Usually, located at /usr/local/cuda/bin
$ nvprof python train_mnist.py
I prefer to use --print-gpu-trace.
centos6 だと epel の libev が古くてめんどくさくなったので centos7 でビルドした。
sudo yum groupinstall "Development Tools"
sudo yum install libev libev-devel zlib zlib-devel openssl openssl-devel git
git clone git@github.com:tatsuhiro-t/nghttp2.git
cd nghttp2
rb_thread_call_without_gvl を使う
例えば普通に書くと
nvrtcResult status = nvrtcCreateProgram(&_prog, _src, _name, _numHeaders, _headers, _includeNames);となる C 関数の呼び出しを、GVLを外して呼び出すには、
| # A trick to embed preprocessors in cython | |
| cdef extern from *: | |
| cdef void EMIT_IF_PYTHON_VERSION_HEX_LT_37 "#if PY_VERSION_HEX < 0x03070000 //" () | |
| cdef void EMIT_ELSE "#else //" () | |
| cdef void EMIT_ENDIF "#endif //" () | |
| EMIT_IF_PYTHON_VERSION_HEX_LT_37() | |
| EMIT_ELSE() |
| 2019-11-28 08:15:11 +0000 [info]: parsing config file is succeeded path="/etc/google-fluentd/google-fluentd.conf" | |
| 2019-11-28 08:15:12 +0000 [info]: Detected GCE platform | |
| 2019-11-28 08:15:12 +0000 [info]: Logs viewer address: https://console.cloud.google.com/logs/viewer?project=teak-kit-226806&resource=container/instance_id/7530681816608554840 | |
| 2019-11-28 08:15:12 +0000 [info]: Detected GCE platform | |
| 2019-11-28 08:15:12 +0000 [info]: Logs viewer address: https://console.cloud.google.com/logs/viewer?project=teak-kit-226806&resource=gce_instance/instance_id/7530681816608554840 | |
| 2019-11-28 08:15:12 +0000 [info]: Detected GCE platform | |
| 2019-11-28 08:15:12 +0000 [info]: Logs viewer address: https://console.cloud.google.com/logs/viewer?project=teak-kit-226806&resource=container/instance_id/7530681816608554840 | |
| 2019-11-28 08:15:13 +0000 [warn]: 'time_format' specified without 'time_key', will be ignored | |
| 2019-11-28 08:15:13 +0000 [warn]: 'filters' parameter is deprecated: filters has been renamed as matches | |
| 2019-11-28 08:1 |
| #include <sys/time.h> | |
| #include <cuda_runtime.h> | |
| #include <stdio.h> | |
| inline double seconds() | |
| { | |
| struct timeval tp; | |
| struct timezone tzp; | |
| int i = gettimeofday(&tp, &tzp); | |
| return ((double)tp.tv_sec + (double)tp.tv_usec * 1.e-6); |
| GOROOT_1_11_X64=/usr/local/go1.11 | |
| ANDROID_HOME=/usr/local/lib/android/sdk | |
| JAVA_HOME_11_X64=/usr/lib/jvm/zulu-11-azure-amd64 | |
| ImageVersion=20190930.2 | |
| LANG=C.UTF-8 | |
| INVOCATION_ID=029524e610de4329ac57a8f3e8b21281 | |
| JAVA_HOME_12_X64=/usr/lib/jvm/zulu-12-azure-amd64 | |
| ANDROID_SDK_ROOT=/usr/local/lib/android/sdk | |
| RUNNER_TOOL_CACHE=/opt/hostedtoolcache | |
| JAVA_HOME=/usr/lib/jvm/zulu-11-azure-amd64 |
| apiVersion: apps/v1 | |
| kind: Deployment | |
| metadata: | |
| name: my-sweet-api | |
| spec: | |
| template: | |
| metadata: | |
| annotations: | |
| fluentd_firehose_delivery_stream_name: "my-sweet-firehose-stream-name" |