Skip to content

Updates to work with CUDA >=12.9 #40

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
prateekchawla168 wants to merge 9 commits into
base: main
Choose a base branch
from
Open

Updates to work with CUDA >=12.9 #40

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
569b46d
Updates to work with CUDA 13.x
prateekchawla168 Mar 23, 2026
85d7e90
update to c++17
prateekchawla168 Mar 24, 2026
50a774e
updates to work with cuda 13
prateekchawla168 Mar 26, 2026
cfeb8bc
changes to accomodate for cuda 13
prateekchawla168 Mar 26, 2026
840123a
Merge pull request #1 from prateekchawla168/06-overlap
prateekchawla168 Apr 8, 2026
193525f
Merge pull request #2 from prateekchawla168/08-nccl
prateekchawla168 Apr 8, 2026
90378b5
use dynamic linking for nvshmem, fixes for cuda 12.9+
prateekchawla168 Apr 9, 2026
e2ee4c2
fix cuda graphs to work with dynamic linked nvshmem and cuda 13
prateekchawla168 Apr 9, 2026
1c49b5b
update cuda graphs makefile
prateekchawla168 Apr 9, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions 03-H_Multi_GPU_Parallelization/.master/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,8 +24,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 03-H_Multi_GPU_Parallelization/.master/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions 03-H_Multi_GPU_Parallelization/solutions/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,8 +24,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
4 changes: 2 additions & 2 deletions 03-H_Multi_GPU_Parallelization/solutions/advanced/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,8 +24,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 03-H_Multi_GPU_Parallelization/solutions/advanced/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
2 changes: 1 addition & 1 deletion 03-H_Multi_GPU_Parallelization/solutions/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions 03-H_Multi_GPU_Parallelization/tasks/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,8 +24,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME) -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 03-H_Multi_GPU_Parallelization/tasks/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
6 changes: 3 additions & 3 deletions 06-H_Overlap_Communication_and_Computation_MPI/.master/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,9 +22,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 06-H_Overlap_Communication_and_Computation_MPI/.master/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
6 changes: 3 additions & 3 deletions 06-H_Overlap_Communication_and_Computation_MPI/solutions/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,9 +22,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 06-H_Overlap_Communication_and_Computation_MPI/solutions/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
6 changes: 3 additions & 3 deletions 06-H_Overlap_Communication_and_Computation_MPI/tasks/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,9 +22,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 06-H_Overlap_Communication_and_Computation_MPI/tasks/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -56,7 +56,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
8 changes: 4 additions & 4 deletions 08-H_NCCL_NVSHMEM/.master/NCCL/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# Copyright (c) 2021-2024, NVIDIA CORPORATION. All rights reserved.
THIS_TASK := 08H-NCCL-@@TASKSOL@@
OUTPUT_NAME := jacobi.$(THIS_TASK)__$(shell date '+%Y%m%d-%H%M')
NP ?= 1
NP ?= 4
NVCC=nvcc
JSC_SUBMIT_CMD ?= srun --cpu-bind=socket --gres=gpu:4 --ntasks-per-node 4
MPICXX=mpicxx
Expand All @@ -23,9 +23,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt -lnccl
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnccl
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/.master/NCCL/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions 08-H_NCCL_NVSHMEM/.master/NVSHMEM/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/.master/NVSHMEM/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -61,7 +61,7 @@
#endif // HAVE_CUB

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
8 changes: 4 additions & 4 deletions 08-H_NCCL_NVSHMEM/solutions/NCCL/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# Copyright (c) 2021-2024, NVIDIA CORPORATION. All rights reserved.
THIS_TASK := 08H-NCCL-sol
OUTPUT_NAME := jacobi.$(THIS_TASK)__$(shell date '+%Y%m%d-%H%M')
NP ?= 1
NP ?= 4
NVCC=nvcc
JSC_SUBMIT_CMD ?= srun --cpu-bind=socket --gres=gpu:4 --ntasks-per-node 4
MPICXX=mpicxx
Expand All @@ -23,9 +23,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt -lnccl
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnccl
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/solutions/NCCL/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions 08-H_NCCL_NVSHMEM/solutions/NVSHMEM/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/solutions/NVSHMEM/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,7 +59,7 @@
#endif // HAVE_CUB

