Submissions ID: 8f442200-3a90-4bc7-b1c5-fb83d8caf302
MIT/MGHPCC Holyoke, MA
Actual Site Name MIT/MGHPCC Holyoke, MA
Site URL https://researchcomputing.mit.edu
Acronym MIT
Segment Academic
City Holyoke
Country United States of America
Satori
Installation Date Sept. 24, 2019
System URL https://researchcomputing.mit.edu/satori
Applications Area Research
Vendor IBM
Component 1: Satori
Model AC922
Nodes 64
Vendor IBM
Processor IBM Power 9 20C 2.4GHz
Speed 2400
Sockets per Node: 2
Cores per Socket: 20
Accelerator/CP: Tesla V100-SMX2-32GB
Accelerators/CP per Node: 4
Cores per Accelerator/CP: 80
Operating System: RHEL 7.6
Primary Interconnect: Infiniband EDR
Peak Power (kW): 105.04
Size of Power Measurement (Cores) 23040
Memory per Node (GB) 1024
Summary of all components
CPU Cores 2560
Accelerators/CP 256
Accelerator/CP Cores
Memory 65536 GB
Power 105.04 kW
Linpack Benchmark
Benchmark Date Oct. 22, 2019
CPU Cores 2560
Accelerator/CP Cores 20480
Rmax 1562.00 GFlop/s
Rpeak 1996.00 GFlop/s
Nmax 2555904
Nhalf None
HPL input
HPLinpack benchmark input file
Innovative Computing Laboratory, University of Tennessee
HPL.out output file name (if any)
6 device out (6=stdout,7=stderr,file)
1 # of problems sizes (N)
2555904 2588160 110592 622080 1244160 2488320 2592768 1769472 1543680 311040 Ns
1 # of NBs
768 NBs
0 PMAP process mapping (0=Row-,1=Column-major)
1 # of process grids (P x Q)
16 4 8 16 Ps
16 2 8 16 Qs
16.0 threshold
1 # of panel fact
2 2 2 PFACTs (0=left, 1=Crout, 2=Right)
1 # of recursive stopping criterium
4 4 4 NBMINs (>= 1)
1 # of panels in recursion
2 NDIVs
1 # of recursive panel fact.
0 0 0 RFACTs (0=left, 1=Crout, 2=Right)
1 # of broadcast
3 1 BCASTs (0=1rg,1=1rM,2=2rg,3=2rM,4=Lng,5=LnM)
1 # of lookahead depth
0 DEPTHs (>=0)
1 SWAP (0=bin-exch,1=long,2=mix)
768 swapping threshold
1 L1 in (0=transposed,1=no-transposed) form
0 U in (0=transposed,1=no-transposed) form
0 Equilibration (0=no,1=yes)
8 memory alignment in double (> 0)
Power information
- in power measurements below (line beginning Power Comments Samples of average power every 180 ) the fields are "date time","timestamp","power" where power is in Watts. So typical power for this benchmark was 106KJ/s.
Software
Compiler NVidia Binary
Compiler Options used
Math Library
MPI Library IBM Spectrum MPI
Power
Power Level 2
Power Comments Samples of average power every 180 secs from Revenue Grade Meter. CSV of measurements is shown below. Timestamp,Average Power(W) 2019-10-22 10:46:00 1571755560,106440 2019-10-22 10:49:00 1571755740,106284 2019-10-22 10:52:00 1571755920,106487 2019-10-22 10:55:00 1571756100,106540 2019-10-22 10:58:00 1571756280,106532 2019-10-22 11:01:00 1571756460,105599 2019-10-22 11:04:00 1571756640,105794 2019-10-22 11:07:00 1571756820,106009 2019-10-22 11:10:00 1571757000,105446 2019-10-22 11:13:00 1571757180,106094 2019-10-22 11:16:00 1571757360,106203 2019-10-22 11:19:00 1571757540,105630 2019-10-22 11:22:00 1571757720,105494 2019-10-22 11:25:00 1571757900,106194 2019-10-22 11:28:00 1571758080,106003 2019-10-22 11:31:00 1571758260,104041 2019-10-22 11:34:00 1571758440,92724 2019-10-22 11:37:00 1571758620,105255 2019-10-22 11:40:00 1571758800,105573 2019-10-22 11:43:00 1571758980,104891 2019-10-22 11:46:00 1571759160,103697 2019-10-22 11:49:00 1571759340,105667 2019-10-22 11:52:00 1571759520,105582 2019-10-22 11:55:00 1571759700,105619 2019-10-22 11:58:00 1571759880,105150 2019-10-22 12:01:00 1571760060,104102 2019-10-22 12:04:00 1571760240,105092 2019-10-22 12:07:00 1571760420,104927 2019-10-22 12:10:00 1571760600,105066 2019-10-22 12:13:00 1571760780,105407 2019-10-22 12:16:00 1571760960,105334 2019-10-22 12:19:00 1571761140,105127 2019-10-22 12:22:00 1571761320,105350 2019-10-22 12:25:00 1571761500,104933 2019-10-22 12:28:00 1571761680,105444 2019-10-22 12:31:00 1571761860,102916 2019-10-22 12:34:00 1571762040,104365 2019-10-22 12:37:00 1571762220,104614
Peak Power 105.00
Measured Cores 23040
Measure Device Measurelogic DTS310 Revenue Grade Meters
Measure Date Oct. 22, 2019
Cooling on False
Storage on False
HPL output