function renaming and redefination of XLeaderPS::MakeAll()

source/train/XLeader.cpp

@@ -317,9 +317,9 @@ void XLeader::Start()
 add a number of job workers (given their device ids) 
 add a number of job workers (given their device ids) 
 >> model - the neural network
 >> model - the neural network
 >> n - number of the models
 >> n - number of the models
->> ids - the array of device ids
+>> devIDs - the array of device ids
 */
 */
-void XLeader::AddJobWorker(XModel * model, int n, int * ids)
+void XLeader::AddJobWorker(XModel * model, int n, const int * devIDs)
 {
 {
     /* we keep the input model */
     /* we keep the input model */
     if (n >= 1) {
     if (n >= 1) {
@@ -331,7 +331,7 @@ void XLeader::AddJobWorker(XModel * model, int n, int * ids)
     /* we clone the input model */
     /* we clone the input model */
     for (int i = 1; i < n; i++) {
     for (int i = 1; i < n; i++) {
         XWorkerJob * worker = new XWorkerJob();
         XWorkerJob * worker = new XWorkerJob();
-        worker->SetModel(model->Clone(ids[i]));
+        worker->SetModel(model->Clone(devIDs[i]));
         jworkers.Add(worker);
         jworkers.Add(worker);
     }
     }
 }
 }
@@ -340,7 +340,7 @@ void XLeader::AddJobWorker(XModel * model, int n, int * ids)
 add a data-collecting worker 
 add a data-collecting worker 
 >> mode - the data-transfer mode of the worker
 >> mode - the data-transfer mode of the worker
 */
 */
-void XLeader::AddJobCollectWorker(DATA_COLLECT_TYPE mode)
+void XLeader::AddCollectWorker(DATA_COLLECT_TYPE mode)
 {
 {
     XWorkerCollect * worker = new XWorkerCollect();
     XWorkerCollect * worker = new XWorkerCollect();
     worker->SetCollectMode(mode);
     worker->SetCollectMode(mode);
@@ -350,17 +350,15 @@ void XLeader::AddJobCollectWorker(DATA_COLLECT_TYPE mode)
 /* 
 /* 
 add a model-update worker 
 add a model-update worker 
 >> model - the model
 >> model - the model
->> optimizer - the optimizer
 */
 */
-void XLeader::AddJobUpdateWorker(XModel * model, XOptimizer * optimizer)
+void XLeader::AddUpdateWorker(XModel * model)
 {
 {
     XWorkerUpdate * worker = new XWorkerUpdate();
     XWorkerUpdate * worker = new XWorkerUpdate();
-    worker->SetOptimizer(optimizer);
     uworkers.Add(worker);
     uworkers.Add(worker);
 }
 }
 
 
 /* add a data-broadcasting worker */
 /* add a data-broadcasting worker */
-void XLeader::AddJobBroadcastWorker()
+void XLeader::AddBroadcastWorker()
 {
 {
     XWorkerBroadcast * worker = new XWorkerBroadcast();
     XWorkerBroadcast * worker = new XWorkerBroadcast();
     bworkers.Add(worker);
     bworkers.Add(worker);
@@ -370,7 +368,7 @@ void XLeader::AddJobBroadcastWorker()
 add a parameter worker (or a pipeline) 
 add a parameter worker (or a pipeline) 
 >> n - number of parameters
 >> n - number of parameters
 */
 */
-void XLeader::AddJobParamterWorker(int n)
+void XLeader::AddParamterWorker(int n)
 {
 {
     for (int i = 0; i < n; i++) {
     for (int i = 0; i < n; i++) {
         XWorker * worker = new XWorker();
         XWorker * worker = new XWorker();

source/train/XLeader.h

@@ -160,19 +160,19 @@ public:
     void SetInstantRun(bool flag = true);
     void SetInstantRun(bool flag = true);
     
     
     /* add a number of job workers (given their device ids) */
     /* add a number of job workers (given their device ids) */
-    void AddJobWorker(XModel * model, int n, int * ids);
+    void AddJobWorker(XModel * model, int n, const int * devIDs);
 
 
     /* add a data-collecting worker */
     /* add a data-collecting worker */
-    void AddJobCollectWorker(DATA_COLLECT_TYPE mode = DATA_COLLECT_P2P);
+    void AddCollectWorker(DATA_COLLECT_TYPE mode = DATA_COLLECT_P2P);
 
 
     /* add a model-update worker */
     /* add a model-update worker */
-    void AddJobUpdateWorker(XModel * model, XOptimizer * optimizer);
+    void AddUpdateWorker(XModel * model);
 
 
     /* add a data-broadcasting worker */
     /* add a data-broadcasting worker */
-    void AddJobBroadcastWorker();
+    void AddBroadcastWorker();
 
 
     /* add a parameter worker (or a pipeline) */
     /* add a parameter worker (or a pipeline) */
-    void AddJobParamterWorker(int n);
+    void AddParamterWorker(int n);
     
     
     /* destroy the parameter map (and gradient map) */
     /* destroy the parameter map (and gradient map) */
     void DestroyParamMap();
     void DestroyParamMap();

source/train/XLeaderPS.cpp

@@ -44,6 +44,25 @@ XLeaderPS::~XLeaderPS()
 }
 }
 
 
 /* 
 /* 
+create workers 
+>> config - configuration
+>> model - the model that we run
+>> devIDs - device ids of the workers (the first id is for server)
+>> jobWorkerNum - number of job workers
+*/
+void XLeaderPS::MakeAll(XConfig * config, XModel * model, const int * devIDs, const int jobWorkerNum)
+{
+    Init();
+    AddJobWorker(model, jobWorkerNum, devIDs);
+    AddCollectWorker();
+    AddUpdateWorker(model);
+    AddBroadcastWorker();
+    AddParamterWorker(model->paramNum);
+
+    XLeader::MakeAll(config, model);
+}
+
+/*
 run the model (for one time). Basically this is a map-reduce process.
 run the model (for one time). Basically this is a map-reduce process.
 >> config - the configuration
 >> config - the configuration
 >> dataDistributor - data distributor
 >> dataDistributor - data distributor

source/train/XLeaderPS.h

@@ -47,6 +47,9 @@ public:
 	/* deconstructor */
 	/* deconstructor */
 	~XLeaderPS();
 	~XLeaderPS();
 
 
+    /* create workers and other stuff used in training */
+    void MakeAll(XConfig * config, XModel * model, const int * devIDs, const int jobWorkerNum);
+
 	/* run the model and update it (for one time) */
 	/* run the model and update it (for one time) */
 	bool Run(XConfig* config, DataDistributeBase* dataDistributor, XOptimizer* optimizer);
 	bool Run(XConfig* config, DataDistributeBase* dataDistributor, XOptimizer* optimizer);
 
 

source/train/XTrainer.cpp

@@ -113,14 +113,8 @@ void XTrainer::Run(XConfig * config, DataDistributeBase * dataDistributor,
 
 
     /* create the server and workers */
     /* create the server and workers */
     XLeaderPS leader;
     XLeaderPS leader;
-    leader.Init();
-    leader.AddJobWorker(model, jobNum, ids);
-    leader.AddJobCollectWorker();
-    leader.AddJobUpdateWorker(model, optimizer);
-    leader.AddJobBroadcastWorker();
-    leader.AddJobParamterWorker(model->paramNum);
+    leader.MakeAll(config, model, ids, jobNum);
     //leader.SetInstantRun();
     //leader.SetInstantRun();
-    leader.MakeAll(config, model);
     leader.Start();
     leader.Start();
 
 
     /* learning rate scheduler */
     /* learning rate scheduler */