cjolivier01 commented on a change in pull request #9933: [MXNET-23] Adding
support to profile kvstore server during distributed training
URL: https://github.com/apache/incubator-mxnet/pull/9933#discussion_r174010221
##########
File path: include/mxnet/kvstore.h
##########
@@ -361,6 +373,17 @@ class KVStore {
*/
virtual void SendCommandToServers(int cmd_id, const std::string& cmd_body) {
}
+ /**
+ * \brief Sends server profiler commands to all server nodes
+ * Only the worker with rank=0 sends the command which will be received by
all servers
+ * \param type ProfilerCommand type
+ * \param params parameters for that command in the form of a string
+ */
+ virtual void SetServerProfilerCommand(const KVStoreServerProfilerCommand
type,
+ const std::string& params) {
+ LOG(FATAL) << "compile with USE_DIST_KVSTORE=1 to use distributed kvstore";
Review comment:
or a CHECK may be a caught error, right? is FATAL catchable? python user
might not know whether kvstore was turned on. i mean, is this something where
execution just can?t continue?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
With regards,
Apache Git Services
