Removed CTS1 to archive folder, fixed plots. Added OSU benchmarks to …

…microbenchmarks folder.
lanl · Sep 27, 2023 · 3450d53 · 3450d53
1 parent bfadbcf
commit 3450d53
Show file tree

Hide file tree

Showing 237 changed files with 149,639 additions and 24 deletions.
diff --git a/doc/sphinx/10_microbenchmarks/M1_STREAM/STREAM.rst b/doc/sphinx/10_microbenchmarks/M1_STREAM/STREAM.rst
@@ -126,9 +126,12 @@ Replace `<num_processes>` with the number of MPI processes you want to use. For
 
 Example Results
 ===============
+Results from Branson are provided on the following systems:
 
-ATS-3 Rocinante HBM
--------------------
+* Crossroads (see :ref:`GlobalSystemATS3`)
+
+Crossroads
+----------
 
 .. csv-table:: STREAM microbenchmark bandwidth measurement
    :file: stream-xrds_ats5cce-cray-mpich.csv
@@ -140,18 +143,3 @@ ATS-3 Rocinante HBM
    :align: center
    :scale: 50%
    :alt: STREAM microbenchmark bandwidth measurement
-
-CTS-1 Snow
------------
-
-.. csv-table:: STREAM microbenchmark bandwidth measurement
-   :file: stream_cts1.csv
-   :align: center
-   :widths: 10, 10, 10
-   :header-rows: 1
-
-.. figure:: stream_cpu_cts1.png
-   :align: center
-   :scale: 50%
-   :alt: STREAM microbenchmark bandwidth measurement
-
diff --git a/doc/sphinx/10_microbenchmarks/M1_STREAM/cpu.gp b/doc/sphinx/10_microbenchmarks/M1_STREAM/cpu.gp
@@ -3,7 +3,7 @@
 set terminal pngcairo enhanced size 1024, 768 dashed font 'Helvetica,18'
 set output "stream_cpu_cts1.png"
 
-set title "STREAM Single node bandwidth" font "serif,22"
+# set title "STREAM Single node bandwidth" font "serif,22"
 set ylabel "Per core triad BW (MB/s)"
 set y2label "FOM: Total triad BW (MB/s)"
 

diff --git a/doc/sphinx/10_microbenchmarks/M3_DGEMM/DGEMM.rst b/doc/sphinx/10_microbenchmarks/M3_DGEMM/DGEMM.rst
@@ -82,8 +82,12 @@ These are positional arguments, so, for instance, R cannot be set without settin
 Example Results
 ===============
 
-ATS-3 Rocinante HBM
--------------------
+Results from Branson are provided on the following systems:
+
+* Crossroads (see :ref:`GlobalSystemATS3`)
+
+Crossroads
+----------
 
 This test was built with the intel 2023.1.0 compiler using the crayOS compiler wrapper where: :math:`N=2500, R=500, \alpha=1.0, \beta=1.0`. The 110 core run (cores are used as OpenMP threads) avoids the OS dedicated cores and takes roughly an hour. All four runs on rocinante hbm take 5-6 hours.
 

diff --git a/doc/sphinx/10_microbenchmarks/M3_DGEMM/cpu.gp b/doc/sphinx/10_microbenchmarks/M3_DGEMM/cpu.gp
@@ -2,7 +2,7 @@
 set terminal pngcairo enhanced size 1024, 768 dashed font 'Helvetica,18'
 set output "dgemm_cts1.png"
 
-set title " Single node Dgemm" font "serif,22"
+# set title " Single node Dgemm" font "serif,22"
 set xlabel "No. Processing Elements"
 set ylabel "Figure of Merit (GFlops)"
 
@@ -24,7 +24,7 @@ set style line 2 linetype 1 dashtype 2 linecolor rgb "#FF0000" linewidth 2
 #plot "dgemm_cts1.csv" using 1:2 with linespoints linestyle 1, "" using 1:3 with line linestyle 2
 
 set output "dgemm_ats3.png"
-set xrange [4:128]
+set xrange [32:128]
 plot "dgemm_ats3.csv" using 1:2 with linespoints linestyle 1
 
 
diff --git a/doc/sphinx/10_microbenchmarks/M4_IOR/IOR.rst b/doc/sphinx/10_microbenchmarks/M4_IOR/IOR.rst
@@ -55,6 +55,8 @@ Input
 
 Sample production input files in `microbenchmarks/ior/inputs.xroads`
 
+Running IOR does not require using the input files. All arguments can be given on the command line.
+
 Example Results
 ===============
 
diff --git a/doc/sphinx/10_microbenchmarks/M4_IOR/cpu.gp b/doc/sphinx/10_microbenchmarks/M4_IOR/cpu.gp
@@ -2,7 +2,7 @@
 set terminal pngcairo enhanced size 1024, 768 dashed font 'Helvetica,18'
 set output "dgemm_cts1.png"
 
-set title " Single node Dgemm" font "serif,22"
+# set title " Single node Dgemm" font "serif,22"
 set xlabel "No. Processing Elements"
 set ylabel "Figure of Merit (GFlops)"
 
