Commit d42c04df authored by Brandon's avatar Brandon Committed by Kelly L. Rowland

change cpu_bind to cpu-bind

parent 889419dd
......@@ -76,7 +76,7 @@ functional calculations:
export OMP_PROC_BIND=spread
module load espresso/6.1
srun --cpu_bind=cores pw.x -nbgrp 8 -input test.in
srun --cpu-bind=cores pw.x -nbgrp 8 -input test.in
```
!!! example "Cori KNL"
......@@ -95,7 +95,7 @@ functional calculations:
export OMP_PROC_BIND=spread
module load espresso/6.1
srun --cpu_bind=cores pw.x -nbgrp 8 -input test.in
srun --cpu-bind=cores pw.x -nbgrp 8 -input test.in
```
!!! tip
......
......@@ -130,4 +130,4 @@ to enable "Remote Resize" in the NX menu:
#### Emacs complains about missing fonts and shows white blocks
This is due to a problem with font server. Please use the following
command instead: `emacs -font 7x14`
\ No newline at end of file
command instead: `emacs -font 7x14`
......@@ -339,4 +339,4 @@ source directory or space to write in the target directory. One common
mistake is to make the files readable, but forget to make the
directory holding them readable. You can check directory permissions
with ```ls -ld```. At NERSC you can make sure you have enough space to
write in a directory by using the ```myquota``` command.
\ No newline at end of file
write in a directory by using the ```myquota``` command.
......@@ -63,9 +63,9 @@ illustrative, not accurate).
1. If #MPI tasks per node is not a divisor of 64 on Haswell (meaning
the node is not fully packed), need to add an srun flag
"--cpu_bind=cores". Add "--cpu_bind=threads" instead if
"--cpu-bind=cores". Add "--cpu-bind=threads" instead if
#MPI_per_node > 32. In most cases for KNL, when use only 64 cores
out of 68 cores, --cpu_bind is needed.
out of 68 cores, --cpu-bind is needed.
1. Set `OMP_NUM_THREADS` environment variable to number of OpenMP
threads per MPI task.
......@@ -103,7 +103,7 @@ and OpenMP threads as what your application will use, and check if the
desired binding is obtained. The Cori binaries can be used to check
for both Haswell or KNL, since binaries are compatible.
Alternatively an srun flag `--cpu_bind=verbose` can be added to report process and thread binding.
Alternatively an srun flag `--cpu-bind=verbose` can be added to report process and thread binding.
Or you can set the following run time environment to obtain affinity
information as part of the job stdout:
......
......@@ -8,4 +8,4 @@ Table of content:
static directory allowing to use Shifter-based local DB to serve
the DB tables to root4star jobs on Cori/Edison
2. Example of interactive use of Shifter at PDSF and Cori, uses CHOS-SL6.4 image, CVMFS, features Atlas simulation job, runs for any PDSF user
\ No newline at end of file
2. Example of interactive use of Shifter at PDSF and Cori, uses CHOS-SL6.4 image, CVMFS, features Atlas simulation job, runs for any PDSF user
......@@ -235,7 +235,7 @@ Slurm Batch Script
#SBATCH -N 1
#SBATCH -t 00:30:00
#SBATCH -C haswell
srun -n 16 -c 4 --cpu_bind=cores --multi-prog –k mpmd_16.conf
srun -n 16 -c 4 --cpu-bind=cores --multi-prog –k mpmd_16.conf
```
mpmd_16.conf
```
......
......@@ -20,7 +20,7 @@ With core specialization, we increased the 32 nodes HDF5 collective IO bandwidth
In case node is not fully packed, process affinity is important in balancing workload and leveraging memory locality.
```shell
nersc$ srun -n 4 -c 64 --cpu_bind=cores
nersc$ srun -n 4 -c 64 --cpu-bind=cores
```
"That does the trick. The time for reading the WRF restart file is now 36 seconds, (was 300 seconds)" --John Michalakes, UCAR
......
# Intel Trace Analyzer and Collector
\ No newline at end of file
# Intel Trace Analyzer and Collector
......@@ -341,4 +341,4 @@ The following results were obtained by executing on Edison for 4 MPI tasks.
* [Cray Reveal Training](http://www.nersc.gov/users/training/events/cray-reveal-tool-training-sept-18-2014/),
presented by Heidi Paxon, September 2014
* [Lecture Slides](http://www.nersc.gov/assets/Uploads/Day-1B-01-Cray-PE-workshop-Reveal-NERSC.pdf)
for talk on Cray Reveal by Heidi Paxon, March 2016
\ No newline at end of file
for talk on Cray Reveal by Heidi Paxon, March 2016
......@@ -161,4 +161,4 @@ There are several [tools](openmp-tools.md) available at NERSC that are useful
for tuning OpenMP codes.
The [Performance and Debugging Tools](../../performance-debugging-tools/index.md)
page also shows other tools that can be useful for OpenMP codes.
\ No newline at end of file
page also shows other tools that can be useful for OpenMP codes.
# False Sharing Detection using VTune Amplifier
\ No newline at end of file
# False Sharing Detection using VTune Amplifier
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment