Skip to content

Commit 4d09c5c

Browse files
jirikrausjirikraus
committed
Updates for 10H for ISC24 instance.
1 parent 3524792 commit 4d09c5c

File tree

13 files changed

+128
-91
lines changed

13 files changed

+128
-91
lines changed

‎.etc/modules.sh

Lines changed: 8 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -1,10 +1,11 @@
11
module purge
2-
module load Stages/2023
3-
module load GCC/11.3.0
4-
module load CUDA/11.7
5-
module load ParaStationMPI/5.7.1-1
6-
module load NCCL/default-CUDA-11.7
7-
module load NVSHMEM/2.8.0-CUDA-11.7
8-
module load Nsight-Systems/2023.2.1
2+
module use $OTHERSTAGES
3+
module load Stages/2024
4+
module load GCC/12.3.0
5+
module load CUDA/12
6+
module load ParaStationMPI/5.9.2-1
7+
module load NCCL/default-CUDA-12
8+
module load NVSHMEM/2.10.1-CUDA-12
9+
module load Nsight-Systems/2023.3.1
910
module load MPI-settings/CUDA
1011
# module use $JSCCOURSE_DIR_GROUP/common/modulefiles

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/.master/Device-initiated_Communication_with_NVSHMEM/Instructions.ipynb

Lines changed: 17 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -2,15 +2,17 @@
22
"cells": [
33
{
44
"cell_type": "markdown",
5+
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb",
56
"metadata": {},
67
"source": [
7-
"# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
8+
"# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
89
"\n",
9-
"- Time: Monday, 13 November 2023 8:30 - 17:00 MST\n",
10-
"- Location: 405, Denver Congress Centre, USA\n",
10+
"- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST\n",
11+
"- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany\n",
1112
"- Program Link:\n",
12-
" https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242\n",
13-
" \\## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
13+
" https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ== \n",
14+
"\n",
15+
"## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
1416
"\n",
1517
"### Task: Using NVSHMEM device API\n",
1618
"\n",
@@ -84,12 +86,15 @@
8486
" variant in the [Multi GPU Programming Models Github\n",
8587
" repository](https://github.com/NVIDIA/multi-gpu-programming-models)\n",
8688
" implements the same strategy."
87-
],
88-
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb"
89+
]
90+
},
91+
{
92+
"cell_type": "markdown",
93+
"id": "a19b1176",
94+
"metadata": {},
95+
"source": []
8996
}
9097
],
91-
"nbformat": 4,
92-
"nbformat_minor": 5,
9398
"metadata": {
9499
"kernelspec": {
95100
"display_name": "Python 3",
@@ -108,5 +113,7 @@
108113
"pygments_lexer": "ipython3",
109114
"version": "3.6.8"
110115
}
111-
}
116+
},
117+
"nbformat": 4,
118+
"nbformat_minor": 5
112119
}

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/.master/Device-initiated_Communication_with_NVSHMEM/Instructions.md

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
1-
# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale
1+
# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale
22

3-
- Time: Monday, 13 November 2023 8:30 - 17:00 MST
4-
- Location: 405, Denver Congress Centre, USA
3+
- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST
4+
- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany
55
- Program Link:
6-
https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242
6+
https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==
7+
78
## Hands-On 10A: Device-initiated Communication with NVSHMEM
89

910
### Task: Using NVSHMEM device API

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/.master/Using_CUDA_Graphs/Instructions.ipynb

