add new runs

Signed-off-by: vsoch <vsoch@users.noreply.github.com>

google/networking/netmark/README.md

@@ -83,8 +83,17 @@ written locally to see what was run. Note that I think for machine type, if we e
 project_id="$(gcloud config get-value core/project)"
 bucket="netmark-experiment-bucket"
 
-# 16 tasks on one instance
+# 16 tasks on one instance (of size 16, but I think this is only 8 cores?)
 $ python run-job.py ${project_id} --cpu-milli 1000 --memory 1000 --tasks 16 --max-run-duration 3600s --bucket ${bucket} --machine-type c2-standard-16 --job-name netmark-experiment-16-02 --netmark-store-trial --parallelism 1
+
+# 8 tasks on one instance (16 vCPU which is 8 cores)
+$ python run-job.py ${project_id} --cpu-milli 1000 --memory 1000 --tasks 8 --max-run-duration 3600s --bucket ${bucket} --machine-type c2-standard-16 --job-name netmark-experiment-16-03 --netmark-store-trial --parallelism 1
+
+# 56 tasks on 112 vCPU instance
+$ python run-job.py ${project_id} --cpu-milli 1000 --memory 1000 --tasks 56 --max-run-duration 3600s --bucket ${bucket} --machine-type c2d-standard-112 --job-name netmark-experiment-112-01 --netmark-store-trial --parallelism 1
+
+# 30 tasks on 60 vCPU instance
+$ python run-job.py ${project_id} --cpu-milli 1000 --memory 1000 --tasks 30 --max-run-duration 3600s --bucket ${bucket} --machine-type c2-standard-60 --job-name netmark-experiment-60-01 --netmark-store-trial --parallelism 1
 ```
 
 ## Feedback for Google Batch

google/networking/netmark/data/netmark-experiment-112-01/metadata.json

@@ -0,0 +1,24 @@
+{
+    "project_id": "llnl-flux",
+    "region": "us-central1",
+    "job_name": "netmark-experiment-112-01",
+    "bucket_name": "netmark-experiment-bucket",
+    "netmark_path": "netmark",
+    "outdir": "data",
+    "tasks": 56,
+    "tasks_per_node": 1,
+    "cpu_milli": 1000,
+    "memory": 1000,
+    "retry_count": 2,
+    "machine_type": "c2d-standard-112",
+    "image_family": "hpc-centos-7",
+    "image_project": "cloud-hpc-image-public",
+    "max_run_duration": "3600s",
+    "mount_path": "/mnt/share",
+    "warmup": 10,
+    "trials": 20,
+    "cycles": 100,
+    "message_size": 0,
+    "store_trial": true,
+    "parallelism": 1
+}

google/networking/netmark/data/netmark-experiment-112-01/netmark_wrapper.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+/mnt/share/netmark/netmark.x -w 10 -t 20 -c 100 -b 0  -s 

google/networking/netmark/data/netmark-experiment-112-01/run.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+source /opt/intel/mpi/latest/env/vars.sh
+
+# Ensure we have the wrapper template and it's correctly populated
+wrapper=/mnt/share/data/netmark-experiment-112-01/netmark_wrapper.sh
+cat ${wrapper}
+chmod +x ${wrapper}
+
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/data/netmark-experiment-112-01
+  ls
+  mpirun -n 56 -ppn 1 -f $BATCH_HOSTS_FILE ${wrapper}
+fi

google/networking/netmark/data/netmark-experiment-112-01/setup.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+export DEBIAN_FRONTEND=noninteractive
+sleep $BATCH_TASK_INDEX
+
+# Note that for this family / image, we are root (do not need sudo)
+yum update -y && yum install -y cmake gcc tuned ethtool python3
+
+# Ensure a python3 executable is found, if does not exist
+which python3 || (ln -s $(which python) /usr/bin/python3)
+
+# This ONLY works on the hpc-* image family images
+google_mpi_tuning --nosmt
+# google_install_mpi --intel_mpi
+google_install_intelmpi --impi_2021
+source /opt/intel/mpi/latest/env/vars.sh
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+outdir=/mnt/share/data/netmark-experiment-112-01
+mkdir -p $outdir
+find /opt/intel -name mpicc
+
+# Only have index 0 compile
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/netmark
+  ls
+  # And only compile if the executable does not exist!   
