SPEC CPU result comparison

15th October 2026 - Normalized report

SPEC2026 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
706.stockfish_r 100.00% 59.68%
707.ntest_r 100.00% 106.49%
708.sqlite_r 100.00% 97.14%
710.omnetpp_r 100.00% 90.58%
714.cpython_r 100.00% 96.62%
721.gcc_r 100.00% 100.00%
723.llvm_r 100.00% 108.09%
727.cppcheck_r 100.00% 102.34%
729.abc_r 100.00% 100.00%
734.vpr_r 100.00% 98.81%
735.gem5_r 100.00% 96.71%
750.sealcrypto_r 100.00% 100.00%
753.ns3_r 100.00% 108.31%
777.zstd_r 100.00% 96.24%
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
706.stockfish_r 100.00% 99.79%
707.ntest_r 100.00% 97.76%
708.sqlite_r 100.00% 85.58%
710.omnetpp_r 100.00% 99.13%
714.cpython_r 100.00% 88.18%
721.gcc_r 100.00% 91.83%
723.llvm_r 100.00% 89.43%
727.cppcheck_r 100.00% 89.47%
729.abc_r 100.00% 92.44%
734.vpr_r 100.00% 93.65%
735.gem5_r 100.00% 96.67%
750.sealcrypto_r 100.00% 88.12%
753.ns3_r 100.00% 81.00%
777.zstd_r 100.00% 99.33%
Smaller binary sizes are always better.

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

Run-times

Benchmark LLVM GCC
706.stockfish_r 100.00% 65.69%
707.ntest_r 100.00% 107.80%
708.sqlite_r 100.00% 97.83%
710.omnetpp_r 100.00% 90.24%
714.cpython_r 100.00% 94.80%
721.gcc_r 100.00% 98.52%
723.llvm_r 100.00% 109.01%
727.cppcheck_r 100.00% 105.34%
729.abc_r 100.00% 98.63%
734.vpr_r 100.00% 97.73%
735.gem5_r 100.00% 99.27%
750.sealcrypto_r 100.00% 114.80%
753.ns3_r 100.00% 101.75%
777.zstd_r 100.00% 98.49%
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
706.stockfish_r 100.00% 99.80%
707.ntest_r 100.00% 97.75%
708.sqlite_r 100.00% 85.58%
710.omnetpp_r 100.00% 99.16%
714.cpython_r 100.00% 88.18%
721.gcc_r 100.00% 91.83%
723.llvm_r 100.00% 89.41%
727.cppcheck_r 100.00% 89.63%
729.abc_r 100.00% 92.45%
734.vpr_r 100.00% 93.92%
735.gem5_r 100.00% 96.63%
750.sealcrypto_r 100.00% 86.63%
753.ns3_r 100.00% 79.41%
777.zstd_r 100.00% 99.34%
Smaller binary sizes are always better.