zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-379372569
@cjolivier01 are you OK with the PR?
This
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-379096904
@cjolivier01 I have removed "auto" as much as possible.
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378440323
Here is the performance before and after the optimizations.
The performance is measured with `python
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378440323
Here is the performance before and after the optimizations.
The performance is measured with `python
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378440323
model | batch size | before | after
-- | -- | -- | --
AlexNet | 1 | 268.63 | 378.96
| 2 |
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378439529
batch size | before | after
-- | -- | --
AlexNet | 1 | 268.63 | 378.96
| 2 | 431.88 | 585.72
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378439529
batch size | before | after
-- | -- | --
AlexNet | 1 | 268.63 | 378.96
| 2 | 431.88 | 585.72
zheng-da commented on issue #10317: [MXNET-264] Improve performance of MKLDNN
in small batch sizes.
URL: https://github.com/apache/incubator-mxnet/pull/10317#issuecomment-378047340
Could you please review this PR? @piiswrong @pengzhao-intel @TaoLv