.wrf_r 100.00% 100.74% 99.81% 100.00% 99.08%
526.blender_r 100.00% 101.79% 101.67% 100.00% 99.88%
527.cam4_r 100.00% 100.24% 100.22% 100.00% 99.98%
538.imagick_r 100.00% 104.40% 104.11% 100.00% 99.71%
544.nab_r 100.00% 100.58% 100.82% 100.00% 100.23%
549.fotonik3d_r 100.00% 102.08% 99.47% 100.00% 97.44%
554.roms_r 100.00% 101.86% 101.58% 100.00% 99.73%
Smaller binary sizes are always better.