SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2026 FPrate: PGO+LTO native -Ofast - Released GCCs vs Trunk

-Ofast -flto -march=native -mtune=native -fprofile-use

Please note that event though the header of this comparison says -Ofast, benchmark 767.nest_r was additionally compiled with -fno-finite-math-only. Furthermore, benchmarks 737.gmsh_r and 748.flightdm_r were compiled with -fno-fast-math, so a big chunk of -Ofast was in fact disabled. (This might put GCC at a disadvantage compared to LLVM-based compilers because the latter defaults to -fno-math-errno and -fno-trapping-math, in future measurements, we will try to enable these flags.)

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
709.cactus_r 100.00% 98.86% 105.32% 100.00% 106.54%
722.palm_r 100.00% 104.01% 96.85% 100.00% 93.11%
731.astcenc_r 100.00% 93.60% 93.07% 100.00% 99.43%
736.ocio_r 100.00% 98.33% 95.56% 100.00% 97.18%
737.gmsh_r 100.00% 100.00% 100.91% 100.00% 100.91%
748.flightdm_r 100.00% 101.12% 101.12% 100.00% 100.00%
749.fotonik3d_r 100.00% 100.70% 100.00% 100.00% 99.30%
765.roms_r 100.00% 99.56% 99.56% 100.00% 100.00%
766.femflow_r 100.00% 88.68% 88.68% 100.00% 100.00%
767.nest_r 100.00% 101.12% 101.12% 100.00% 100.00%
772.marian_r 100.00% 118.18% 100.00% 100.00% 84.62%
782.lbm_r 100.00% 98.05% 98.05% 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
709.cactus_r 100.00% 101.71% 102.53% 100.00% 100.80%
722.palm_r 100.00% 102.39% 101.77% 100.00% 99.39%
731.astcenc_r 100.00% 97.31% 93.30% 100.00% 95.87%
736.ocio_r 100.00% 100.70% 100.58% 100.00% 99.88%
737.gmsh_r 100.00% 101.72% 100.96% 100.00% 99.26%
748.flightdm_r 100.00% 101.90% 100.89% 100.00% 99.01%
749.fotonik3d_r 100.00% 104.74% 104.44% 100.00% 99.71%
765.roms_r 100.00% 101.02% 98.90% 100.00% 97.90%
766.femflow_r 100.00% 101.88% 102.38% 100.00% 100.49%
767.nest_r 100.00% 100.61% 101.07% 100.00% 100.46%
772.marian_r 100.00% 102.17% 102.23% 100.00% 100.06%
782.lbm_r 100.00% 100.37% 101.68% 100.00% 101.31%
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
709.cactus_r 100.00% 100.00% 100.57% 100.00% 100.57%
722.palm_r 100.00% 102.58% 95.61% 100.00% 93.20%
731.astcenc_r 100.00% 93.49% 96.88% 100.00% 103.62%
736.ocio_r 100.00% 95.48% 96.01% 100.00% 100.56%
737.gmsh_r 100.00% 101.59% 101.59% 100.00% 100.00%
748.flightdm_r 100.00% 97.06% 97.06% 100.00% 100.00%
749.fotonik3d_r 100.00% 99.52% 100.48% 100.00% 100.96%
765.roms_r 100.00% 100.00% 99.64% 100.00% 99.64%
766.femflow_r 100.00% 93.98% 91.57% 100.00% 97.44%
767.nest_r 100.00% 100.00% 100.00% 100.00% 100.00%
772.marian_r 100.00% 113.33% 103.33% 100.00% 91.18%
782.lbm_r 100.00% 98.00% 98.57% 100.00% 100.58%
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
709.cactus_r 100.00% 101.67% 102.51% 100.00% 100.83%
722.palm_r 100.00% 102.35% 101.97% 100.00% 99.63%
731.astcenc_r 100.00% 104.11% 97.55% 100.00% 93.70%
736.ocio_r 100.00% 100.71% 100.47% 100.00% 99.76%
737.gmsh_r 100.00% 101.80% 101.10% 100.00% 99.31%
748.flightdm_r 100.00% 102.08% 101.01% 100.00% 98.95%
749.fotonik3d_r 100.00% 105.04% 104.86% 100.00% 99.82%
765.roms_r 100.00% 102.21% 100.66% 100.00% 98.48%
766.femflow_r 100.00% 101.95% 102.23% 100.00% 100.27%
767.nest_r 100.00% 100.75% 101.21% 100.00% 100.46%
772.marian_r 100.00% 102.15% 102.26% 100.00% 100.10%
782.lbm_r 100.00% 101.36% 101.44% 100.00% 100.08%
Smaller binary sizes are always better.