SPEC2017 INTrate: Standard -Ofast native - LLVM, ICX and GCC
-Ofast -march=native
AMD EPYC 8635P (zen5 based Sorano)
Run-times
Benchmark
ICX
GCC
500.perlbench_r
100.00%
96.88%
502.gcc_r
100.00%
96.15%
505.mcf_r
100.00%
98.61%
520.omnetpp_r
100.00%
99.31%
523.xalancbmk_r
100.00%
98.06%
525.x264_r
100.00%
101.55%
531.deepsjeng_r
100.00%
100.96%
541.leela_r
100.00%
97.41%
548.exchange2_r
100.00%
79.38%
557.xz_r
100.00%
103.00%
| Benchmark | ICX | GCC |
|---|---|---|
| 500.perlbench_r | 100.00% | 96.88% |
| 502.gcc_r | 100.00% | 96.15% |
| 505.mcf_r | 100.00% | 98.61% |
| 520.omnetpp_r | 100.00% | 99.31% |
| 523.xalancbmk_r | 100.00% | 98.06% |
| 525.x264_r | 100.00% | 101.55% |
| 531.deepsjeng_r | 100.00% | 100.96% |
| 541.leela_r | 100.00% | 97.41% |
| 548.exchange2_r | 100.00% | 79.38% |
| 557.xz_r | 100.00% | 103.00% |
| LLVM | clang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc), flang version 23.1.3 (/labs/mjambor/git/llvm 0d261d1ca552c95a8f007e061c787ac7132fbcbc) |
| ICX | Intel(R) oneAPI DPC++/C++ Compiler 2026.1.1 (2026.1.1.20260724), ifx (IFX) 2026.1.1 20260724 |
| GCC | GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d) |
Non-debug executable sizes
Benchmark
ICX
GCC
500.perlbench_r
100.00%
102.47%
502.gcc_r
100.00%
106.07%
505.mcf_r
100.00%
95.73%
520.omnetpp_r
100.00%
107.09%
523.xalancbmk_r
100.00%
127.25%
525.x264_r
100.00%
76.01%
531.deepsjeng_r
100.00%
99.86%
541.leela_r
100.00%
91.27%
548.exchange2_r
100.00%
23.30%
557.xz_r
100.00%
81.98%
Smaller binary sizes are always better.
| Benchmark | ICX | GCC |
|---|---|---|
| 500.perlbench_r | 100.00% | 102.47% |
| 502.gcc_r | 100.00% | 106.07% |
| 505.mcf_r | 100.00% | 95.73% |
| 520.omnetpp_r | 100.00% | 107.09% |
| 523.xalancbmk_r | 100.00% | 127.25% |
| 525.x264_r | 100.00% | 76.01% |
| 531.deepsjeng_r | 100.00% | 99.86% |
| 541.leela_r | 100.00% | 91.27% |
| 548.exchange2_r | 100.00% | 23.30% |
| 557.xz_r | 100.00% | 81.98% |
AMD EPYC 75F3 (zen3 based high-frequency Milan)
Run-times
Benchmark
LLVM
ICX
GCC
ICX
GCC
500.perlbench_r
100.00%
95.12%
90.65%
100.00%
95.30%
502.gcc_r
100.00%
101.22%
95.11%
100.00%
93.96%
505.mcf_r
100.00%
88.73%
86.72%
100.00%
97.73%
520.omnetpp_r
100.00%
101.26%
97.28%
100.00%
96.07%
523.xalancbmk_r
100.00%
89.92%
89.92%
100.00%
100.00%
525.x264_r
100.00%
89.44%
97.03%
100.00%
108.49%
531.deepsjeng_r
100.00%
96.94%
91.02%
100.00%
93.89%
541.leela_r
100.00%
104.43%
103.16%
100.00%
98.79%
548.exchange2_r
100.00%
165.14%
82.39%
100.00%
49.89%
557.xz_r
100.00%
100.86%
104.29%
100.00%
103.40%
| Benchmark | LLVM | ICX | GCC | ICX | GCC |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 95.12% | 90.65% | 100.00% | 95.30% |
| 502.gcc_r | 100.00% | 101.22% | 95.11% | 100.00% | 93.96% |
| 505.mcf_r | 100.00% | 88.73% | 86.72% | 100.00% | 97.73% |
| 520.omnetpp_r | 100.00% | 101.26% | 97.28% | 100.00% | 96.07% |
| 523.xalancbmk_r | 100.00% | 89.92% | 89.92% | 100.00% | 100.00% |
| 525.x264_r | 100.00% | 89.44% | 97.03% | 100.00% | 108.49% |
| 531.deepsjeng_r | 100.00% | 96.94% | 91.02% | 100.00% | 93.89% |
| 541.leela_r | 100.00% | 104.43% | 103.16% | 100.00% | 98.79% |
| 548.exchange2_r | 100.00% | 165.14% | 82.39% | 100.00% | 49.89% |
| 557.xz_r | 100.00% | 100.86% | 104.29% | 100.00% | 103.40% |
| 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) |
| ICX | Intel(R) oneAPI DPC++/C++ Compiler 2026.1.1 (2026.1.1.20260724), ifx (IFX) 2026.1.1 20260724 |
| GCC | GCC 17.0.0 20261015 (experimental) (revision 84348add236100e5f69a3e9ec4f7b1038bf2b83d) |
Non-debug executable sizes
Benchmark
LLVM
ICX
GCC
ICX
GCC
500.perlbench_r
100.00%
112.11%
115.34%
100.00%
102.87%
502.gcc_r
100.00%
102.66%
108.84%
100.00%
106.03%
505.mcf_r
100.00%
190.32%
90.08%
100.00%
47.33%
520.omnetpp_r
100.00%
104.04%
110.75%
100.00%
106.44%
523.xalancbmk_r
100.00%
98.63%
129.04%
100.00%
130.83%
525.x264_r
100.00%
136.92%
106.22%
100.00%
77.57%
531.deepsjeng_r
100.00%
100.21%
99.98%
100.00%
99.78%
541.leela_r
100.00%
147.02%
137.14%
100.00%
93.28%
548.exchange2_r
100.00%
91.72%
22.23%
100.00%
24.23%
557.xz_r
100.00%
135.80%
112.86%
100.00%
83.11%
Smaller binary sizes are always better.
| Benchmark | LLVM | ICX | GCC | ICX | GCC |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 112.11% | 115.34% | 100.00% | 102.87% |
| 502.gcc_r | 100.00% | 102.66% | 108.84% | 100.00% | 106.03% |
| 505.mcf_r | 100.00% | 190.32% | 90.08% | 100.00% | 47.33% |
| 520.omnetpp_r | 100.00% | 104.04% | 110.75% | 100.00% | 106.44% |
| 523.xalancbmk_r | 100.00% | 98.63% | 129.04% | 100.00% | 130.83% |
| 525.x264_r | 100.00% | 136.92% | 106.22% | 100.00% | 77.57% |
| 531.deepsjeng_r | 100.00% | 100.21% | 99.98% | 100.00% | 99.78% |
| 541.leela_r | 100.00% | 147.02% | 137.14% | 100.00% | 93.28% |
| 548.exchange2_r | 100.00% | 91.72% | 22.23% | 100.00% | 24.23% |
| 557.xz_r | 100.00% | 135.80% | 112.86% | 100.00% | 83.11% |