SPEC2017 INTrate: PGO LTO -Ofast - Different x86_64 levels
-flto -Ofast
AMD EPYC 7543P (zen3 based Milan)
Run-times
Benchmark
x86-64
x86-64-v2
x86-64-v3
x86-64-v2
x86-64-v3
500.perlbench_r
100.00%
99.11%
99.11%
100.00%
100.00%
502.gcc_r
100.00%
100.00%
100.00%
100.00%
100.00%
505.mcf_r
100.00%
100.00%
101.18%
100.00%
101.18%
520.omnetpp_r
100.00%
99.49%
98.46%
100.00%
98.97%
523.xalancbmk_r
100.00%
98.88%
98.88%
100.00%
100.00%
525.x264_r
100.00%
95.71%
94.64%
100.00%
98.88%
531.deepsjeng_r
100.00%
95.74%
93.10%
100.00%
97.25%
541.leela_r
100.00%
100.44%
100.00%
100.00%
99.56%
548.exchange2_r
100.00%
100.36%
102.54%
100.00%
102.17%
557.xz_r
100.00%
100.63%
100.42%
100.00%
99.79%
| Benchmark | x86-64 | x86-64-v2 | x86-64-v3 | x86-64-v2 | x86-64-v3 |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 99.11% | 99.11% | 100.00% | 100.00% |
| 502.gcc_r | 100.00% | 100.00% | 100.00% | 100.00% | 100.00% |
| 505.mcf_r | 100.00% | 100.00% | 101.18% | 100.00% | 101.18% |
| 520.omnetpp_r | 100.00% | 99.49% | 98.46% | 100.00% | 98.97% |
| 523.xalancbmk_r | 100.00% | 98.88% | 98.88% | 100.00% | 100.00% |
| 525.x264_r | 100.00% | 95.71% | 94.64% | 100.00% | 98.88% |
| 531.deepsjeng_r | 100.00% | 95.74% | 93.10% | 100.00% | 97.25% |
| 541.leela_r | 100.00% | 100.44% | 100.00% | 100.00% | 99.56% |
| 548.exchange2_r | 100.00% | 100.36% | 102.54% | 100.00% | 102.17% |
| 557.xz_r | 100.00% | 100.63% | 100.42% | 100.00% | 99.79% |
Naturally, the smaller the run-times, the better. On the other hand, rates and especially their geometric means are better when they are greater.
See also continuous gcc tuning benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/tuning.
Non-debug executable sizes
Click to display/hide
Benchmark
x86-64
x86-64-v2
x86-64-v3
x86-64-v2
x86-64-v3
500.perlbench_r
100.00%
99.98%
100.01%
100.00%
100.03%
502.gcc_r
100.00%
99.99%
100.16%
100.00%
100.17%
505.mcf_r
100.00%
101.38%
105.48%
100.00%
104.04%
520.omnetpp_r
100.00%
99.97%
100.80%
100.00%
100.83%
523.xalancbmk_r
100.00%
99.97%
100.73%
100.00%
100.76%
525.x264_r
100.00%
97.47%
95.22%
100.00%
97.69%
531.deepsjeng_r
100.00%
99.99%
99.98%
100.00%
99.98%
541.leela_r
100.00%
100.01%
101.96%
100.00%
101.95%
548.exchange2_r
100.00%
100.00%
103.97%
100.00%
103.97%
557.xz_r
100.00%
100.03%
100.28%
100.00%
100.24%
Smaller binary sizes are always better.
