From dev-return-2955-archive-asf-public=cust-asf.ponee.io@singa.incubator.apache.org  Wed Jul  3 14:03:21 2019
Return-Path: <dev-return-2955-archive-asf-public=cust-asf.ponee.io@singa.incubator.apache.org>
X-Original-To: archive-asf-public@cust-asf.ponee.io
Delivered-To: archive-asf-public@cust-asf.ponee.io
Received: from mail.apache.org (hermes.apache.org [207.244.88.153])
	by mx-eu-01.ponee.io (Postfix) with SMTP id 239AA180109
	for <archive-asf-public@cust-asf.ponee.io>; Wed,  3 Jul 2019 16:03:20 +0200 (CEST)
Received: (qmail 12619 invoked by uid 500); 3 Jul 2019 14:03:19 -0000
Mailing-List: contact dev-help@singa.incubator.apache.org; run by ezmlm
Precedence: bulk
List-Help: <mailto:dev-help@singa.incubator.apache.org>
List-Unsubscribe: <mailto:dev-unsubscribe@singa.incubator.apache.org>
List-Post: <mailto:dev@singa.incubator.apache.org>
List-Id: <dev.singa.incubator.apache.org>
Reply-To: dev@singa.incubator.apache.org
Delivered-To: mailing list dev@singa.incubator.apache.org
Received: (qmail 12350 invoked by uid 99); 3 Jul 2019 14:03:19 -0000
Received: from ec2-52-202-80-70.compute-1.amazonaws.com (HELO gitbox.apache.org) (52.202.80.70)
    by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 03 Jul 2019 14:03:19 +0000
From: GitBox <git@apache.org>
To: dev@singa.apache.org
Subject: [GitHub] [incubator-singa] nudles commented on a change in pull request
 #468: Distributted module
Message-ID: <156216259902.18407.1952009714746603573.gitbox@gitbox.apache.org>
Date: Wed, 03 Jul 2019 14:03:19 -0000
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit

nudles commented on a change in pull request #468: Distributted module
URL: https://github.com/apache/incubator-singa/pull/468#discussion_r299973462
 
 

 ##########
 File path: src/dist/communicator.cc
 ##########
 @@ -0,0 +1,150 @@
+#include "singa/dist/communicator.h"
+#include<iostream>
+namespace singa{
+
+static uint64_t getHostHash(const char* string) {
+  // Based on DJB2, result = result * 33 + char
+  uint64_t result = 5381;
+  for (int c = 0; string[c] != '\0'; c++){
+    result = ((result << 5) + result) + string[c];
+  }
+  return result;
+}
+
+
+static void getHostName(char* hostname, int maxlen) {
+  gethostname(hostname, maxlen);
+  for (int i=0; i< maxlen; i++) {
+    if (hostname[i] == '.') {
+        hostname[i] = '\0';
+        return;
+    }
+  }
+}
+
+
+Communicator::Communicator(int nDev): nDev(nDev){
+  MPICHECK(MPI_Init(NULL, NULL));
+  // get MPI Global Ranks and total Ranks
+  MPICHECK(MPI_Comm_rank(MPI_COMM_WORLD, &MPIRankInGlobal));
+  MPICHECK(MPI_Comm_size(MPI_COMM_WORLD, &totalMPIRanksInGlobal));
+  //std::cout<<"g rank " << MPIRankInGlobal << "\n";
+
+  //calculating MPIRankInLocal which is used in selecting a GPU
+  MPIRankInLocal=0;
+  uint64_t hostHashs[totalMPIRanksInGlobal];
+  char hostname[1024];
+  getHostName(hostname, 1024);
+  hostHashs[MPIRankInGlobal] = getHostHash(hostname);
+  MPICHECK(MPI_Allgather(MPI_IN_PLACE, 0, MPI_DATATYPE_NULL, hostHashs,
+    		 sizeof(uint64_t), MPI_BYTE, MPI_COMM_WORLD));
+  for (int p=0; p<totalMPIRanksInGlobal; p++) {
+     if (p == MPIRankInGlobal) break;
+     if (hostHashs[p] == hostHashs[MPIRankInGlobal]) MPIRankInLocal++;
+  }
+
+  //std::cout<<"l rank " << MPIRankInLocal << "\n";
+
+  //picking GPUs based on MPIRankInLocal
+  //create cuda stream s
+  s = (cudaStream_t*)malloc(sizeof(cudaStream_t)*nDev);
+  for (int i = 0; i < nDev; ++i) {
+    CUDACHECK(cudaSetDevice(MPIRankInLocal*nDev + i));
 
 Review comment:
   what's the relationship between MPIRankInLocal and nDev?

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services