issues
search
undertherain
/
benchmarker
modular framework for [not only] deep learning performance benchmarking
http://blackbird.pw/performance
Mozilla Public License 2.0
9
stars
5
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Fp16 flops
#147
shwetasalaria
closed
3 years ago
0
don't convert int tensors to dnnl
#146
undertherain
closed
3 years ago
0
perf flops do not measure correctly on Ryzen CPU
#145
undertherain
opened
3 years ago
2
estimate ops for bert_custom
#144
undertherain
closed
3 years ago
0
preheat
#143
vatai
closed
3 years ago
0
Bert flops
#142
vatai
closed
3 years ago
0
FP16 flops are not being measured
#141
undertherain
closed
3 years ago
0
Add flop per joule
#140
vatai
closed
3 years ago
3
Add barriers to abstract process invocation
#139
vatai
closed
3 years ago
0
Fix freeze
#138
undertherain
closed
3 years ago
0
Profile pytorch
#137
shwetasalaria
closed
3 years ago
0
(first merge this this!) Remove redundant duplicate args
#136
vatai
closed
3 years ago
0
Fapp
#135
vatai
closed
3 years ago
0
Measure flops optionally
#134
shwetasalaria
closed
3 years ago
0
refactoring efficiency and use builting amp
#133
undertherain
closed
3 years ago
0
gpu power monitoring thread does not exist when main thread crashes
#132
undertherain
closed
3 years ago
1
Ops per second for dl
#131
undertherain
closed
3 years ago
0
need to update download paths for travis
#130
undertherain
opened
3 years ago
1
nb_epoch for gemm kernels
#129
undertherain
closed
3 years ago
0
integrate perf measurement nicely
#128
undertherain
closed
3 years ago
0
Conv2d operations
#127
vatai
closed
3 years ago
0
better measureent of cublas power
#126
undertherain
closed
3 years ago
0
report power in cublas
#125
undertherain
closed
3 years ago
0
Cublas16
#124
undertherain
closed
3 years ago
0
Cudnn
#123
vatai
closed
3 years ago
1
[cudnn?] Do both channel first and channel last
#122
vatai
opened
3 years ago
0
Spearate/factor out `params[]` related code
#121
vatai
opened
3 years ago
0
[cudnn] (and maybe [cublas]) figure out automatically maximal -arch=smXX
#120
vatai
opened
3 years ago
0
[cudnn] implement multiple epochs
#119
vatai
closed
3 years ago
0
[cudnn] implement activations
#118
vatai
opened
3 years ago
0
[cudnn] 1d conv is not working
#117
vatai
opened
3 years ago
0
[cudnn] implement different precission
#116
vatai
opened
3 years ago
0
[cudnn] implement strings algorithm names
#115
vatai
closed
3 years ago
0
estimate operations for conv2d kernel
#114
shwetasalaria
closed
3 years ago
0
Cudnn
#113
vatai
closed
3 years ago
0
Power
#112
undertherain
closed
3 years ago
0
muplty-gpu scaling sucks
#111
undertherain
opened
3 years ago
0
remove redundant ssd
#110
undertherain
opened
3 years ago
0
resnet50 fails on inference / not detected by unittests
#109
undertherain
opened
3 years ago
1
Fp16
#108
undertherain
closed
3 years ago
0
multi-head self attention
#107
undertherain
closed
3 years ago
0
mixed precision with TF
#106
undertherain
closed
3 years ago
1
bert does not work with DNNL
#105
undertherain
opened
3 years ago
0
Unparsed args
#104
undertherain
closed
3 years ago
0
Bert large
#103
undertherain
closed
4 years ago
0
add onnxruntime support
#102
undertherain
opened
4 years ago
0
Edgetpu
#101
vatai
closed
3 years ago
0
support of DL models fully in FP16
#100
undertherain
opened
4 years ago
0
imagenette as a small image classification dataset
#99
undertherain
opened
4 years ago
0
Cudnn benchmark param
#98
vatai
closed
4 years ago
0
Previous
Next