Benchmarks: micro benchmarks - add nvbandwidth build (#665)

**Description**
Add nvbandwidth build to repo

---------
Co-authored-by: hongtaozhang <hongtaozhang@microsoft.com>

.gitmodules

@@ -30,3 +30,6 @@
 [submodule "third_party/Megatron/Megatron-DeepSpeed"]
 	path = third_party/Megatron/Megatron-DeepSpeed
 	url = https://github.com/microsoft/Megatron-DeepSpeed.git
+[submodule "third_party/nvbandwidth"]
+	path = third_party/nvbandwidth
+	url = https://github.com/NVIDIA/nvbandwidth.git

dockerfile/cuda11.1.1.dockerfile

@@ -149,7 +149,7 @@ ADD dockerfile/etc /opt/microsoft/
 WORKDIR ${SB_HOME}
 
 ADD third_party third_party
-RUN make -C third_party cuda
+RUN make -C third_party cuda -o nvbandwidth
 
 ADD . .
 RUN python3 -m pip install --upgrade setuptools==65.7 && \

third_party/Makefile

@@ -16,12 +16,12 @@ ROCM_VER ?= $(shell hipconfig -R | grep -oP '\d+\.\d+\.\d+' || echo "0.0.0")
 
 NUM_MAKE_JOBS ?= $(shell nproc --ignore=2)
 
-.PHONY: all cuda_with_msccl cuda rocm common cuda_cutlass cuda_bandwidthTest cuda_nccl_tests cuda_perftest cuda_msccl rocm_perftest fio rocm_rccl_tests rocm_rocblas rocm_bandwidthTest gpcnet cuda_gpuburn cpu_stream cpu_hpl directx_amf_encoding_latency directx_amd rocm_hipblaslt megatron_lm megatron_deepspeed apex_rocm
+.PHONY: all cuda_with_msccl cuda rocm common cuda_cutlass cuda_bandwidthTest cuda_nccl_tests cuda_perftest cuda_msccl rocm_perftest fio rocm_rccl_tests rocm_rocblas rocm_bandwidthTest gpcnet cuda_gpuburn cpu_stream cpu_hpl directx_amf_encoding_latency directx_amd rocm_hipblaslt megatron_lm megatron_deepspeed apex_rocm nvbandwidth
 
 # Build targets.
 all: cuda rocm
 cuda_with_msccl: cuda cuda_msccl
-cuda: common cuda_cutlass cuda_bandwidthTest cuda_nccl_tests cuda_perftest gpcnet cuda_gpuburn megatron_lm megatron_deepspeed
+cuda: common cuda_cutlass cuda_bandwidthTest cuda_nccl_tests cuda_perftest gpcnet cuda_gpuburn megatron_lm megatron_deepspeed nvbandwidth
 rocm: common rocm_perftest rocm_rccl_tests rocm_rocblas rocm_bandwidthTest rocm_hipblaslt megatron_deepspeed apex_rocm
 cpu: common cpu_perftest cpu_stream
 common: fio
@@ -239,3 +239,8 @@ ifneq (,$(wildcard msccl/tests/msccl-tests-nccl/Makefile))
 	mkdir -p $(SB_MICRO_PATH)/bin/msccl-tests-nccl && \
 	cp -r -v ./msccl/tests/msccl-tests-nccl/build/* $(SB_MICRO_PATH)/bin/msccl-tests-nccl/
 endif
+
+# Build nvbandwidth.
+nvbandwidth: sb_micro_path
+	cd ./nvbandwidth && cmake . && make && cd ..
+	cp -v ./nvbandwidth/nvbandwidth $(SB_MICRO_PATH)/bin

nvbandwidth @ 445d8aef

+Subproject commit 445d8aef742e8a48a69779a939996f9e8863df9d