SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2017 INTrate: LTO -O3 x86-64-v3 - LLVM and GCC

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

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% 94.21%
502.gcc_r 100.00% 99.14%
505.mcf_r 100.00% 90.77%
520.omnetpp_r 100.00% 98.28%
523.xalancbmk_r 100.00% 84.55%
525.x264_r 100.00% 94.57%
531.deepsjeng_r 100.00% 95.58%
541.leela_r 100.00% 101.70%
548.exchange2_r 100.00% 95.02%
557.xz_r 100.00% 102.32%
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% 115.66%
502.gcc_r 100.00% 101.74%
505.mcf_r 100.00% 92.16%
520.omnetpp_r 100.00% 100.23%
523.xalancbmk_r 100.00% 104.62%
525.x264_r 100.00% 117.48%
531.deepsjeng_r 100.00% 102.45%
541.leela_r 100.00% 123.67%
548.exchange2_r 100.00% 19.98%
557.xz_r 100.00% 113.08%
Smaller binary sizes are always better.

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

Run-times

Benchmark LLVM GCC
500.perlbench_r 100.00% 92.02%
502.gcc_r 100.00% 97.00%
505.mcf_r 100.00% 94.79%
520.omnetpp_r 100.00% 96.68%
523.xalancbmk_r 100.00% 87.72%
525.x264_r 100.00% 89.64%
531.deepsjeng_r 100.00% 94.69%
541.leela_r 100.00% 102.38%
548.exchange2_r 100.00% 90.69%
557.xz_r 100.00% 103.42%
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% 115.66%
502.gcc_r 100.00% 101.74%
505.mcf_r 100.00% 92.21%
520.omnetpp_r 100.00% 100.11%
523.xalancbmk_r 100.00% 104.62%
525.x264_r 100.00% 117.52%
531.deepsjeng_r 100.00% 102.45%
541.leela_r 100.00% 123.71%
548.exchange2_r 100.00% 19.99%
557.xz_r 100.00% 113.12%
Smaller binary sizes are always better.