Lines changed: 17 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -2,15 +2,17 @@
22
"cells": [
33
{
44
"cell_type": "markdown",
5+
"id": "bc28bb54-3dda-4b0a-9465-f6609051d0cc",
56
"metadata": {},
67
"source": [
7-
"# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
8+
"# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
89
"\n",
9-
"- Time: Monday, 13 November 2023 8:30 - 17:00 MST\n",
10-
"- Location: 405, Denver Congress Centre, USA\n",
10+
"- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST\n",
11+
"- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany\n",
1112
"- Program Link:\n",
12-
" https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242\n",
13-
" \\## Hands-On 10B: Using CUDA Graphs\n",
13+
" https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ== \n",
14+
"\n",
15+
"## Hands-On 10B: Using CUDA Graphs\n",
1416
"\n",
1517
"### Task: Combining CUDA Graphs with NCCL for Inter-GPU Communication\n",
1618
"\n",
@@ -63,12 +65,15 @@
6365
"Study the performance by glimpsing at the profile generated with\n",
6466
"`make profile`. For `make run` and `make profile` the environment\n",
6567
"variable `NP` can be set to change the number of processes."
66-
],
67-
"id": "bc28bb54-3dda-4b0a-9465-f6609051d0cc"
68+
]
69+
},
70+
{
71+
"cell_type": "markdown",
72+
"id": "0ce20fd8",
73+
"metadata": {},
74+
"source": []
6875
}
6976
],
70-
"nbformat": 4,
71-
"nbformat_minor": 5,
7277
"metadata": {
7378
"kernelspec": {
7479
"display_name": "Python 3",
@@ -87,5 +92,7 @@
8792
"pygments_lexer": "ipython3",
8893
"version": "3.6.8"
8994
}
90-
}
95+
},
96+
"nbformat": 4,
97+
"nbformat_minor": 5
9198
}

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/.master/Using_CUDA_Graphs/Instructions.md

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
1-
# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale
1+
# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale
22

3-
- Time: Monday, 13 November 2023 8:30 - 17:00 MST
4-
- Location: 405, Denver Congress Centre, USA
3+
- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST
4+
- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany
55
- Program Link:
6-
https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242
6+
https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==
7+
78
## Hands-On 10B: Using CUDA Graphs
89

910
### Task: Combining CUDA Graphs with NCCL for Inter-GPU Communication

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Device-initiated_Communication_with_NVSHMEM/Instructions.ipynb

Lines changed: 11 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -2,15 +2,17 @@
22
"cells": [
33
{
44
"cell_type": "markdown",
5+
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb",
56
"metadata": {},
67
"source": [
7-
"# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
8+
"# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
89
"\n",
9-
"- Time: Monday, 13 November 2023 8:30 - 17:00 MST\n",
10-
"- Location: 405, Denver Congress Centre, USA\n",
10+
"- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST\n",
11+
"- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany\n",
1112
"- Program Link:\n",
12-
" https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242\n",
13-
" \\## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
13+
" https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==\n",
14+
" \n",
15+
"## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
1416
"\n",
1517
"### Task: Using NVSHMEM device API\n",
1618
"\n",
@@ -84,12 +86,9 @@
8486
" variant in the [Multi GPU Programming Models Github\n",
8587
" repository](https://github.com/NVIDIA/multi-gpu-programming-models)\n",
8688
" implements the same strategy."
87-
],
88-
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb"
89+
]
8990
}
9091
],
91-
"nbformat": 4,
92-
"nbformat_minor": 5,
9392
"metadata": {
9493
"kernelspec": {
9594
"display_name": "Python 3",
@@ -108,5 +107,7 @@
108107
"pygments_lexer": "ipython3",
109108
"version": "3.6.8"
110109
}
111-
}
110+
},
111+
"nbformat": 4,
112+
"nbformat_minor": 5
112113
}

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Device-initiated_Communication_with_NVSHMEM/Instructions.md

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
1-
# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale
1+
# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale
22

3-
- Time: Monday, 13 November 2023 8:30 - 17:00 MST
4-
- Location: 405, Denver Congress Centre, USA
3+
- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST
4+
- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany
55
- Program Link:
6-
https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242
6+
https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==
7+
78
## Hands-On 10A: Device-initiated Communication with NVSHMEM
89

910
### Task: Using NVSHMEM device API

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Using_CUDA_Graphs/Instructions.ipynb