+  # Makefile content plus adding include directories
+  if [[ ! -f "netmark.x" ]]; then
+      mpicc -std=c99 -lmpi -lmpifort -O3 netmark.c -DTRACE -I/opt/intel/mpi/latest/include -I/opt/intel/mpi/2021.8.0/include -L/opt/intel/mpi/2021.8.0/lib/release -L/opt/intel/mpi/2021.8.0/lib -o netmark.x 
+   fi
+fi

google/networking/netmark/data/netmark-experiment-16-03/RTT.csv

@@ -0,0 +1,8 @@
+0.000,0.913,0.896,0.928,0.965,0.925,0.890,0.932
+0.971,0.000,0.914,0.903,0.938,0.973,0.894,0.935
+1.036,0.954,0.000,0.907,0.968,1.041,0.846,0.907
+1.050,0.979,0.860,0.000,0.955,1.055,0.870,0.891
+1.100,1.022,0.984,1.009,0.000,1.081,0.967,0.967
+1.178,1.095,1.053,1.051,1.119,0.000,1.024,1.045
+1.049,1.042,1.016,1.026,1.035,1.085,0.000,1.025
+1.030,1.050,1.102,1.018,1.001,1.015,1.033,0.000

google/networking/netmark/data/netmark-experiment-16-03/metadata.json

@@ -0,0 +1,24 @@
+{
+    "project_id": "llnl-flux",
+    "region": "us-central1",
+    "job_name": "netmark-experiment-16-03",
+    "bucket_name": "netmark-experiment-bucket",
+    "netmark_path": "netmark",
+    "outdir": "data",
+    "tasks": 8,
+    "tasks_per_node": 1,
+    "cpu_milli": 1000,
+    "memory": 1000,
+    "retry_count": 2,
+    "machine_type": "c2-standard-16",
+    "image_family": "hpc-centos-7",
+    "image_project": "cloud-hpc-image-public",
+    "max_run_duration": "3600s",
+    "mount_path": "/mnt/share",
+    "warmup": 10,
+    "trials": 20,
+    "cycles": 100,
+    "message_size": 0,
+    "store_trial": true,
+    "parallelism": 1
+}

google/networking/netmark/data/netmark-experiment-16-03/netmark_wrapper.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+/mnt/share/netmark/netmark.x -w 10 -t 20 -c 100 -b 0  -s 

google/networking/netmark/data/netmark-experiment-16-03/run.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+source /opt/intel/mpi/latest/env/vars.sh
+
+# Ensure we have the wrapper template and it's correctly populated
+wrapper=/mnt/share/data/netmark-experiment-16-03/netmark_wrapper.sh
+cat ${wrapper}
+chmod +x ${wrapper}
+
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/data/netmark-experiment-16-03
+  ls
+  mpirun -n 8 -ppn 1 -f $BATCH_HOSTS_FILE ${wrapper}
+fi

google/networking/netmark/data/netmark-experiment-16-03/setup.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+export DEBIAN_FRONTEND=noninteractive
+sleep $BATCH_TASK_INDEX
+
+# Note that for this family / image, we are root (do not need sudo)
+yum update -y && yum install -y cmake gcc tuned ethtool python3
+
+# Ensure a python3 executable is found, if does not exist
+which python3 || (ln -s $(which python) /usr/bin/python3)
+
+# This ONLY works on the hpc-* image family images
+google_mpi_tuning --nosmt
+# google_install_mpi --intel_mpi
+google_install_intelmpi --impi_2021
+source /opt/intel/mpi/latest/env/vars.sh
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+outdir=/mnt/share/data/netmark-experiment-16-03
+mkdir -p $outdir
+find /opt/intel -name mpicc
+
+# Only have index 0 compile
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/netmark
+  ls
+  # And only compile if the executable does not exist!   
