SPEC CPU result comparison

15th October 2026 - Normalized report

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

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

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% 101.92% 110.38% 100.00% 108.30%
722.palm_r 100.00% 101.92% 95.62% 100.00% 93.82%
731.astcenc_r 100.00% 96.77% 82.53% 100.00% 85.28%
736.ocio_r 100.00% 98.35% 95.59% 100.00% 97.20%
737.gmsh_r 100.00% 100.00% 100.00% 100.00% 100.00%
748.flightdm_r 100.00% 100.00% 100.00% 100.00% 100.00%
749.fotonik3d_r 100.00% 101.31% 100.00% 100.00% 98.71%
765.roms_r 100.00% 101.79% 99.10% 100.00% 97.36%
766.femflow_r 100.00% 90.38% 90.38% 100.00% 100.00%
767.nest_r 100.00% 101.99% 101.99% 100.00% 100.00%
772.marian_r 100.00% 100.00% 75.00% 100.00% 75.00%
782.lbm_r 100.00% 98.54% 100.00% 100.00% 101.48%
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
709.cactus_r 100.00% 100.20% 98.71% 100.00% 98.52%
722.palm_r 100.00% 99.94% 97.70% 100.00% 97.76%
731.astcenc_r 100.00% 117.76% 98.90% 100.00% 83.98%
736.ocio_r 100.00% 100.25% 99.92% 100.00% 99.67%
737.gmsh_r 100.00% 101.72% 99.75% 100.00% 98.06%
748.flightdm_r 100.00% 101.31% 99.16% 100.00% 97.88%
749.fotonik3d_r 100.00% 101.97% 95.34% 100.00% 93.50%
765.roms_r 100.00% 98.58% 94.13% 100.00% 95.49%
766.femflow_r 100.00% 103.57% 102.40% 100.00% 98.87%
767.nest_r 100.00% 100.73% 99.95% 100.00% 99.22%
772.marian_r 100.00% 100.62% 100.46% 100.00% 99.84%
782.lbm_r 100.00% 107.29% 104.63% 100.00% 97.52%
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% 100.28% 100.00% 100.28%
722.palm_r 100.00% 101.75% 95.99% 100.00% 94.33%
731.astcenc_r 100.00% 92.62% 86.26% 100.00% 93.13%
736.ocio_r 100.00% 94.97% 94.97% 100.00% 100.00%
737.gmsh_r 100.00% 100.00% 100.00% 100.00% 100.00%
748.flightdm_r 100.00% 101.92% 98.32% 100.00% 96.47%
749.fotonik3d_r 100.00% 99.23% 100.00% 100.00% 100.78%
765.roms_r 100.00% 106.88% 100.00% 100.00% 93.56%
766.femflow_r 100.00% 92.50% 93.75% 100.00% 101.35%
767.nest_r 100.00% 100.00% 100.00% 100.00% 100.00%
772.marian_r 100.00% 97.06% 85.29% 100.00% 87.88%
782.lbm_r 100.00% 96.07% 94.76% 100.00% 98.64%
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% 102.58% 101.07% 100.00% 98.52%
722.palm_r 100.00% 101.80% 100.81% 100.00% 99.03%
731.astcenc_r 100.00% 125.44% 109.81% 100.00% 87.54%
736.ocio_r 100.00% 101.08% 100.23% 100.00% 99.16%
737.gmsh_r 100.00% 103.13% 101.80% 100.00% 98.72%
748.flightdm_r 100.00% 101.36% 99.23% 100.00% 97.90%
749.fotonik3d_r 100.00% 102.64% 101.82% 100.00% 99.20%
765.roms_r 100.00% 100.89% 98.52% 100.00% 97.65%
766.femflow_r 100.00% 103.84% 102.06% 100.00% 98.29%
767.nest_r 100.00% 101.45% 100.66% 100.00% 99.23%
772.marian_r 100.00% 102.48% 102.18% 100.00% 99.71%
782.lbm_r 100.00% 109.31% 105.19% 100.00% 96.24%
Smaller binary sizes are always better.