Lines changed: 17 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -2,15 +2,17 @@
22
"cells": [
33
{
44
"cell_type": "markdown",
5+
"id": "bc28bb54-3dda-4b0a-9465-f6609051d0cc",
56
"metadata": {},
67
"source": [
7-
"# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
8+
"# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
89
"\n",
9-
"- Time: Monday, 13 November 2023 8:30 - 17:00 MST\n",
10-
"- Location: 405, Denver Congress Centre, USA\n",
10+
"- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST\n",
11+
"- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany\n",
1112
"- Program Link:\n",
12-
" https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242\n",
13-
" \\## Hands-On 10B: Using CUDA Graphs\n",
13+
" https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==\n",
14+
"\n",
15+
"## Hands-On 10B: Using CUDA Graphs\n",
1416
"\n",
1517
"### Task: Combining CUDA Graphs with NCCL for Inter-GPU Communication\n",
1618
"\n",
@@ -63,12 +65,15 @@
6365
"Study the performance by glimpsing at the profile generated with\n",
6466
"`make profile`. For `make run` and `make profile` the environment\n",
6567
"variable `NP` can be set to change the number of processes."
66-
],
67-
"id": "bc28bb54-3dda-4b0a-9465-f6609051d0cc"
68+
]
69+
},
70+
{
71+
"cell_type": "markdown",
72+
"id": "60dfd2ee",
73+
"metadata": {},
74+
"source": []
6875
}
6976
],
70-
"nbformat": 4,
71-
"nbformat_minor": 5,
7277
"metadata": {
7378
"kernelspec": {
7479
"display_name": "Python 3",
@@ -87,5 +92,7 @@
8792
"pygments_lexer": "ipython3",
8893
"version": "3.6.8"
8994
}
90-
}
95+
},
96+
"nbformat": 4,
97+
"nbformat_minor": 5
9198
}

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/solutions/Using_CUDA_Graphs/Instructions.md

Lines changed: 5 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,10 @@
1-
# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale
1+
# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale
22

3-
- Time: Monday, 13 November 2023 8:30 - 17:00 MST
4-
- Location: 405, Denver Congress Centre, USA
3+
- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST
4+
- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany
55
- Program Link:
6-
https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242
6+
https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==
7+
78
## Hands-On 10B: Using CUDA Graphs
89

910
### Task: Combining CUDA Graphs with NCCL for Inter-GPU Communication

‎10-H_CUDA_Graphs_and_Device-initiated_Communication_with_NVSHMEM/tasks/Device-initiated_Communication_with_NVSHMEM/Instructions.ipynb

Lines changed: 11 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -2,15 +2,17 @@
22
"cells": [
33
{
44
"cell_type": "markdown",
5+
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb",
56
"metadata": {},
67
"source": [
7-
"# SC23 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
8+
"# ISC24 Tutorial: Efficient Distributed GPU Programming for Exascale\n",
89
"\n",
9-
"- Time: Monday, 13 November 2023 8:30 - 17:00 MST\n",
10-
"- Location: 405, Denver Congress Centre, USA\n",
10+
"- Time: Sunday, May 12, 2024 9:00 AM to 6:00 PM CEST\n",
11+
"- Location: Hall Y8 - 2nd floor, Congress Center Hamburg, Germany\n",
1112
"- Program Link:\n",
12-
" https://sc23.supercomputing.org/presentation/?id=tut140&sess=sess242\n",
13-
" \\## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
13+
" https://app.swapcard.com/widget/event/isc-high-performance-2024/planning/UGxhbm5pbmdfMTgyNTY0MQ==\n",
14+
" \n",
15+
"## Hands-On 10A: Device-initiated Communication with NVSHMEM\n",
1416
"\n",
1517
"### Task: Using NVSHMEM device API\n",
1618
"\n",
@@ -84,12 +86,9 @@
8486
" variant in the [Multi GPU Programming Models Github\n",
8587
" repository](https://github.com/NVIDIA/multi-gpu-programming-models)\n",
8688
" implements the same strategy."
87-
],
88-
"id": "3d1de1bf-0ebb-4709-bddc-9e34484b03bb"
89+
]
8990
}
9091
],
91-
"nbformat": 4,
92-
"nbformat_minor": 5,
9392
"metadata": {
9493
"kernelspec": {
9594
"display_name": "Python 3",
@@ -108,5 +107,7 @@
108107
"pygments_lexer": "ipython3",
109108
"version": "3.6.8"
110109
}
111-
}
110+
},
111+
"nbformat": 4,
112+
"nbformat_minor": 5
112113
}

0 commit comments

Comments
 (0)