SPEC2026 FPrate: Trunk generic -O2 - standard vs LTO vs LTO+PGO
-O2
AMD EPYC 8635P (zen5 based Sorano)
Run-times
Benchmark
Normal
LTO
LTO+PGO
LTO
LTO+PGO
709.cactus_r
100.00%
105.77%
99.48%
100.00%
94.04%
722.palm_r
100.00%
99.35%
95.42%
100.00%
96.05%
731.astcenc_r
100.00%
101.46%
97.50%
100.00%
96.10%
736.ocio_r
100.00%
99.39%
93.51%
100.00%
94.08%
737.gmsh_r
100.00%
99.14%
97.41%
100.00%
98.26%
748.flightdm_r
100.00%
94.90%
80.68%
100.00%
85.01%
749.fotonik3d_r
100.00%
100.00%
103.71%
100.00%
103.71%
765.roms_r
100.00%
100.76%
86.36%
100.00%
85.71%
766.femflow_r
100.00%
100.00%
27.69%
100.00%
27.69%
767.nest_r
100.00%
100.00%
103.03%
100.00%
103.03%
772.marian_r
100.00%
100.00%
58.82%
100.00%
58.82%
782.lbm_r
100.00%
100.89%
103.11%
100.00%
102.20%
| Benchmark | Normal | LTO | LTO+PGO | LTO | LTO+PGO |
|---|---|---|---|---|---|
| 709.cactus_r | 100.00% | 105.77% | 99.48% | 100.00% | 94.04% |
| 722.palm_r | 100.00% | 99.35% | 95.42% | 100.00% | 96.05% |
| 731.astcenc_r | 100.00% | 101.46% | 97.50% | 100.00% | 96.10% |
| 736.ocio_r | 100.00% | 99.39% | 93.51% | 100.00% | 94.08% |
| 737.gmsh_r | 100.00% | 99.14% | 97.41% | 100.00% | 98.26% |
| 748.flightdm_r | 100.00% | 94.90% | 80.68% | 100.00% | 85.01% |
| 749.fotonik3d_r | 100.00% | 100.00% | 103.71% | 100.00% | 103.71% |
| 765.roms_r | 100.00% | 100.76% | 86.36% | 100.00% | 85.71% |
| 766.femflow_r | 100.00% | 100.00% | 27.69% | 100.00% | 27.69% |
| 767.nest_r | 100.00% | 100.00% | 103.03% | 100.00% | 103.03% |
| 772.marian_r | 100.00% | 100.00% | 58.82% | 100.00% | 58.82% |
| 782.lbm_r | 100.00% | 100.89% | 103.11% | 100.00% | 102.20% |
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
LTO
LTO+PGO
LTO
LTO+PGO
709.cactus_r
100.00%
87.04%
76.27%
100.00%
87.63%
722.palm_r
100.00%
88.89%
83.56%
100.00%
94.01%
731.astcenc_r
100.00%
93.16%
74.69%
100.00%
80.17%
736.ocio_r
100.00%
55.47%
35.73%
100.00%
64.42%
737.gmsh_r
100.00%
64.30%
42.76%
100.00%
66.51%
748.flightdm_r
100.00%
73.87%
57.54%
100.00%
77.90%
749.fotonik3d_r
100.00%
96.67%
78.36%
100.00%
81.06%
765.roms_r
100.00%
83.91%
83.24%
100.00%
99.20%
766.femflow_r
100.00%
8.51%
5.72%
100.00%
67.22%
767.nest_r
100.00%
79.36%
52.04%
100.00%
65.57%
772.marian_r
100.00%
68.00%
45.56%
100.00%
67.00%
782.lbm_r
100.00%
81.90%
90.48%
100.00%
110.47%
Smaller binary sizes are always better.
