SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 INTrate: LTO x86-64-v3 -O3 - Released GCCs vs Trunk

-O3 -flto -march=x86-64-v3

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark GCC 15 GCC 16 Trunk GCC 16 Trunk
500.perlbench_r 100.00% 98.04% 99.35% 100.00% 101.33%
502.gcc_r 100.00% 99.14% 99.14% 100.00% 100.00%
505.mcf_r 100.00% 98.90% 100.55% 100.00% 101.68%
520.omnetpp_r 100.00% 99.00% 100.00% 100.00% 101.01%
523.xalancbmk_r 100.00% 96.57% 96.57% 100.00% 100.00%
525.x264_r 100.00% 97.29% 94.57% 100.00% 97.21%
531.deepsjeng_r 100.00% 100.00% 98.02% 100.00% 98.02%
541.leela_r 100.00% 95.81% 97.21% 100.00% 101.46%
548.exchange2_r 100.00% 100.00% 98.52% 100.00% 98.52%
557.xz_r 100.00% 101.49% 103.41% 100.00% 101.89%
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% 101.65% 101.21% 100.00% 99.56%
502.gcc_r 100.00% 103.93% 103.23% 100.00% 99.33%
505.mcf_r 100.00% 97.53% 98.63% 100.00% 101.13%
520.omnetpp_r 100.00% 104.42% 100.87% 100.00% 96.60%
523.xalancbmk_r 100.00% 108.71% 107.44% 100.00% 98.84%
525.x264_r 100.00% 125.66% 119.63% 100.00% 95.20%
531.deepsjeng_r 100.00% 100.02% 100.04% 100.00% 100.02%
541.leela_r 100.00% 97.97% 100.17% 100.00% 102.25%
548.exchange2_r 100.00% 105.60% 107.79% 100.00% 102.08%
557.xz_r 100.00% 102.96% 101.57% 100.00% 98.66%
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% 96.85% 98.65% 100.00% 101.86%
502.gcc_r 100.00% 98.68% 99.34% 100.00% 100.67%
505.mcf_r 100.00% 100.00% 101.11% 100.00% 101.11%
520.omnetpp_r 100.00% 97.48% 100.00% 100.00% 102.58%
523.xalancbmk_r 100.00% 96.32% 96.32% 100.00% 100.00%
525.x264_r 100.00% 96.00% 92.33% 100.00% 96.18%
531.deepsjeng_r 100.00% 100.00% 98.99% 100.00% 98.99%
541.leela_r 100.00% 95.33% 95.56% 100.00% 100.23%
548.exchange2_r 100.00% 99.12% 98.25% 100.00% 99.12%
557.xz_r 100.00% 101.27% 102.11% 100.00% 100.83%
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% 101.66% 101.21% 100.00% 99.56%
502.gcc_r 100.00% 103.93% 103.23% 100.00% 99.33%
505.mcf_r 100.00% 97.53% 98.64% 100.00% 101.13%
520.omnetpp_r 100.00% 104.42% 100.87% 100.00% 96.60%
523.xalancbmk_r 100.00% 108.71% 107.45% 100.00% 98.84%
525.x264_r 100.00% 125.65% 119.62% 100.00% 95.20%
531.deepsjeng_r 100.00% 100.02% 100.04% 100.00% 100.02%
541.leela_r 100.00% 97.93% 100.12% 100.00% 102.25%
548.exchange2_r 100.00% 105.59% 107.79% 100.00% 102.08%
557.xz_r 100.00% 102.95% 101.57% 100.00% 98.66%
Smaller binary sizes are always better.