BinAC/Software/Alphafold - Revision history

F Bartusch at 08:46, 22 April 2024

2024-04-22T08:46:31Z

← Older revision		Revision as of 10:46, 22 April 2024
Line 71:		Line 71:
	</pre>		</pre>

			<!-- This is a comment

	= Benchmark on BinAC =		= Benchmark on BinAC =

Line 87:		Line 87:
	\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ...		\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ...
	\|}		\|}
			-->

F Bartusch at 11:55, 18 April 2024

2024-04-18T11:55:39Z

← Older revision		Revision as of 13:55, 18 April 2024
Line 64:		Line 64:

	Alphafold's module contains two example jobscripts. One computes the MSA on CPU followed by inference on GPU. The other example used a precomputed MSA and runs inference on the GPU.		Alphafold's module contains two example jobscripts. One computes the MSA on CPU followed by inference on GPU. The other example used a precomputed MSA and runs inference on the GPU.

			<pre>
			module load bio/alphafold/2.3.2
			qsub ${ALPHAFOLD_EXA_DIR}/binac-alphafold-2.3.2-bwhpc-examples.pbs
			qsub ${ALPHAFOLD_EXA_DIR}/binac-alphafold-2.3.2-bwhpc-examples_precomputed_msa.pbs
			</pre>


	= Benchmark on BinAC =		= Benchmark on BinAC =

F Bartusch at 11:52, 18 April 2024

2024-04-18T11:52:30Z

← Older revision		Revision as of 13:52, 18 April 2024
Line 62:		Line 62:

	= Example on BinAC =		= Example on BinAC =

			Alphafold's module contains two example jobscripts. One computes the MSA on CPU followed by inference on GPU. The other example used a precomputed MSA and runs inference on the GPU.

	= Benchmark on BinAC =		= Benchmark on BinAC =

F Bartusch at 10:11, 20 February 2024

2024-02-20T10:11:56Z

← Older revision		Revision as of 12:11, 20 February 2024
Line 64:		Line 64:

	= Benchmark on BinAC =		= Benchmark on BinAC =

			We ran some CASP14 targets with the <code>--benchmark=true</code> on BinAC. The following table gives you some guidance for choosing meaningful memory and walltime values.

			{\| class="wikitable" style="margin:auto"
			\|+ Benchmark results on BinAC (work in progress)
			\|-
			! Target !! #Residues !! jackhmmer UniRef90 [s] !! jackhmmer MGnify [s] !! HHblits on BFD [s] !! Inference [s] !! Memory Usage [GB]
			\|-
			\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ...
			\|-
			\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ...
			\|-
			\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ... \|\| ...
			\|}

F Bartusch at 09:52, 20 February 2024

2024-02-20T09:52:03Z

← Older revision		Revision as of 11:52, 20 February 2024
Line 24:		Line 24:
	= Usage =		= Usage =

	The BinAC provides Alphafold via an Apptainer Container. Both, the container and the AlphaFold database is stored on the WORK filesystem.		The BinAC provides Alphafold via an Apptainer Container. Both, the container and the AlphaFold database is stored on the <code>WORK</code> filesystem.
	The module `bio/alphafold` provides a wrapper script called <code>alphafold</code>.		The module <code>bio/alphafold</code> provides a wrapper script called <code>alphafold</code>.
	Upon loading the module, the wrapper <code>alphafold</code> is in <code>PATH</code> and can be directly used.		Upon loading the module, the wrapper <code>alphafold</code> is in <code>PATH</code> and can be directly used.
	The wrapper behaves like the script used in DeepMind's AlphaFold GitHub repository. Thus, all options explained ~~there~~ are also applicable for our <code>alphafold</code> wrapper.		The wrapper behaves like the script used in DeepMind's AlphaFold GitHub repository. Thus, all options explained in DeepMind's AlphaFold GitHub repository are also applicable for our <code>alphafold</code> wrapper.

	= Parallel ~~computing~~ =		= Parallel Computing =

			AlphaFold's algorithm results in two optimal resource profile regarding the number of cores and GPUs, depending on the way you run AlphaFold.
⚫	~~The~~ ~~first~~ ~~steps of~~ AlphaFold ~~compute~~ three multiple sequence alignments (MSA).
			The memory requirement depends on protein size.

			== Compute MSAs ==

		⚫	In the beginning, AlphaFold computes three multiple sequence alignments (MSA).
	These MSAs are computed on the CPU sequentially and the number of threads are hard-coded:		These MSAs are computed on the CPU sequentially and the number of threads are hard-coded:

	jackhmmer on UniRef90 using 8 threads</b>		jackhmmer on UniRef90 using 8 threads</br>
	jackhmmer on MGnify using 8 threads</b>		jackhmmer on MGnify using 8 threads</br>
	HHblits on BFD + Uniclust30 using 4 threads</b>		HHblits on BFD + Uniclust30 using 4 threads</br>

			Aftere computing the MSAs, AlphaFold then performs model inference on the GPU. Only one GPU is used.
	Thus, it does not make sense to use more than 8 cores in your job!
			This use case has this optimal resource profile:
	The additional cores will be idle.

			<pre>
			#PBS -l nodes=1:ppn=8:gpus=1
			</pre>

	The three MSAs are stored in the directory specified by <code>--output_dir</code> and can be reused with <code>--use_precomputed_msas=true</code>.		The three MSAs are stored in the directory specified by <code>--output_dir</code> and can be reused with <code>--use_precomputed_msas=true</code>.

			== Use existing MSAs ==
	Only the model inference will run on the GPUs.

			There is a switch (<code>--use_precomputed_msas=true</code>) that lets you use MSAs that were computed by an earlier AlphaFold run.
			As AlphaFold skips the computation of the MSAs. The model inference step will run on only one GPU. Thus the optimal resource profile is;

			<pre>
			#PBS -l nodes=1:ppn=1:gpus=1
			</pre>

			= Example on BinAC =

			= Benchmark on BinAC =

F Bartusch at 13:27, 16 February 2024

2024-02-16T13:27:02Z

@@ Line 15: / Line 15: @@
 |-
 | Links
-| Deepmind Alphafold Website: [https://deepmind.google/technologies/alphafold/]
+| DeepMind AlphaFold Website: [https://deepmind.google/technologies/alphafold/]
 |}
 = Description =
 = Parallel computing =

F Bartusch: Created page with "{{Softwarepage|bio/alphafold}} {| width=600px class="wikitable" |- ! Description !! Content |- | module load | bio/alphafold |- | License | Apache License 2.0 - see [https://..."

2024-02-16T12:57:54Z

Created page with "{{Softwarepage|bio/alphafold}} {| width=600px class="wikitable" |- ! Description !! Content |- | module load | bio/alphafold |- | License | Apache License 2.0 - see [https://..."

New page

{{Softwarepage|bio/alphafold}}

{| width=600px class="wikitable"
|-
! Description !! Content
|-
| module load
| bio/alphafold
|-
| License
| Apache License 2.0 - see [https://github.com/google-deepmind/alphafold/blob/main/LICENSE]
|-
| Citing
| See [https://github.com/google-deepmind/alphafold?tab=readme-ov-file#citing-this-work]
|-
| Links
| Deepmind Alphafold Website: [https://deepmind.google/technologies/alphafold/]
|}

= Description =

<br>
<br>

= Parallel computing =

= Usage =