Skip to content

Commit

Permalink
Merge pull request #1245 from benjaminulmer/release/rocm-rel-5.1
Browse files Browse the repository at this point in the history
cherry-pick aldebaran fp16 and fp32 tuning (#1118)
  • Loading branch information
TorreZuk authored Apr 27, 2022
2 parents 06fde28 + d2fede0 commit f0273f2
Show file tree
Hide file tree
Showing 11 changed files with 8,272 additions and 8 deletions.

Large diffs are not rendered by default.

Original file line number Diff line number Diff line change
Expand Up @@ -45887,4 +45887,152 @@
- [136, 82.8815]
- - [1568, 1024, 1, 512]
- [137, 67.3595]
- - [1280, 100, 1, 1792]
- [112, 26.37]
- - [4096, 4096, 1, 32]
- [57, 27.778]
- - [768, 768, 1, 7168]
- [121, 76.642]
- - [3136, 96, 32, 128]
- [15, 54.217]
- - [49, 1024, 32, 128]
- [6, 26.164]
- - [49, 1216, 32, 128]
- [7, 30.133]
- - [49, 1376, 32, 128]
- [7, 30.58]
- - [49, 992, 32, 128]
- [6, 25.063]
- - [224, 64, 160, 224]
- [15, 43.854]
- - [224, 64, 384, 224]
- [11, 52.18]
- - [224, 64, 512, 224]
- [107, 54.643]
- - [512, 512, 8, 8192]
- [46, 84.119]
- - [2048, 2048, 2, 8192]
- [15, 89.3]
- - [1000, 16, 1, 1664]
- [23, 3.846]
- - [1000, 16, 1, 1920]
- [119, 4.273]
- - [1000, 256, 1, 1664]
- [23, 39.019]
- - [1000, 256, 1, 1920]
- [20, 39.626]
- - [1000, 256, 1, 2048]
- [20, 39.979]
- - [1000, 32, 1, 4096]
- [123, 12.925]
- - [1000, 64, 1, 1664]
- [23, 15.333]
- - [1000, 96, 1, 2048]
- [50, 24.342]
- - [1000, 96, 1, 4096]
- [123, 29.584]
- - [100, 1024, 1, 1000]
- [20, 16.481]
- - [100, 16, 1, 1024]
- [120, 0.294]
- - [100, 16, 1, 768]
- [44, 0.253]
- - [100, 224, 1, 768]
- [23, 3.623]
- - [100, 256, 1, 1000]
- [23, 4.596]
- - [100, 48, 1, 1024]
- [34, 0.865]
- - [100, 48, 1, 768]
- [20, 0.759]
- - [100, 4, 1, 1024]
- [118, 0.08]
- - [100, 4, 1, 768]
- [5, 0.063]
- - [100, 5376, 1, 1280]
- [15, 44.62]
- - [100, 5376, 1, 768]
- [15, 39.421]
- - [100, 896, 1, 1280]
- [24, 17.171]
- - [100, 896, 1, 768]
- [23, 13.969]
- - [1024, 48, 1, 1024]
- [23, 8.159]
- - [1024, 896, 1, 4096]
- [90, 72.47]
- - [1280, 224, 1, 1280]
- [23, 40.537]
- - [1280, 896, 1, 1280]
- [107, 67.097]
- - [1280, 896, 1, 3840]
- [22, 74.418]
- - [1280, 896, 1, 5120]
- [15, 75.458]
- - [4096, 256, 1, 25088]
- [128, 82.541]
- - [4096, 256, 1, 4096]
- [105, 73.179]
- - [4096, 32, 1, 25088]
- [112, 37.488]
- - [4096, 32, 1, 4096]
- [10, 34.173]
- - [4096, 3584, 1, 1024]
- [15, 88.776]
- - [4096, 896, 1, 1024]
- [15, 78.119]
- - [4096, 96, 1, 25088]
- [112, 59.451]
- - [4096, 96, 1, 4096]
- [112, 64.576]
- - [768, 10752, 1, 768]
- [15, 85.059]
- - [768, 16, 1, 768]
- [28, 1.841]
- - [768, 224, 1, 2304]
- [105, 40.141]
- - [768, 224, 1, 3072]
- [105, 42.228]
- - [768, 224, 1, 768]
- [23, 25.337]
- - [768, 3584, 1, 3072]
- [90, 79.972]
- - [768, 48, 1, 768]
- [28, 5.596]
- - [768, 4, 1, 768]
- [28, 0.451]
- - [768, 5376, 1, 3072]
- [11, 83.384]
- - [768, 896, 1, 2304]
- [22, 63.118]
- - [768, 896, 1, 3072]
- [22, 65.251]
- - [224, 224, 12, 64]
- [71, 10.589]
- - [224, 224, 20, 64]
- [80, 14.329]
- - [224, 224, 192, 64]
- [15, 35.946]
- - [224, 224, 256, 64]
- [7, 38.512]
- - [224, 224, 288, 64]
- [15, 39.062]
- - [224, 224, 48, 64]
- [43, 22.775]
- - [224, 224, 480, 64]
- [15, 41.179]
- - [224, 224, 576, 64]
- [15, 41.472]
- - [224, 224, 64, 64]
- [7, 25.657]
- - [224, 224, 768, 64]
- [15, 42.26]
- - [224, 224, 80, 64]
- [7, 28.586]
- - [64, 224, 48, 224]
- [80, 27.676]
- - [64, 224, 576, 224]
- [107, 52.36]
- - [64, 224, 768, 224]
- [107, 54.482]
- null
Loading

0 comments on commit f0273f2

Please sign in to comment.