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
8 changes: 4 additions & 4 deletions 08-H_NCCL_NVSHMEM/tasks/NCCL/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
# Copyright (c) 2021-2024, NVIDIA CORPORATION. All rights reserved.
THIS_TASK := 08H-NCCL-task
OUTPUT_NAME := jacobi.$(THIS_TASK)__$(shell date '+%Y%m%d-%H%M')
NP ?= 1
NP ?= 4
NVCC=nvcc
JSC_SUBMIT_CMD ?= srun --cpu-bind=socket --gres=gpu:4 --ntasks-per-node 4
MPICXX=mpicxx
Expand All @@ -23,9 +23,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt -lnccl
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17 -I$(CUDA_HOME)/include
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnccl
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/tasks/NCCL/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions 08-H_NCCL_NVSHMEM/tasks/NVSHMEM/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,8 +29,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include -I$(CUDA_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion 08-H_NCCL_NVSHMEM/tasks/NVSHMEM/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,7 +57,7 @@
#endif // HAVE_CUB

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions ...ommunication_with_NVSHMEM/.master/Device-initiated_Communication_with_NVSHMEM/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion ..._Communication_with_NVSHMEM/.master/Device-initiated_Communication_with_NVSHMEM/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,7 @@
#endif // HAVE_CUB

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
6 changes: 3 additions & 3 deletions ...phs_and_Device-initiated_Communication_with_NVSHMEM/.master/Using_CUDA_Graphs/Makefile.in
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,9 +23,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt -lnccl
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnccl
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion ...aphs_and_Device-initiated_Communication_with_NVSHMEM/.master/Using_CUDA_Graphs/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions ...Communication_with_NVSHMEM/solutions/Device-initiated_Communication_with_NVSHMEM/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
2 changes: 1 addition & 1 deletion ...ommunication_with_NVSHMEM/solutions/Device-initiated_Communication_with_NVSHMEM/jacobi.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,7 +60,7 @@
#endif // HAVE_CUB

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
6 changes: 3 additions & 3 deletions ...aphs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Using_CUDA_Graphs/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,9 +23,9 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++14
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++14
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnvToolsExt -lnccl
NVCC_FLAGS += -lineinfo $(GENCODE_FLAGS) -std=c++17
MPICXX_FLAGS = -DUSE_NVTX -I$(CUDA_HOME)/include -I$(NCCL_HOME)/include -std=c++17
LD_FLAGS = -L$(CUDA_HOME)/lib64 -lcudart -lnccl
jacobi: Makefile jacobi.cpp jacobi_kernels.o
$(MPICXX) $(MPICXX_FLAGS) jacobi.cpp jacobi_kernels.o $(LD_FLAGS) -o jacobi

Expand Down
2 changes: 1 addition & 1 deletion ...hs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Using_CUDA_Graphs/jacobi.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,7 +51,7 @@
#include <cuda_runtime.h>

#ifdef USE_NVTX
#include <nvToolsExt.h>
#include <nvtx3/nvToolsExt.h>

const uint32_t colors[] = {0x0000ff00, 0x000000ff, 0x00ffff00, 0x00ff00ff,
0x0000ffff, 0x00ff0000, 0x00ffffff};
Expand Down
4 changes: 2 additions & 2 deletions ...ted_Communication_with_NVSHMEM/tasks/Device-initiated_Communication_with_NVSHMEM/Makefile
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,8 +28,8 @@ ifdef DISABLE_CUB
else
NVCC_FLAGS = -DHAVE_CUB
endif
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++14 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvToolsExt -lnvidia-ml
NVCC_FLAGS += -dc -Xcompiler -fopenmp -lineinfo -DUSE_NVTX -lnvToolsExt $(GENCODE_FLAGS) -std=c++17 -I$(NVSHMEM_HOME)/include -I$(MPI_HOME)/include
NVCC_LDFLAGS = -ccbin=mpic++ -L$(NVSHMEM_HOME)/lib -lnvshmem_host -lnvshmem_device -L$(MPI_HOME)/lib -lmpi -L$(CUDA_HOME)/lib64 -lcuda -lcudart -lnvidia-ml
jacobi: Makefile jacobi.cu
$(NVCC) $(NVCC_FLAGS) jacobi.cu -c -o jacobi.o
$(NVCC) $(GENCODE_FLAGS) jacobi.o -o jacobi $(NVCC_LDFLAGS)
Expand Down
Loading