@@ -24,7 +24,7 @@ set style line 2 linetype 1 dashtype 2 linecolor rgb "#FF0000" linewidth 2
 #plot "dgemm_cts1.csv" using 1:2 with linespoints linestyle 1, "" using 1:3 with line linestyle 2
 
 set output "dgemm_ats3.png"
-set xrange [4:128]
+set xrange [32:128]
 plot "dgemm_ats3.csv" using 1:2 with linespoints linestyle 1
 
 
diff --git a/doc/sphinx/archive/cts1/stream/cpu.gp b/doc/sphinx/archive/cts1/stream/cpu.gp
@@ -0,0 +1,33 @@
+#!/usr/bin/gnuplot
+#STREAM
+set terminal pngcairo enhanced size 1024, 768 dashed font 'Helvetica,18'
+set output "stream_cpu_cts1.png"
+
+# set title "STREAM Single node bandwidth" font "serif,22"
+set ylabel "Per core triad BW (MB/s)"
+set y2label "FOM: Total triad BW (MB/s)"
+
+set xrange [1:40]
+#set yrange [3000:15000]
+
+# set logscale x 2
+set logscale y 2
+
+set grid
+show grid
+set key left top
+
+set datafile separator comma
+set key autotitle columnheader
+
+set style line 1 linetype 6 dashtype 1 linecolor rgb "#FF0000" linewidth 2 pointtype 6 pointsize 3
+set style line 2 linetype 1 dashtype 2 linecolor rgb "#FF0000" linewidth 2
+
+plot "stream_cts1.csv" using 1:2 with linespoints linestyle 1 axis x1y1, "" using 1:3 with line linestyle 2 axis x1y2
+
+set output "stream_cpu_ats3.png"
+set xrange [4:115]
+plot "stream-xrds_ats5cce-cray-mpich.csv" using 1:2 with linespoints linestyle 1 axis x1y1, "" using 1:3 with line linestyle 2 axis x1y2
+
+
+
diff --git a/doc/sphinx/archive/cts1/stream/cpu_cts1.png b/doc/sphinx/archive/cts1/stream/cpu_cts1.png
diff --git a/doc/sphinx/archive/cts1/stream/stream_archive.rst b/doc/sphinx/archive/cts1/stream/stream_archive.rst
@@ -0,0 +1,14 @@
+CTS-1 Snow
+-----------
+
+.. csv-table:: STREAM microbenchmark bandwidth measurement
+   :file: stream_cts1.csv
+   :align: center
+   :widths: 10, 10, 10
+   :header-rows: 1
+
+.. figure:: stream_cpu_cts1.png
+   :align: center
+   :scale: 50%
+   :alt: STREAM microbenchmark bandwidth measurement
+
diff --git a/doc/sphinx/archive/cts1/stream/stream_cts1.csv b/doc/sphinx/archive/cts1/stream/stream_cts1.csv
@@ -0,0 +1,8 @@
+No. Cores,Per Core Bandwidth (MB/s),Total Bandwidth (MB/s)
+1,10690.1,10690.1
+2,10701.3,21402.6
+4,9316.5,37266.0
+8,7884.5,63076.0
+16,7747.5,123960.0
+32,5510.3,176329.6
+36,3189.2,114811.2
diff --git a/microbenchmarks/ior/Xrds_acceptance_script.sh b/microbenchmarks/ior/Xrds_acceptance_script.sh
@@ -0,0 +1,36 @@
+#!/bin/bash
+find /lustre/xrscratch1/${USER}/ior/ -type f -delete
+for taskPerNode in 1 8 32 56 88 110; do
+  segments=16
+  size=2G
+  if [ $numNodes -lt 3070 ];  then
+          taskPerNode=20
+  fi
+  if [ $numNodes -lt 614 ];  then
+          taskPerNode=50
+  fi
+        nTasks=$(( taskPerNode * numNodes ))
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -k -e -a POSIX -F -v  -b 4G -s 16 -t 1M -D 180 -w -o /lustre/xrscratch1/aparga/ior/${numNodes}_fpr_posix
+        sleep 3
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -k -e -a MPIIO -F -v  -b 4G -s 16 -t 1M -D 180 -w -o /lustre/xrscratch1/aparga/ior/${numNodes}_fpr_MPIIO
+        sleep 3
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -C -Q $taskPerNode -k -E -a POSIX -F -v  -b 4G -s 16 -t 1M -D 30 -r -o /lustre/xrscratch1/aparga/ior/${numNodes}_fpr_posix
+        sleep 3
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -C -Q $taskPerNode -k -E -a MPIIO -F -v  -b 4G -s 16 -t 4M -D 30 -r -o /lustre/xrscratch1/aparga/ior/${numNodes}_fpr_MPIIO
+        sleep 3
+
+
+        taskPerNode=10
+        nTasks=$(( taskPerNode * numNodes ))
+        lfs setstripe -c 4 /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_posix
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -k -e -E -a POSIX -v  -b $size -s $segments -t 1M -D 180 -w -o /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_posix
+        sleep 3
+        lfs setstripe -c 4 /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_MPIIO
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -k -e -E -a MPIIO -v  -b $size -s $segments -t 1M -D 180 -w -o /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_MPIIO
+        sleep 3
+
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -C -Q $taskPerNode -k -E -a POSIX -v  -b $size -s $segments -t 1M -D 45 -r -o /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_posix
+        sleep 3
+        srun -N $numNodes --ntasks=$nTasks /users/aparga/xr/bin/ior -C -Q $taskPerNode -k -E -a MPIIO -v  -b $size -s $segments -t 1M -D 45 -r -o /lustre/xrscratch1/aparga/ior/${numNodes}_nto1_MPIIO
+        sleep 3
+done