Nsight-Compute-CLI-Sapelo2 - Revision history

Shtsai at 21:21, 14 May 2026

2026-05-14T21:21:23Z

← Older revision		Revision as of 17:21, 14 May 2026
Line 75:		Line 75:

	<pre class="gcomment">		<pre class="gcomment">
	A100: ra4-2, b6-3, b8-4		A100: ra4-1, ra4-2, b6-1, b6-2, b6-3, b6-4, b7-1, b7-2,b7-4, b8-4

	H100: ~~ra8~~-3, ra7-2, ra8-4		H100: ra5-2, ra5-3, ra7-2, ra8-3, ra8-4
	</pre>		</pre>

Shtsai at 01:07, 5 April 2026

2026-04-05T01:07:26Z

← Older revision		Revision as of 21:07, 4 April 2026
Line 1:		Line 1:
	[[Category:Sapelo2]][[Category:Software]][[Category:Other]]		[[Category:Sapelo2]][[Category:Software]][[Category:Other]]

	=== Category ===		== Category ==

	Other		Other

	=== Program On ===		== Program On ==

	Sapelo2		Sapelo2

	=== Version ===		== Version ==

	2023.1.1.0, 2024.1.0.0, 2024.3.0.0, 2025.1.0.0, 2025.4.0.0		2023.1.1.0, 2024.1.0.0, 2024.3.0.0, 2025.1.0.0, 2025.4.0.0

	=== Author / Distributor ===		== Author / Distributor ==

	NVIDIA		NVIDIA

	=== Description ===		== Description ==

	From https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html: "NVIDIA Nsight Compute CLI (ncu) provides a non-interactive way to profile applications from the command line. "		From https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html: "NVIDIA Nsight Compute CLI (ncu) provides a non-interactive way to profile applications from the command line. "

	=== Running Program ===		== Running Program ==

	Also refer to [[Running Jobs on Sapelo2]]		Also refer to [[Running Jobs on Sapelo2]]
Line 95:		Line 95:
	</pre>		</pre>

	=== Documentation ===		== Documentation ==

	Please see https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html		Please see https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html

	=== Installation ===		== Installation ==

	Downloaded from NVIDIA site.		Downloaded from NVIDIA site.

	=== System ===		== System ==

	64-bit Linux		64-bit Linux

Shtsai at 16:55, 20 February 2026

2026-02-20T16:55:37Z

← Older revision		Revision as of 12:55, 20 February 2026
Line 75:		Line 75:

	<pre class="gcomment">		<pre class="gcomment">
	A100: ra4-2, b6-3		A100: ra4-2, b6-3, b8-4

	H100: ra8-3, ra7-2, ra8-4		H100: ra8-3, ra7-2, ra8-4

Shtsai: /* Running Program */

2026-02-11T14:41:00Z

Running Program

← Older revision		Revision as of 10:41, 11 February 2026
Line 75:		Line 75:

	<pre class="gcomment">		<pre class="gcomment">
	A100: ra4-2		A100: ra4-2, b6-3

	H100: ra8-3, ra7-2, ra8-4		H100: ra8-3, ra7-2, ra8-4
Line 94:		Line 94:
	#SBATCH --nodelist=ra8-3		#SBATCH --nodelist=ra8-3
	</pre>		</pre>


	=== Documentation ===		=== Documentation ===

Shtsai at 18:49, 3 February 2026

2026-02-03T18:49:32Z

← Older revision		Revision as of 14:49, 3 February 2026
Line 63:		Line 63:


	Please note:		'''Please note:'''

	1. Before running ncu to profile an application, please set the environment variable TMPDIR to point to a directory of yours. For example, use		'''1.''' Before running ncu to profile an application, please set the environment variable TMPDIR to point to a directory of yours. For example, use
	<pre class="gcommand">		<pre class="gcommand">
	export TMPDIR=/scratch/$USER		export TMPDIR=/scratch/$USER
	</pre>		</pre>

	2. To profile applications on the H100 devices, please use the ncu in CUDA version 12.6.0 or higher. The older versions of ncu do not support the Hopper architecture.		'''2.''' To profile applications on the H100 devices, please use the ncu in CUDA version 12.6.0 or higher. The older versions of ncu do not support the Hopper architecture.

	3. Currently the ncu command only works on the following nodes:		'''3.''' Currently the ncu command only works on the following nodes:

			<pre class="gcomment">
	A100: ra4-2		A100: ra4-2

	H100: ra8-3, ra7-2, ra8-4		H100: ra8-3, ra7-2, ra8-4
			</pre>

	More nodes will be added to the list above, as ncu is enabled on them (it requires a node reboot). Please check back, if of interest.		More nodes will be added to the list above, as ncu is enabled on them (it requires a node reboot). Please check back, if of interest.

Moses: /* Running Program */

2026-02-03T18:43:07Z

Running Program

← Older revision		Revision as of 14:43, 3 February 2026
Line 25:		Line 25:
	Also refer to [[Running Jobs on Sapelo2]]		Also refer to [[Running Jobs on Sapelo2]]

	Compute nodes equipped with GPU cards have access to all applications installed in ~~<code>~~/apps~~</code>~~.		Compute nodes equipped with GPU cards have access to all applications installed in /apps.

	* Version 2025.4.0.0 is installed with CUDA v. 13.1.0 in /apps/eb/CUDA/13.1.0		* Version 2025.4.0.0 is installed with CUDA v. 13.1.0 in /apps/eb/CUDA/13.1.0

Moses: /* Running Program */

2026-02-03T18:41:33Z

Running Program

← Older revision		Revision as of 14:41, 3 February 2026
Line 25:		Line 25:
	Also refer to [[Running Jobs on Sapelo2]]		Also refer to [[Running Jobs on Sapelo2]]

	~~The compute~~ nodes ~~that are connected to the~~ GPU cards ~~can~~ access all ~~the~~ applications installed in /apps.		Compute nodes equipped with GPU cards have access to all applications installed in <code>/apps</code>.