+  # Makefile content plus adding include directories
+  if [[ ! -f "netmark.x" ]]; then
+      mpicc -std=c99 -lmpi -lmpifort -O3 netmark.c -DTRACE -I/opt/intel/mpi/latest/include -I/opt/intel/mpi/2021.8.0/include -L/opt/intel/mpi/2021.8.0/lib/release -L/opt/intel/mpi/2021.8.0/lib -o netmark.x 
+   fi
+fi

google/networking/netmark/data/netmark-experiment-60-01/RTT.csv

@@ -0,0 +1,30 @@
+0.000,1.358,1.091,1.159,1.130,1.100,1.169,1.103,1.123,1.154,1.093,1.134,1.090,1.138,1.151,1.407,1.383,1.404,1.410,1.452,1.487,1.444,1.428,1.400,1.404,1.394,1.415,1.417,1.481,1.366
+1.302,0.000,1.128,1.140,1.129,1.083,1.100,1.090,1.125,1.155,1.188,1.150,1.110,1.126,1.098,1.363,1.392,1.436,1.392,1.418,1.373,1.378,1.372,1.340,1.498,1.456,1.450,1.360,1.385,1.394
+1.092,1.238,0.000,1.020,0.915,0.901,0.933,1.117,0.977,0.888,0.909,0.936,0.943,1.008,0.898,1.291,1.201,1.281,1.258,1.284,1.201,1.225,1.240,1.277,1.209,1.221,1.239,1.232,1.200,1.228
+0.997,1.171,0.989,0.000,1.015,1.007,0.982,0.990,1.010,1.008,0.962,1.019,1.006,1.038,1.122,1.349,1.333,1.332,1.303,1.329,1.270,1.310,1.332,1.287,1.288,1.307,1.349,1.336,1.318,1.372
+0.932,1.089,0.913,0.959,0.000,0.855,0.887,0.925,0.954,0.878,0.900,0.917,0.941,0.918,1.069,1.227,1.265,1.244,1.271,1.258,1.206,1.235,1.236,1.299,1.211,1.224,1.213,1.225,1.243,1.231
+0.936,1.115,0.922,0.997,0.883,0.000,0.878,0.955,0.939,1.005,0.926,0.970,0.930,0.958,1.110,1.268,1.265,1.285,1.298,1.240,1.259,1.262,1.258,1.321,1.271,1.237,1.267,1.261,1.288,1.256
+0.922,1.106,0.934,0.980,0.885,0.874,0.000,0.948,0.983,1.016,0.893,0.944,0.961,0.955,1.092,1.259,1.241,1.301,1.268,1.271,1.215,1.228,1.238,1.261,1.232,1.247,1.242,1.226,1.238,1.251
+0.886,1.108,0.892,0.925,0.896,0.901,0.907,0.000,0.989,0.890,0.928,0.952,0.923,0.884,1.036,1.274,1.247,1.300,1.318,1.275,1.249,1.261,1.267,1.292,1.304,1.285,1.279,1.272,1.226,1.254
+0.891,1.112,1.008,0.893,0.909,0.912,0.933,0.888,0.000,0.917,1.073,0.995,0.947,0.932,0.870,1.252,1.270,1.426,1.280,1.252,1.254,1.270,1.262,1.288,1.226,1.248,1.255,1.273,1.273,1.267
+0.937,1.003,0.998,0.910,0.902,0.902,0.889,0.911,0.967,0.000,1.092,0.979,0.933,0.940,0.911,1.209,1.237,1.285,1.259,1.232,1.216,1.257,1.268,1.170,1.258,1.226,1.212,1.257,1.234,1.232
+1.108,1.072,1.094,1.066,1.082,1.130,1.083,1.069,1.202,1.096,0.000,1.123,1.093,1.109,1.082,1.353,1.409,1.436,1.393,1.359,1.402,1.429,1.411,1.332,1.417,1.404,1.421,1.427,1.391,1.368
+0.971,0.991,0.982,0.967,0.982,0.956,1.081,0.987,1.101,0.977,1.162,0.000,1.050,0.963,0.965,1.292,1.375,1.305,1.278,1.306,1.345,1.360,1.315,1.348,1.357,1.278,1.325,1.462,1.348,1.374
+1.006,1.064,0.924,0.882,0.904,0.887,0.887,0.885,0.933,0.882,1.085,0.991,0.000,0.943,0.882,1.258,1.270,1.286,1.366,1.271,1.246,1.315,1.262,1.220,1.211,1.216,1.219,1.238,1.231,1.275
+0.935,1.091,0.958,0.968,0.965,0.936,1.001,0.948,1.024,0.925,1.093,0.945,0.970,0.000,0.944,1.254,1.242,1.261,1.263,1.346,1.199,1.240,1.241,1.222,1.207,1.266,1.241,1.271,1.187,1.248
