SPEC2026 FPrate: Trunk generic -O2 - standard vs PGO vs LTO+PGO
-O2
AMD EPYC 8635P (zen5 based Sorano)
Run-times
Benchmark
Normal
PGO
LTO+PGO
PGO
LTO+PGO
709.cactus_r
100.00%
102.10%
99.48%
100.00%
97.43%
722.palm_r
100.00%
95.42%
95.42%
100.00%
100.00%
731.astcenc_r
100.00%
97.08%
97.50%
100.00%
100.43%
736.ocio_r
100.00%
93.51%
93.51%
100.00%
100.00%
737.gmsh_r
100.00%
99.14%
97.41%
100.00%
98.26%
748.flightdm_r
100.00%
91.51%
80.68%
100.00%
88.17%
749.fotonik3d_r
100.00%
105.34%
103.71%
100.00%
98.46%
765.roms_r
100.00%
85.61%
86.36%
100.00%
100.88%
766.femflow_r
100.00%
38.68%
27.69%
100.00%
71.59%
767.nest_r
100.00%
103.90%
103.03%
100.00%
99.17%
772.marian_r
100.00%
58.48%
58.82%
100.00%
100.59%
782.lbm_r
100.00%
99.56%
103.11%
100.00%
103.57%
| Benchmark | Normal | PGO | LTO+PGO | PGO | LTO+PGO |
|---|---|---|---|---|---|
| 709.cactus_r | 100.00% | 102.10% | 99.48% | 100.00% | 97.43% |
| 722.palm_r | 100.00% | 95.42% | 95.42% | 100.00% | 100.00% |
| 731.astcenc_r | 100.00% | 97.08% | 97.50% | 100.00% | 100.43% |
| 736.ocio_r | 100.00% | 93.51% | 93.51% | 100.00% | 100.00% |
| 737.gmsh_r | 100.00% | 99.14% | 97.41% | 100.00% | 98.26% |
| 748.flightdm_r | 100.00% | 91.51% | 80.68% | 100.00% | 88.17% |
| 749.fotonik3d_r | 100.00% | 105.34% | 103.71% | 100.00% | 98.46% |
| 765.roms_r | 100.00% | 85.61% | 86.36% | 100.00% | 100.88% |
| 766.femflow_r | 100.00% | 38.68% | 27.69% | 100.00% | 71.59% |
| 767.nest_r | 100.00% | 103.90% | 103.03% | 100.00% | 99.17% |
| 772.marian_r | 100.00% | 58.48% | 58.82% | 100.00% | 100.59% |
| 782.lbm_r | 100.00% | 99.56% | 103.11% | 100.00% | 103.57% |
Naturally, the smaller the run-times, the better.
See also continuous gcc options benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/options.
Non-debug executable sizes
Benchmark
Normal
PGO
LTO+PGO
PGO
LTO+PGO
709.cactus_r
100.00%
89.84%
76.27%
100.00%
84.90%
722.palm_r
100.00%
94.05%
83.56%
100.00%
88.85%
731.astcenc_r
100.00%
87.05%
74.69%
100.00%
85.80%
736.ocio_r
100.00%
57.21%
35.73%
100.00%
62.45%
737.gmsh_r
100.00%
69.51%
42.76%
100.00%
61.52%
748.flightdm_r
100.00%
68.66%
57.54%
100.00%
83.80%
749.fotonik3d_r
100.00%
85.87%
78.36%
100.00%
91.25%
765.roms_r
100.00%
102.54%
83.24%
100.00%
81.18%
766.femflow_r
100.00%
67.27%
5.72%
100.00%
8.51%
767.nest_r
100.00%
64.74%
52.04%
100.00%
80.38%
772.marian_r
100.00%
65.67%
45.56%
100.00%
69.38%
782.lbm_r
100.00%
106.42%
90.48%
100.00%
85.02%
Smaller binary sizes are always better.
