SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 INTrate: LTO -O2 generic - LLVM and GCC

-O2 -flto

Please note that the same "optimization levels" mean different things to different compilers. For example, LLVM and LLVM-based compiler tend to be much more aggressive at -O2, often with the corresponding consequences to size of the generated binaries and compilation time. GCC is sometimes more aggressive at -O3. LLVM is also often by default less strict when it comes to standard compliance by default and performs optimizations GCC would only do with a special compilation option or at -Ofast. Keep this is mind when comparing different compilers.

AMD EPYC 8635P (zen5 based Sorano)

Run-times

Benchmark LLVM GCC
500.perlbench_r 100.00% 94.14%
502.gcc_r 100.00% 100.85%
505.mcf_r 100.00% 104.99%
520.omnetpp_r 100.00% 98.52%
523.xalancbmk_r 100.00% 96.78%
525.x264_r 100.00% 141.21%
531.deepsjeng_r 100.00% 97.65%
541.leela_r 100.00% 108.76%
548.exchange2_r 100.00% 160.85%
557.xz_r 100.00% 101.49%
LLVM clang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc)
GCC GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.

Non-debug executable sizes

Benchmark LLVM GCC
500.perlbench_r 100.00% 100.54%
502.gcc_r 100.00% 93.36%
505.mcf_r 100.00% 72.02%
520.omnetpp_r 100.00% 95.39%
523.xalancbmk_r 100.00% 99.54%
525.x264_r 100.00% 76.46%
531.deepsjeng_r 100.00% 102.50%
541.leela_r 100.00% 110.34%
548.exchange2_r 100.00% 6.34%
557.xz_r 100.00% 106.94%
Smaller binary sizes are always better.

AMD EPYC 75F3 (zen3 based high-frequency Milan)

Run-times

Benchmark LLVM GCC
500.perlbench_r 100.00% 90.21%
502.gcc_r 100.00% 100.00%
505.mcf_r 100.00% 109.38%
520.omnetpp_r 100.00% 95.77%
523.xalancbmk_r 100.00% 100.00%
525.x264_r 100.00% 142.29%
531.deepsjeng_r 100.00% 95.37%
541.leela_r 100.00% 110.09%
548.exchange2_r 100.00% 174.50%
557.xz_r 100.00% 102.58%
LLVM clang version 23.1.3 (https://github.com/llvm/llvm-project.git 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (https://github.com/llvm/llvm-project.git 0d261d1ca552c95a8f007e061c787ac7132fbcbc)
GCC GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d)
Naturally, the smaller the run-times, the better.

Non-debug executable sizes

Benchmark LLVM GCC
500.perlbench_r 100.00% 100.55%
502.gcc_r 100.00% 93.37%
505.mcf_r 100.00% 72.10%
520.omnetpp_r 100.00% 95.38%
523.xalancbmk_r 100.00% 99.54%
525.x264_r 100.00% 76.48%
531.deepsjeng_r 100.00% 102.50%
541.leela_r 100.00% 110.32%
548.exchange2_r 100.00% 6.34%
557.xz_r 100.00% 106.95%
Smaller binary sizes are always better.