Merge pull request #4821 from nrspruit/OFI_mtl_multi_event_progress

MTL OFI: Added support for reading multiple CQ events in ofi progress

ompi/mca/mtl/ofi/mtl_ofi.h

@@ -63,8 +63,7 @@ __opal_attribute_always_inline__ static inline int
 ompi_mtl_ofi_progress(void)
 {
     ssize_t ret;
-    int count = 0;
+    int count = 0, i, events_read;
-    struct fi_cq_tagged_entry wc = { 0 };
     struct fi_cq_err_entry error = { 0 };
     ompi_mtl_ofi_request_t *ofi_req = NULL;
 
@@ -74,19 +73,23 @@ ompi_mtl_ofi_progress(void)
      * Call the request's callback.
      */
     while (true) {
-        ret = fi_cq_read(ompi_mtl_ofi.cq, (void *)&wc, 1);
+        ret = fi_cq_read(ompi_mtl_ofi.cq, ompi_mtl_ofi.progress_entries,
+                ompi_mtl_ofi.ofi_progress_event_count);
         if (ret > 0) {
-            count++;
+            count+= ret;
-            if (NULL != wc.op_context) {
+            events_read = ret;
-                ofi_req = TO_OFI_REQ(wc.op_context);
+            for (i = 0; i < events_read; i++) {
-                assert(ofi_req);
+                if (NULL != ompi_mtl_ofi.progress_entries[i].op_context) {
-                ret = ofi_req->event_callback(&wc, ofi_req);
+                    ofi_req = TO_OFI_REQ(ompi_mtl_ofi.progress_entries[i].op_context);
-                if (OMPI_SUCCESS != ret) {
+                    assert(ofi_req);
-                    opal_output(0, "%s:%d: Error returned by request event callback: %zd.\n"
+                    ret = ofi_req->event_callback(&ompi_mtl_ofi.progress_entries[i], ofi_req);
-                                   "*** The Open MPI OFI MTL is aborting the MPI job (via exit(3)).\n",
+                    if (OMPI_SUCCESS != ret) {
-                                   __FILE__, __LINE__, ret);
+                        opal_output(0, "%s:%d: Error returned by request event callback: %zd.\n"
-                    fflush(stderr);
+                                       "*** The Open MPI OFI MTL is aborting the MPI job (via exit(3)).\n",
-                    exit(1);
+                                       __FILE__, __LINE__, ret);
+                        fflush(stderr);
+                        exit(1);
+                    }
                 }
             }
         } else if (OPAL_UNLIKELY(ret == -FI_EAVAIL)) {

ompi/mca/mtl/ofi/mtl_ofi_component.c

@@ -98,6 +98,7 @@ ompi_mtl_ofi_component_register(void)
 {
     int ret;
     mca_base_var_enum_t *new_enum = NULL;
+    char *desc;
 
     param_priority = 25;   /* for now give a lower priority than the psm mtl */
     mca_base_component_var_register(&mca_mtl_ofi_component.super.mtl_version,
@@ -125,6 +126,18 @@ ompi_mtl_ofi_component_register(void)
                                     MCA_BASE_VAR_SCOPE_READONLY,
                                     &prov_exclude);
 
+    ompi_mtl_ofi.ofi_progress_event_count = 100;
+    asprintf(&desc, "Max number of events to read each call to OFI progress (default: %d events will be read per OFI progress call)", ompi_mtl_ofi.ofi_progress_event_count);
+    mca_base_component_var_register(&mca_mtl_ofi_component.super.mtl_version,
+                                    "progress_event_cnt",
+                                    desc,
+                                    MCA_BASE_VAR_TYPE_INT, NULL, 0, 0,
+                                    OPAL_INFO_LVL_6,
+                                    MCA_BASE_VAR_SCOPE_READONLY,
+                                    &ompi_mtl_ofi.ofi_progress_event_count);
+
+     free(desc);
+
     ret = mca_base_var_enum_create ("control_prog_type", control_prog_type, &new_enum);
     if (OPAL_SUCCESS != ret) {
         return ret;
@@ -465,6 +478,19 @@ ompi_mtl_ofi_component_init(bool enable_progress_threads,
      *     - dynamic memory-spanning memory region
      */
     cq_attr.format = FI_CQ_FORMAT_TAGGED;
+
+    /**
+     * If a user has set an ofi_progress_event_count > the default, then
+     * the CQ size hint is set to the user's desired value such that
+     * the CQ created will have enough slots to store up to
+     * ofi_progress_event_count events. If a user has not set the
+     * ofi_progress_event_count, then the provider is trusted to set a
+     * default high CQ size and the CQ size hint is left unspecified.
+     */
+    if (ompi_mtl_ofi.ofi_progress_event_count > 100) {
+        cq_attr.size = ompi_mtl_ofi.ofi_progress_event_count;
+    }
+
     ret = fi_cq_open(ompi_mtl_ofi.domain, &cq_attr, &ompi_mtl_ofi.cq, NULL);
     if (ret) {
         opal_output_verbose(1, ompi_mtl_base_framework.framework_output,
@@ -473,6 +499,17 @@ ompi_mtl_ofi_component_init(bool enable_progress_threads,
         goto error;
     }
 
+    /**
+     * Allocate memory for storing the CQ events read in OFI progress.
+     */
+    ompi_mtl_ofi.progress_entries = calloc(ompi_mtl_ofi.ofi_progress_event_count, sizeof(struct fi_cq_tagged_entry));
+    if (OPAL_UNLIKELY(!ompi_mtl_ofi.progress_entries)) {
+        opal_output_verbose(1, ompi_mtl_base_framework.framework_output,
+                            "%s:%d: alloc of CQ event storage failed: %s\n",
+                            __FILE__, __LINE__, strerror(errno));
+        goto error;
+    }
+
     /**
      * The remote fi_addr will be stored in the ofi_endpoint struct.
      */
@@ -595,6 +632,10 @@ error:
     if (ompi_mtl_ofi.fabric) {
         (void) fi_close((fid_t)ompi_mtl_ofi.fabric);
     }
+    if (ompi_mtl_ofi.progress_entries) {
+        free(ompi_mtl_ofi.progress_entries);
+    }
+
     return NULL;
 }
 
@@ -626,6 +667,8 @@ ompi_mtl_ofi_finalize(struct mca_mtl_base_module_t *mtl)
         goto finalize_err;
     }
 
+    free(ompi_mtl_ofi.progress_entries);
+
     return OMPI_SUCCESS;
 
 finalize_err:

ompi/mca/mtl/ofi/mtl_ofi_types.h

@@ -49,6 +49,12 @@ typedef struct mca_mtl_ofi_module_t {
     /** Maximum inject size */
     size_t max_inject_size;
 
+    /** Maximum number of CQ events to read in OFI Progress */
+    int ofi_progress_event_count;
+
+    /** CQ event storage */
+    struct fi_cq_tagged_entry *progress_entries;
+
 } mca_mtl_ofi_module_t;
 
 extern mca_mtl_ofi_module_t ompi_mtl_ofi;