| Benchmark | Normal | LTO | LTO+PGO | LTO | LTO+PGO |
|---|---|---|---|---|---|
| 709.cactus_r | 100.00% | 87.04% | 76.27% | 100.00% | 87.63% |
| 722.palm_r | 100.00% | 88.89% | 83.56% | 100.00% | 94.01% |
| 731.astcenc_r | 100.00% | 93.16% | 74.69% | 100.00% | 80.17% |
| 736.ocio_r | 100.00% | 55.47% | 35.73% | 100.00% | 64.42% |
| 737.gmsh_r | 100.00% | 64.30% | 42.76% | 100.00% | 66.51% |
| 748.flightdm_r | 100.00% | 73.87% | 57.54% | 100.00% | 77.90% |
| 749.fotonik3d_r | 100.00% | 96.67% | 78.36% | 100.00% | 81.06% |
| 765.roms_r | 100.00% | 83.91% | 83.24% | 100.00% | 99.20% |
| 766.femflow_r | 100.00% | 8.51% | 5.72% | 100.00% | 67.22% |
| 767.nest_r | 100.00% | 79.36% | 52.04% | 100.00% | 65.57% |
| 772.marian_r | 100.00% | 68.00% | 45.56% | 100.00% | 67.00% |
| 782.lbm_r | 100.00% | 81.90% | 90.48% | 100.00% | 110.47% |
AMD EPYC 75F3 (zen3 based high-frequency Milan)
Run-times
Benchmark
LTO
LTO+PGO
709.cactus_r
100.00%
99.14%
722.palm_r
100.00%
96.89%
731.astcenc_r
100.00%
95.57%
736.ocio_r
100.00%
95.58%
737.gmsh_r
100.00%
97.85%
748.flightdm_r
100.00%
83.37%
749.fotonik3d_r
100.00%
101.58%
765.roms_r
100.00%
84.11%
766.femflow_r
100.00%
37.15%
767.nest_r
100.00%
104.12%
772.marian_r
100.00%
64.85%
782.lbm_r
100.00%
101.56%
| Benchmark | LTO | LTO+PGO |
|---|---|---|
| 709.cactus_r | 100.00% | 99.14% |
| 722.palm_r | 100.00% | 96.89% |
| 731.astcenc_r | 100.00% | 95.57% |
| 736.ocio_r | 100.00% | 95.58% |
| 737.gmsh_r | 100.00% | 97.85% |
| 748.flightdm_r | 100.00% | 83.37% |
| 749.fotonik3d_r | 100.00% | 101.58% |
| 765.roms_r | 100.00% | 84.11% |
| 766.femflow_r | 100.00% | 37.15% |
| 767.nest_r | 100.00% | 104.12% |
| 772.marian_r | 100.00% | 64.85% |
| 782.lbm_r | 100.00% | 101.56% |
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
LTO
LTO+PGO
709.cactus_r
100.00%
87.63%
722.palm_r
100.00%
94.01%
731.astcenc_r
100.00%
80.16%
736.ocio_r
100.00%
64.42%
737.gmsh_r
100.00%
66.51%
748.flightdm_r
100.00%
77.91%
749.fotonik3d_r
100.00%
81.06%
765.roms_r
100.00%
99.20%
766.femflow_r
100.00%
67.22%
767.nest_r
100.00%
65.57%
772.marian_r
100.00%
67.00%
782.lbm_r
100.00%
110.42%
Smaller binary sizes are always better.
| Benchmark | LTO | LTO+PGO |
|---|---|---|
| 709.cactus_r | 100.00% | 87.63% |
| 722.palm_r | 100.00% | 94.01% |
| 731.astcenc_r | 100.00% | 80.16% |
| 736.ocio_r | 100.00% | 64.42% |
| 737.gmsh_r | 100.00% | 66.51% |
| 748.flightdm_r | 100.00% | 77.91% |
| 749.fotonik3d_r | 100.00% | 81.06% |
| 765.roms_r | 100.00% | 99.20% |
| 766.femflow_r | 100.00% | 67.22% |
| 767.nest_r | 100.00% | 65.57% |
| 772.marian_r | 100.00% | 67.00% |
| 782.lbm_r | 100.00% | 110.42% |