	* Version 2025.4.0.0 is installed with CUDA v. 13.1.0 in /apps/eb/CUDA/13.1.0		* Version 2025.4.0.0 is installed with CUDA v. 13.1.0 in /apps/eb/CUDA/13.1.0

Shtsai at 16:53, 3 February 2026

2026-02-03T16:53:57Z

← Older revision		Revision as of 12:53, 3 February 2026
Line 75:		Line 75:

	A100: ra4-2		A100: ra4-2

	H100: ra8-3, ra7-2, ra8-4		H100: ra8-3, ra7-2, ra8-4

	More nodes will be added to the list above, as ncu is enabled on them (it requires a node reboot). Please check back, if of interest.		More nodes will be added to the list above, as ncu is enabled on them (it requires a node reboot). Please check back, if of interest.


	You can request a specific node with the Slurm <code>--nodelist</code> option. For example, <code>--nodelist=ra8-3 </code>.		You can request a specific node with the Slurm <code>--nodelist</code> option. For example, <code>--nodelist=ra8-3 </code>.

Shtsai at 16:45, 3 February 2026

2026-02-03T16:45:34Z

← Older revision		Revision as of 12:45, 3 February 2026
Line 54:		Line 54:
	module load CUDA/12.4.0		module load CUDA/12.4.0
	</pre>		</pre>


	* Version 2023.1.1.0 is installed with CUDA v. 12.1.1 in /apps/eb/CUDA/12.1.1		* Version 2023.1.1.0 is installed with CUDA v. 12.1.1 in /apps/eb/CUDA/12.1.1

Shtsai: Created page with "Category:Sapelo2Category:SoftwareCategory:Other === Category === Other === Program On === Sapelo2 === Version === 2023.1.1.0, 2024.1.0.0, 2024.3.0.0, 2025.1.0.0, 2025.4.0.0 === Author / Distributor === NVIDIA === Description === From https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html: "NVIDIA Nsight Compute CLI (ncu) provides a non-interactive way to profile applications from the command line. " === Running Program === Also ref..."

2026-02-03T16:40:10Z

Created page with "Category:Sapelo2 Category:Software Category:Other === Category === Other === Program On === Sapelo2 === Version === 2023.1.1.0, 2024.1.0.0, 2024.3.0.0, 2025.1.0.0, 2025.4.0.0 === Author / Distributor === NVIDIA === Description === From https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html: "NVIDIA Nsight Compute CLI (ncu) provides a non-interactive way to profile applications from the command line. " === Running Program === Also ref..."

New page

[[Category:Sapelo2]][[Category:Software]][[Category:Other]]

=== Category ===

Other

=== Program On ===

Sapelo2

=== Version ===

2023.1.1.0, 2024.1.0.0, 2024.3.0.0, 2025.1.0.0, 2025.4.0.0

=== Author / Distributor ===

NVIDIA

=== Description ===

From https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html: "NVIDIA Nsight Compute CLI (ncu) provides a non-interactive way to profile applications from the command line. "

=== Running Program ===

Also refer to [[Running Jobs on Sapelo2]]

The compute nodes that are connected to the GPU cards can access all the applications installed in /apps.

* Version 2025.4.0.0 is installed with CUDA v. 13.1.0 in /apps/eb/CUDA/13.1.0

To use it, please first load the module with:
<pre class="gcommand">
module load CUDA/13.1.0
</pre>

* Version 2025.1.0.0 is installed with CUDA v. 12.8.0 in /apps/eb/CUDA/12.8.0

To use it, please first load the module with:
<pre class="gcommand">
module load CUDA/12.8.0
</pre>

* Version 2024.3.0.0 is installed with CUDA v. 12.6.0 in /apps/eb/CUDA/12.6.0

To use it, please first load the module with:
<pre class="gcommand">
module load CUDA/12.6.0
</pre>

* Version 2024.1.0.0 is installed with CUDA v. 12.4.0 in /apps/eb/CUDA/12.4.0

To use it, please first load the module with:
<pre class="gcommand">
module load CUDA/12.4.0
</pre>

* Version 2023.1.1.0 is installed with CUDA v. 12.1.1 in /apps/eb/CUDA/12.1.1

To use it, please first load the module with:
<pre class="gcommand">
module load CUDA/12.1.1
</pre>

Please note:

1. Before running ncu to profile an application, please set the environment variable TMPDIR to point to a directory of yours. For example, use
<pre class="gcommand">
export TMPDIR=/scratch/$USER
</pre>

2. To profile applications on the H100 devices, please use the ncu in CUDA version 12.6.0 or higher. The older versions of ncu do not support the Hopper architecture.

3. Currently the ncu command only works on the following nodes:

A100: ra4-2
H100: ra8-3, ra7-2, ra8-4

More nodes will be added to the list above, as ncu is enabled on them (it requires a node reboot). Please check back, if of interest.

You can request a specific node with the Slurm <code>--nodelist</code> option. For example, <code>--nodelist=ra8-3 </code>.

'''Sample command to request an interactive job on a specific node:'''
<pre class="gcommand">
interact -p gpu_p --gres=gpu:H100:1 --nodelist=ra8-3
</pre>

'''Sample Slurm header line to request a specific node for a batch job:'''
<pre class="gcommand">
#SBATCH --nodelist=ra8-3
</pre>

=== Documentation ===

Please see https://docs.nvidia.com/nsight-compute/NsightComputeCli/index.html

=== Installation ===

Downloaded from NVIDIA site.

=== System ===

64-bit Linux