Click to display/hide
| Benchmark | x86-64 | x86-64-v2 | x86-64-v3 | x86-64-v2 | x86-64-v3 |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 99.98% | 100.01% | 100.00% | 100.03% |
| 502.gcc_r | 100.00% | 99.99% | 100.16% | 100.00% | 100.17% |
| 505.mcf_r | 100.00% | 101.38% | 105.48% | 100.00% | 104.04% |
| 520.omnetpp_r | 100.00% | 99.97% | 100.80% | 100.00% | 100.83% |
| 523.xalancbmk_r | 100.00% | 99.97% | 100.73% | 100.00% | 100.76% |
| 525.x264_r | 100.00% | 97.47% | 95.22% | 100.00% | 97.69% |
| 531.deepsjeng_r | 100.00% | 99.99% | 99.98% | 100.00% | 99.98% |
| 541.leela_r | 100.00% | 100.01% | 101.96% | 100.00% | 101.95% |
| 548.exchange2_r | 100.00% | 100.00% | 103.97% | 100.00% | 103.97% |
| 557.xz_r | 100.00% | 100.03% | 100.28% | 100.00% | 100.24% |
Intel Cascade Lake Xeon
Run-times
Benchmark
x86-64
x86-64-v2
x86-64-v3
x86-64-v2
x86-64-v3
500.perlbench_r
100.00%
101.11%
102.00%
100.00%
100.88%
502.gcc_r
100.00%
99.32%
99.32%
100.00%
100.00%
505.mcf_r
100.00%
100.48%
102.61%
100.00%
102.13%
520.omnetpp_r
100.00%
101.25%
101.25%
100.00%
100.00%
523.xalancbmk_r
100.00%
100.43%
100.86%
100.00%
100.43%
525.x264_r
100.00%
94.04%
90.93%
100.00%
96.69%
531.deepsjeng_r
100.00%
96.35%
92.09%
100.00%
95.58%
541.leela_r
100.00%
100.22%
101.32%
100.00%
101.10%
548.exchange2_r
100.00%
98.99%
102.02%
100.00%
103.06%
557.xz_r
100.00%
101.27%
99.15%
100.00%
97.91%
| Benchmark | x86-64 | x86-64-v2 | x86-64-v3 | x86-64-v2 | x86-64-v3 |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 101.11% | 102.00% | 100.00% | 100.88% |
| 502.gcc_r | 100.00% | 99.32% | 99.32% | 100.00% | 100.00% |
| 505.mcf_r | 100.00% | 100.48% | 102.61% | 100.00% | 102.13% |
| 520.omnetpp_r | 100.00% | 101.25% | 101.25% | 100.00% | 100.00% |
| 523.xalancbmk_r | 100.00% | 100.43% | 100.86% | 100.00% | 100.43% |
| 525.x264_r | 100.00% | 94.04% | 90.93% | 100.00% | 96.69% |
| 531.deepsjeng_r | 100.00% | 96.35% | 92.09% | 100.00% | 95.58% |
| 541.leela_r | 100.00% | 100.22% | 101.32% | 100.00% | 101.10% |
| 548.exchange2_r | 100.00% | 98.99% | 102.02% | 100.00% | 103.06% |
| 557.xz_r | 100.00% | 101.27% | 99.15% | 100.00% | 97.91% |
Naturally, the smaller the run-times, the better. On the other hand, rates and especially their geometric means are better when they are greater.
See also continuous gcc tuning benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/tuning.
Non-debug executable sizes
Click to display/hide
Benchmark
x86-64
x86-64-v2
x86-64-v3
x86-64-v2
x86-64-v3
500.perlbench_r
100.00%
99.98%
100.01%
100.00%
100.03%
502.gcc_r
100.00%
99.99%
100.17%
100.00%
100.18%
505.mcf_r
100.00%
101.38%
105.48%
100.00%
104.04%
520.omnetpp_r
100.00%
99.97%
100.80%
100.00%
100.83%
523.xalancbmk_r
100.00%
99.97%
100.73%
100.00%
100.76%
525.x264_r
100.00%
97.47%
95.22%
100.00%
97.69%
531.deepsjeng_r
100.00%
99.99%
99.98%
100.00%
99.98%
541.leela_r
100.00%
100.01%
101.96%
100.00%
101.95%
548.exchange2_r
100.00%
100.00%
103.97%
100.00%
103.97%
557.xz_r
100.00%
100.03%
100.28%
100.00%
100.24%
Smaller binary sizes are always better.
Click to display/hide
| Benchmark | x86-64 | x86-64-v2 | x86-64-v3 | x86-64-v2 | x86-64-v3 |
|---|---|---|---|---|---|
| 500.perlbench_r | 100.00% | 99.98% | 100.01% | 100.00% | 100.03% |
| 502.gcc_r | 100.00% | 99.99% | 100.17% | 100.00% | 100.18% |
| 505.mcf_r | 100.00% | 101.38% | 105.48% | 100.00% | 104.04% |
| 520.omnetpp_r | 100.00% | 99.97% | 100.80% | 100.00% | 100.83% |
| 523.xalancbmk_r | 100.00% | 99.97% | 100.73% | 100.00% | 100.76% |
| 525.x264_r | 100.00% | 97.47% | 95.22% | 100.00% | 97.69% |
| 531.deepsjeng_r | 100.00% | 99.99% | 99.98% | 100.00% | 99.98% |
| 541.leela_r | 100.00% | 100.01% | 101.96% | 100.00% | 101.95% |
| 548.exchange2_r | 100.00% | 100.00% | 103.97% | 100.00% | 103.97% |
| 557.xz_r | 100.00% | 100.03% | 100.28% | 100.00% | 100.24% |