 |
 |
 |
|
|
|
[Sponsors] | ||||
Problem in parallel processing [Process affinity not being set] |
 |
|
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
November 27, 2017, 04:08
|
Problem in parallel processing [Process affinity not being set]
|
#1 |
|
Senior Member
Join Date: Sep 2017
Posts: 130
Rep Power: 9  |
Hey guys
I'm trying to do some parallel processing using Ansys fluent 18 but I have a problem. When I set processs to e.g. 35 I get this error: Code:
hp-reep is already loaded (92.9758%). Process affinity not being set. Code:
Host spawning Node 0 on machine "hp-reep" (win64).
Host 0 -- ip 192.178.26.82 -- ranks 0 - 34
host | 0
======|======
0 : SHM
Prot - All Intra-node communication is: SHM
-------------------------------------------------------------------------------
ID Hostname Core O.S. PID Vendor
-------------------------------------------------------------------------------
n34 hp-reep 35/2 Windows-x64 8968 Intel(R) Xeon(R) E5-2699
n33 hp-reep 34/2 Windows-x64 3396 Intel(R) Xeon(R) E5-2699
n32 hp-reep 33/2 Windows-x64 4968 Intel(R) Xeon(R) E5-2699
n31 hp-reep 32/2 Windows-x64 6080 Intel(R) Xeon(R) E5-2699
n30 hp-reep 31/2 Windows-x64 11000 Intel(R) Xeon(R) E5-2699
n29 hp-reep 30/2 Windows-x64 260 Intel(R) Xeon(R) E5-2699
n28 hp-reep 29/2 Windows-x64 4340 Intel(R) Xeon(R) E5-2699
n27 hp-reep 28/2 Windows-x64 3496 Intel(R) Xeon(R) E5-2699
n26 hp-reep 27/2 Windows-x64 11072 Intel(R) Xeon(R) E5-2699
n25 hp-reep 26/2 Windows-x64 480 Intel(R) Xeon(R) E5-2699
n24 hp-reep 25/2 Windows-x64 8740 Intel(R) Xeon(R) E5-2699
n23 hp-reep 24/2 Windows-x64 8704 Intel(R) Xeon(R) E5-2699
n22 hp-reep 23/2 Windows-x64 828 Intel(R) Xeon(R) E5-2699
n21 hp-reep 22/2 Windows-x64 5800 Intel(R) Xeon(R) E5-2699
n20 hp-reep 21/2 Windows-x64 8376 Intel(R) Xeon(R) E5-2699
n19 hp-reep 20/2 Windows-x64 8280 Intel(R) Xeon(R) E5-2699
n18 hp-reep 19/2 Windows-x64 8224 Intel(R) Xeon(R) E5-2699
n17 hp-reep 18/2 Windows-x64 6528 Intel(R) Xeon(R) E5-2699
n16 hp-reep 17/2 Windows-x64 8916 Intel(R) Xeon(R) E5-2699
n15 hp-reep 16/2 Windows-x64 8176 Intel(R) Xeon(R) E5-2699
-------------------------------------------------------------------------------
ID Hostname Core O.S. PID Vendor
-------------------------------------------------------------------------------
n14 hp-reep 15/2 Windows-x64 8980 Intel(R) Xeon(R) E5-2699
n13 hp-reep 14/2 Windows-x64 1460 Intel(R) Xeon(R) E5-2699
n12 hp-reep 13/2 Windows-x64 9164 Intel(R) Xeon(R) E5-2699
n11 hp-reep 12/2 Windows-x64 9540 Intel(R) Xeon(R) E5-2699
n10 hp-reep 11/2 Windows-x64 9336 Intel(R) Xeon(R) E5-2699
n9 hp-reep 10/2 Windows-x64 4388 Intel(R) Xeon(R) E5-2699
n8 hp-reep 9/2 Windows-x64 5512 Intel(R) Xeon(R) E5-2699
n7 hp-reep 8/2 Windows-x64 6156 Intel(R) Xeon(R) E5-2699
n6 hp-reep 7/2 Windows-x64 8428 Intel(R) Xeon(R) E5-2699
n5 hp-reep 6/2 Windows-x64 10028 Intel(R) Xeon(R) E5-2699
n4 hp-reep 5/2 Windows-x64 9192 Intel(R) Xeon(R) E5-2699
n3 hp-reep 4/2 Windows-x64 9328 Intel(R) Xeon(R) E5-2699
n2 hp-reep 3/2 Windows-x64 9412 Intel(R) Xeon(R) E5-2699
n1 hp-reep 2/2 Windows-x64 9512 Intel(R) Xeon(R) E5-2699
n0* hp-reep 1/2 Windows-x64 8512 Intel(R) Xeon(R) E5-2699
host hp-reep Windows-x64 8316 Intel(R) Xeon(R) E5-2699
MPI Option Selected: ibmmpi
Selected system interconnect: default
-------------------------------------------------------------------------------
Cleanup script file is C:\Users\Administrator\Desktop\Aerospike-mesh-quad-1-2-new\cleanup-fluent-hp-reep-8316.bat
>
hp-reep is already loaded (92.9758%). Process affinity not being set.
