Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
H
hpvm-release
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Wiki
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Snippets
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Model registry
Operate
Environments
Monitor
Incidents
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
llvm
hpvm-release
Commits
6c6fe1eb
Commit
6c6fe1eb
authored
4 years ago
by
Hashim Sharif
Browse files
Options
Downloads
Patches
Plain Diff
Fixing buggy LeNet configuration
parent
938c0253
No related branches found
No related tags found
No related merge requests found
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
hpvm/test/dnn_benchmarks/hpvm-c/benchmarks/lenet_mnist/data/tuner_confs.txt
+100
-100
100 additions, 100 deletions
...hmarks/hpvm-c/benchmarks/lenet_mnist/data/tuner_confs.txt
with
100 additions
and
100 deletions
hpvm/test/dnn_benchmarks/hpvm-c/benchmarks/lenet_mnist/data/tuner_confs.txt
+
100
−
100
View file @
6c6fe1eb
...
@@ -9,400 +9,400 @@ conf1 1 1 98.7 0.0
...
@@ -9,400 +9,400 @@ conf1 1 1 98.7 0.0
-----
-----
+++++
+++++
conf2 1.828613181003043 2.071721708828981 98.65 0.04999999999999716
conf2 1.828613181003043 2.071721708828981 98.65 0.04999999999999716
1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 156 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 261 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf3 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
conf3 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 152 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf4 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
conf4 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 152 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf5 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
conf5 1.8936889628815377 2.139779619692146 98.65 0.04999999999999716
1 gpu conv perf_fp16 152 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 152 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf6 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
conf6 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 261 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf7 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
conf7 1.8247639611533713 2.0227145446958756 98.64 0.060000000000002274
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 261 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf8 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
conf8 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf9 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
conf9 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf10 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
conf10 1.8406161850501603 2.037849502542524 98.64 0.060000000000002274
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf11 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
conf11 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 155 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf12 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
conf12 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 155 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf13 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
conf13 1.8663357888260776 2.115790921611576 98.64 0.060000000000002274
1 gpu conv perf_fp16 155 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 155 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf14 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
conf14 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 167 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf15 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
conf15 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 167 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf16 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
conf16 1.8645645142051612 2.1037012333044935 98.61999999999999 0.0800000000000125
1 gpu conv perf_fp16 167 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 167 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf17 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
conf17 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv samp_fp16 264 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf18 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
conf18 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv samp_fp16 264 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf19 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
conf19 2.2168527051833635 2.453341076720038 98.61999999999999 0.0800000000000125
1 gpu conv samp_fp16 264 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv samp_fp16 264 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf20 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf20 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 12 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 12 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf21 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf21 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 12 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 12 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf22 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf22 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 12 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 12 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf23 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf23 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf24 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf24 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf25 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
conf25 1.8406161850501603 2.037849502542524 98.6 0.10000000000000853
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf26 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
conf26 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
1 gpu conv samp_fp16 266 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv samp_fp16 266 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf27 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
conf27 2.200653361151419 2.425091789360736 98.6 0.10000000000000853
1 gpu conv samp_fp16 266 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv samp_fp16 266 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf28 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
conf28 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf29 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
conf29 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf30 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
conf30 1.8406161850501603 2.037849502542524 98.58 0.12000000000000455
1 gpu conv fp16 11 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv fp16 11 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf31 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
conf31 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 156 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf32 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
conf32 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 156 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf33 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
conf33 1.8445326456180258 2.087601822059355 98.58 0.12000000000000455
1 gpu conv perf_fp16 156 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 156 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf34 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
conf34 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 158 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf35 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
conf35 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 158 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf36 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
conf36 1.8916677984300285 2.155437579874673 98.58 0.12000000000000455
1 gpu conv perf_fp16 158 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 158 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf37 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
conf37 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 168 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf38 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
conf38 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 168 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf39 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
conf39 1.8649226857257986 2.1076025277601325 98.56 0.14000000000000057
1 gpu conv perf_fp16 168 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 168 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf40 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
conf40 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 157 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf41 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
conf41 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 157 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf42 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
conf42 1.8463058650555446 2.067271423078985 98.56 0.14000000000000057
1 gpu conv perf_fp16 157 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 157 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf43 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
conf43 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 153 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf44 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
conf44 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 153 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf45 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
conf45 1.9234076467497994 2.1864740913112275 98.56 0.14000000000000057
1 gpu conv perf_fp16 153 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 153 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf46 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
conf46 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 159 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf47 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
conf47 1.8698191484268973 2.13979218727595 98.54 0.1599999999999966
1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 159 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf48 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
conf48 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 165 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf49 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
conf49 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 165 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf50 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
conf50 1.8575043605938137 2.092057786757256 98.52 0.18000000000000682
1 gpu conv perf_fp16 165 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 165 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 262 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 262 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
-----
-----
+++++
+++++
conf51 1.8534621507951072 2.1231113105788597 98.44000000000001 0.2599999999999909
conf51 1.8534621507951072 2.1231113105788597 98.44000000000001 0.2599999999999909
1 gpu conv perf_fp16 159 add fp16 1 pool_max fp16 1
tanh fp16 1
1 gpu conv perf_fp16 159 add fp16 1
tanh fp16 12
pool_max fp16 1
2
2 gpu conv samp_fp16 261 add fp16 1 pool_max fp16 1
tanh fp16 1
2 gpu conv samp_fp16 261 add fp16 1
tanh fp16 12
pool_max fp16 1
2
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
3 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
4 gpu mul fp16 12 add fp16 1 tanh fp16 1
5 gpu softmax fp32 1
5 gpu softmax fp32 1
...
...
This diff is collapsed.
Click to expand it.
Preview
0%
Loading
Try again
or
attach a new file
.
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Save comment
Cancel
Please
register
or
sign in
to comment