SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 INTrate: 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
500.perlbench_r 100.00% 101.32% 99.34% 100.00% 98.05%
502.gcc_r 100.00% 99.14% 99.14% 100.00% 100.00%
505.mcf_r 100.00% 100.56% 101.68% 100.00% 101.11%
520.omnetpp_r 100.00% 100.00% 97.81% 100.00% 97.81%
523.xalancbmk_r 100.00% 98.97% 98.97% 100.00% 100.00%
525.x264_r 100.00% 94.72% 92.08% 100.00% 97.21%
531.deepsjeng_r 100.00% 98.58% 98.02% 100.00% 99.43%
541.leela_r 100.00% 99.29% 99.29% 100.00% 100.00%
548.exchange2_r 100.00% 98.52% 100.00% 100.00% 101.50%
557.xz_r 100.00% 100.85% 100.85% 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
500.perlbench_r 100.00% 100.90% 100.63% 100.00% 99.73%
502.gcc_r 100.00% 103.27% 102.80% 100.00% 99.55%
505.mcf_r 100.00% 99.18% 97.25% 100.00% 98.06%
520.omnetpp_r 100.00% 101.83% 100.77% 100.00% 98.95%
523.xalancbmk_r 100.00% 107.92% 106.60% 100.00% 98.78%
525.x264_r 100.00% 121.38% 115.43% 100.00% 95.10%
531.deepsjeng_r 100.00% 100.03% 100.05% 100.00% 100.02%
541.leela_r 100.00% 99.35% 101.58% 100.00% 102.24%
548.exchange2_r 100.00% 91.76% 93.23% 100.00% 101.60%
557.xz_r 100.00% 99.33% 99.36% 100.00% 100.04%
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
500.perlbench_r 100.00% 97.32% 98.66% 100.00% 101.38%
502.gcc_r 100.00% 100.00% 100.00% 100.00% 100.00%
505.mcf_r 100.00% 102.26% 101.69% 100.00% 99.45%
520.omnetpp_r 100.00% 97.74% 96.16% 100.00% 98.38%
523.xalancbmk_r 100.00% 97.06% 97.06% 100.00% 100.00%
525.x264_r 100.00% 95.00% 91.33% 100.00% 96.14%
531.deepsjeng_r 100.00% 98.75% 98.75% 100.00% 100.00%
541.leela_r 100.00% 95.55% 96.21% 100.00% 100.70%
548.exchange2_r 100.00% 96.55% 96.55% 100.00% 100.00%
557.xz_r 100.00% 102.52% 101.26% 100.00% 98.77%
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
500.perlbench_r 100.00% 101.26% 100.91% 100.00% 99.66%
502.gcc_r 100.00% 103.78% 103.27% 100.00% 99.50%
505.mcf_r 100.00% 97.65% 97.51% 100.00% 99.86%
520.omnetpp_r 100.00% 102.05% 101.03% 100.00% 99.00%
523.xalancbmk_r 100.00% 109.54% 108.12% 100.00% 98.71%
525.x264_r 100.00% 126.87% 119.82% 100.00% 94.44%
531.deepsjeng_r 100.00% 100.02% 100.03% 100.00% 100.01%
541.leela_r 100.00% 100.01% 102.59% 100.00% 102.57%
548.exchange2_r 100.00% 105.02% 106.21% 100.00% 101.13%
557.xz_r 100.00% 100.46% 100.07% 100.00% 99.62%
Smaller binary sizes are always better.