aws: Do not skip the WRITE_IN_ORDER_ALIGNED_128_BYTES check for P5

EFA now reports WRITE_IN_ORDER_ALIGNED_128_BYTES for this platform and
the plugin no longer has to circumvent that check to determine support
for the low-latency protocols.

Signed-off-by: Raghu Raja <[email protected]>

src/platform-aws.c

@@ -301,6 +301,7 @@ static int configure_ep_inorder(struct fid_ep *ep, int optname, const char* optn
 			      optname_name, ret, fi_strerror(-ret));
 		return ret;
 	} else {
+		NCCL_OFI_INFO(NCCL_INIT, "Setting %s have_ordering is true.", optname_name);
 		*have_ordering = true;
 	}
 
@@ -667,31 +668,6 @@ int platform_config_endpoint(struct fi_info *info, struct fid_ep* endpoint) {
 	 * ordering.  If we're not expecting ordering, we don't really
 	 * care if ordering is on or off for the endpoint.
 	 */
-
-	/* TODO: This is a temporary hack to disable setting
-	 * NCCL_PROTO=simple on P5 when using the RDMA protocol.  EFA
-	 * on P5 does not currently report
-	 * WRITE_IN_ORDER_ALIGNED_128_BYTES because it can deliver the
-	 * (correct) payload twice.  This violates the meaning of the
-	 * WRITE_IN_ORDER_ALIGNED_128_BYTES flag in rdma-core, but
-	 * does not violate any assumptions about buffer reuse in
-	 * NCCL.  We have confirmed that the EFA provider in Libfabric
-	 * will not segment messages for fi_write(), so this is safe.
-	 * Note that the SENDRECV protocol does have segmentation
-	 * challenges that require us to obey the
-	 * SENDRECV_IN_ORDER_ALIGNED_128_BYTES flag, so we only skip
-	 * the check when using the RDMA protocol.
-	 */
-	if ((NULL == getenv("NCCL_PROTO")) &&
-	    (0 == strcasecmp("RDMA", nccl_ofi_selected_protocol)) &&
-	    (0 == strcmp(get_platform_type(), "p5.48xlarge"))) {
-		if (!nccl_proto_configured) {
-			NCCL_OFI_INFO(NCCL_INIT, "Skipping NCCL_PROTO checks on P5 + RDMA");
-			need_ordering = false;
-			nccl_proto_configured = true;
-		}
-	}
-
 	if (need_ordering || !nccl_proto_configured) {
 		bool have_ordering = false;