+0.870,1.087,0.926,0.892,0.898,0.893,0.895,0.958,0.901,0.910,1.057,1.045,0.875,0.932,0.000,1.217,1.225,1.290,1.330,1.263,1.224,1.282,1.232,1.236,1.278,1.260,1.238,1.228,1.226,1.254
+1.316,1.275,1.283,1.283,1.313,1.295,1.265,1.284,1.283,1.279,1.299,1.334,1.286,1.286,1.332,0.000,0.957,0.946,0.959,0.983,0.985,1.029,0.927,0.904,0.966,0.963,1.053,0.958,0.972,0.963
+1.475,1.427,1.392,1.409,1.470,1.458,1.419,1.415,1.417,1.410,1.376,1.509,1.456,1.436,1.415,1.214,0.000,1.182,1.199,1.200,1.216,1.206,1.216,1.222,1.202,1.153,1.168,1.179,1.197,1.204
+1.315,1.240,1.243,1.255,1.243,1.297,1.263,1.255,1.259,1.266,1.245,1.341,1.292,1.263,1.224,1.044,1.089,0.000,0.915,0.955,0.954,0.901,0.895,0.880,0.926,0.871,0.893,0.905,0.866,0.902
+1.272,1.206,1.218,1.185,1.181,1.214,1.252,1.216,1.230,1.269,1.211,1.317,1.244,1.209,1.281,1.042,1.084,0.880,0.000,0.971,0.935,0.881,0.847,0.897,0.879,0.855,0.894,0.894,0.857,0.884
+1.343,1.299,1.288,1.270,1.269,1.300,1.345,1.266,1.296,1.322,1.285,1.393,1.284,1.334,1.298,1.073,1.133,1.007,0.966,0.000,0.947,0.948,0.992,1.006,0.996,0.987,0.966,1.001,0.981,0.982
+1.229,1.215,1.222,1.227,1.208,1.266,1.261,1.196,1.225,1.216,1.176,1.319,1.203,1.269,1.206,1.044,1.107,0.834,0.871,0.856,0.000,0.884,0.857,0.882,0.931,0.875,0.890,0.895,0.874,0.892
+1.295,1.201,1.204,1.242,1.209,1.233,1.247,1.165,1.209,1.190,1.205,1.330,1.231,1.292,1.222,1.046,1.098,0.836,0.873,0.891,0.873,0.000,0.887,0.855,0.869,0.842,0.897,0.857,0.912,0.853
+1.293,1.228,1.219,1.241,1.218,1.289,1.290,1.226,1.217,1.310,1.236,1.373,1.239,1.273,1.209,1.036,1.093,0.898,0.884,0.898,0.883,0.897,0.000,0.855,0.894,0.895,0.872,0.888,0.895,0.899
+1.274,1.274,1.230,1.166,1.205,1.202,1.245,1.166,1.220,1.199,1.209,1.316,1.224,1.236,1.259,1.048,1.094,0.858,0.885,0.908,0.859,0.907,0.827,0.000,0.880,0.922,0.857,0.863,0.898,0.873
+1.207,1.317,1.209,1.220,1.180,1.259,1.217,1.196,1.235,1.195,1.222,1.310,1.217,1.242,1.246,1.052,1.086,0.908,0.832,0.850,0.914,0.848,0.885,0.856,0.000,0.895,0.895,0.909,0.891,0.888
+1.273,1.294,1.231,1.198,1.228,1.186,1.209,1.175,1.208,1.233,1.207,1.282,1.185,1.199,1.218,1.056,0.980,0.863,0.882,0.862,1.058,0.868,0.891,0.877,0.873,0.000,0.920,0.914,0.904,0.857
+1.270,1.234,1.249,1.206,1.233,1.238,1.236,1.203,1.229,1.228,1.272,1.342,1.237,1.298,1.261,1.075,0.908,0.886,0.851,0.858,1.079,0.863,0.896,0.906,0.877,0.892,0.000,0.936,0.871,0.878
+1.285,1.241,1.268,1.232,1.211,1.287,1.230,1.243,1.211,1.256,1.228,1.295,1.226,1.344,1.234,1.041,0.883,0.865,0.857,0.885,1.123,0.872,0.861,0.893,0.901,0.860,0.881,0.000,0.907,0.886
+1.267,1.223,1.245,1.306,1.230,1.262,1.207,1.240,1.217,1.196,1.197,1.331,1.242,1.209,1.250,1.047,0.890,0.852,0.875,0.891,1.082,0.878,0.897,0.878,0.914,0.885,0.866,0.872,0.000,0.869