| Benchmark | Normal | PGO | LTO+PGO | PGO | LTO+PGO |
|---|---|---|---|---|---|
| 709.cactus_r | 100.00% | 89.84% | 76.27% | 100.00% | 84.90% |
| 722.palm_r | 100.00% | 94.05% | 83.56% | 100.00% | 88.85% |
| 731.astcenc_r | 100.00% | 87.05% | 74.69% | 100.00% | 85.80% |
| 736.ocio_r | 100.00% | 57.21% | 35.73% | 100.00% | 62.45% |
| 737.gmsh_r | 100.00% | 69.51% | 42.76% | 100.00% | 61.52% |
| 748.flightdm_r | 100.00% | 68.66% | 57.54% | 100.00% | 83.80% |
| 749.fotonik3d_r | 100.00% | 85.87% | 78.36% | 100.00% | 91.25% |
| 765.roms_r | 100.00% | 102.54% | 83.24% | 100.00% | 81.18% |
| 766.femflow_r | 100.00% | 67.27% | 5.72% | 100.00% | 8.51% |
| 767.nest_r | 100.00% | 64.74% | 52.04% | 100.00% | 80.38% |
| 772.marian_r | 100.00% | 65.67% | 45.56% | 100.00% | 69.38% |
| 782.lbm_r | 100.00% | 106.42% | 90.48% | 100.00% | 85.02% |
AMD EPYC 75F3 (zen3 based high-frequency Milan)
Run-times
Benchmark
PGO
LTO+PGO
709.cactus_r
100.00%
100.00%
722.palm_r
100.00%
100.86%
731.astcenc_r
100.00%
101.06%
736.ocio_r
100.00%
102.15%
737.gmsh_r
100.00%
97.23%
748.flightdm_r
100.00%
87.29%
749.fotonik3d_r
100.00%
101.58%
765.roms_r
100.00%
100.53%
766.femflow_r
100.00%
67.83%
767.nest_r
100.00%
98.97%
772.marian_r
100.00%
105.78%
782.lbm_r
100.00%
101.11%
| Benchmark | PGO | LTO+PGO |
|---|---|---|
| 709.cactus_r | 100.00% | 100.00% |
| 722.palm_r | 100.00% | 100.86% |
| 731.astcenc_r | 100.00% | 101.06% |
| 736.ocio_r | 100.00% | 102.15% |
| 737.gmsh_r | 100.00% | 97.23% |
| 748.flightdm_r | 100.00% | 87.29% |
| 749.fotonik3d_r | 100.00% | 101.58% |
| 765.roms_r | 100.00% | 100.53% |
| 766.femflow_r | 100.00% | 67.83% |
| 767.nest_r | 100.00% | 98.97% |
| 772.marian_r | 100.00% | 105.78% |
| 782.lbm_r | 100.00% | 101.11% |
Naturally, the smaller the run-times, the better.
See also continuous gcc options benchmarking at https://lnt.opensuse.org/db_default/v4/SPEC/spec_report/options.
Non-debug executable sizes
Benchmark
PGO
LTO+PGO
709.cactus_r
100.00%
84.90%
722.palm_r
100.00%
88.85%
731.astcenc_r
100.00%
85.79%
736.ocio_r
100.00%
62.46%
737.gmsh_r
100.00%
61.52%
748.flightdm_r
100.00%
83.80%
749.fotonik3d_r
100.00%
91.26%
765.roms_r
100.00%
81.18%
766.femflow_r
100.00%
8.51%
767.nest_r
100.00%
80.39%
772.marian_r
100.00%
69.38%
782.lbm_r
100.00%
85.07%
Smaller binary sizes are always better.
| Benchmark | PGO | LTO+PGO |
|---|---|---|
| 709.cactus_r | 100.00% | 84.90% |
| 722.palm_r | 100.00% | 88.85% |
| 731.astcenc_r | 100.00% | 85.79% |
| 736.ocio_r | 100.00% | 62.46% |
| 737.gmsh_r | 100.00% | 61.52% |
| 748.flightdm_r | 100.00% | 83.80% |
| 749.fotonik3d_r | 100.00% | 91.26% |
| 765.roms_r | 100.00% | 81.18% |
| 766.femflow_r | 100.00% | 8.51% |
| 767.nest_r | 100.00% | 80.39% |
| 772.marian_r | 100.00% | 69.38% |
| 782.lbm_r | 100.00% | 85.07% |