+1.290,1.227,1.269,1.306,1.167,1.243,1.253,1.210,1.212,1.218,1.231,1.314,1.233,1.237,1.284,1.062,0.880,0.849,0.876,0.861,1.098,0.885,0.876,0.886,0.881,0.864,0.868,0.860,0.892,0.000

google/networking/netmark/data/netmark-experiment-60-01/metadata.json

@@ -0,0 +1,24 @@
+{
+    "project_id": "llnl-flux",
+    "region": "us-central1",
+    "job_name": "netmark-experiment-60-01",
+    "bucket_name": "netmark-experiment-bucket",
+    "netmark_path": "netmark",
+    "outdir": "data",
+    "tasks": 30,
+    "tasks_per_node": 1,
+    "cpu_milli": 1000,
+    "memory": 1000,
+    "retry_count": 2,
+    "machine_type": "c2-standard-60",
+    "image_family": "hpc-centos-7",
+    "image_project": "cloud-hpc-image-public",
+    "max_run_duration": "3600s",
+    "mount_path": "/mnt/share",
+    "warmup": 10,
+    "trials": 20,
+    "cycles": 100,
+    "message_size": 0,
+    "store_trial": true,
+    "parallelism": 1
+}

google/networking/netmark/data/netmark-experiment-60-01/netmark_wrapper.sh

@@ -0,0 +1,2 @@
+#!/bin/bash
+/mnt/share/netmark/netmark.x -w 10 -t 20 -c 100 -b 0  -s 

google/networking/netmark/data/netmark-experiment-60-01/run.sh

@@ -0,0 +1,15 @@
+#!/bin/bash
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+source /opt/intel/mpi/latest/env/vars.sh
+
+# Ensure we have the wrapper template and it's correctly populated
+wrapper=/mnt/share/data/netmark-experiment-60-01/netmark_wrapper.sh
+cat ${wrapper}
+chmod +x ${wrapper}
+
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/data/netmark-experiment-60-01
+  ls
+  mpirun -n 30 -ppn 1 -f $BATCH_HOSTS_FILE ${wrapper}
+fi

google/networking/netmark/data/netmark-experiment-60-01/setup.sh

@@ -0,0 +1,31 @@
+#!/bin/bash
+export DEBIAN_FRONTEND=noninteractive
+sleep $BATCH_TASK_INDEX
+
+# Note that for this family / image, we are root (do not need sudo)
+yum update -y && yum install -y cmake gcc tuned ethtool python3
+
+# Ensure a python3 executable is found, if does not exist
+which python3 || (ln -s $(which python) /usr/bin/python3)
+
+# This ONLY works on the hpc-* image family images
+google_mpi_tuning --nosmt
+# google_install_mpi --intel_mpi
+google_install_intelmpi --impi_2021
+source /opt/intel/mpi/latest/env/vars.sh
+export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/intel/mpi/latest/lib:/opt/intel/mpi/latest/lib/release
+export PATH=/opt/intel/mpi/latest/bin:$PATH
+outdir=/mnt/share/data/netmark-experiment-60-01
+mkdir -p $outdir
+find /opt/intel -name mpicc
+
+# Only have index 0 compile
+if [ $BATCH_TASK_INDEX = 0 ]; then
+  cd /mnt/share/netmark
+  ls
+  # And only compile if the executable does not exist!   
+  # Makefile content plus adding include directories
+  if [[ ! -f "netmark.x" ]]; then
+      mpicc -std=c99 -lmpi -lmpifort -O3 netmark.c -DTRACE -I/opt/intel/mpi/latest/include -I/opt/intel/mpi/2021.8.0/include -L/opt/intel/mpi/2021.8.0/lib/release -L/opt/intel/mpi/2021.8.0/lib -o netmark.x 
+   fi
+fi