Reading "\"| gunzip -c \"C:\Users\Administrator\Desktop\Aerospike-mesh-quad-1-2-new\Aerospike_files\dp0\FLU-1\Fluent\SYS.1-Setup-Output.cas.gz\"\""...
Buffering for file scan...
500500 quadrilateral cells, zone 3, binary.
995095 2D interior faces, zone 1, binary.
2804 2D interior faces, zone 2, binary.
2716 2D pressure-outlet faces, zone 6, binary.
100 2D velocity-inlet faces, zone 7, binary.
1680 2D wall faces, zone 8, binary.
1706 2D axis faces, zone 9, binary.
503602 nodes, binary.
503602 node flags, binary.
Building...
mesh
auto partitioning mesh by Metis (fast),
distributing mesh
parts...................................,
faces...................................,
nodes...................................,
cells...................................,
bandwidth reduction using Reverse Cuthill-McKee: 14350/138 = 103.986
materials,
interface,
domains,
mixture
zones,
axis
wall
inlet
outlet
surface_body
interior-surface_body
solid-surface_bodySetting solid-surface_body (mixture) ... Done.
Setting zone id of solid-surface_body to 3.
Setting zone id of interior-surface_body to 1.
Setting zone id of surface_body to 2.
Setting zone id of outlet to 6.
Setting zone id of inlet to 7.
Setting zone id of wall to 8.
Setting zone id of axis to 9.
Done.
Setting solid-surface_body (mixture) ... Done.
Setting inlet (mixture) (zone type changed to pressure-inlet) ... Done.
Setting interior-surface_body (mixture) ... Done.
Setting surface_body (mixture) ... Done.
Setting outlet (mixture) ... Done.
Setting wall (mixture) ... Done.
Setting axis (mixture) ... Done.
warning: for compressible (ideal and real) gas models with buoyancy,
it is recommended to use a specified operating density value of zero.
GUI access: Cell Zone/Boundary Conditions -> Operating Conditions...
TUI access: define operating-conditions operating-density?
parallel,
Done.
Preparing mesh for display...
Done.
Setting Post Processing and Surfaces information ... Done.
By the way, when I run the calculation, it works as serial processing. I already ran calculation on this device for this project with parallel processing and everything was fine but now it doesn't work. What's the problem? Edit: just tried 2 procces and again got this error. |
|
|
|
|
|
October 3, 2018, 17:57
|
|
#2 |
|
New Member
Join Date: Sep 2018
Posts: 2
Rep Power: 0 |
I've got the same problem
Replying only to get this post up |
|
|
|
|
|
|
October 4, 2018, 13:14
|
|
#3 |
|
Senior Member
Lucky
Join Date: Apr 2011
Location: Orlando, FL USA
Posts: 5,762
Rep Power: 66 |
It is a warning message. It's a job scheduling problem and there's almost nothing you can do about it from inside Fluent. It basically means your cpu is busy doing other things and will not prioritize your Fluent processes. Fluent will be running in the background compared to whatever else is being done by your system.
What you can do, is stop running other programs. Figure out what is utilizing the cpu and shut it down, or run on a fewer number of available cores. This message is a hint: Code:
hostname is already loaded (##%). Last edited by LuckyTran; October 5, 2018 at 11:12. |
|
|
|
|
|
|
October 4, 2018, 19:10
|
|
#4 |
|
New Member
Join Date: Sep 2018
Posts: 2
Rep Power: 0 |
Thank you!
I suspected this should be the problem, but I needed someone who really knows this. I've read some thread about this, turned off hyperthreading as suggested and now barely got this message. |
|
|
|
|
|
|
October 26, 2023, 04:42
|
|
#5 | |
|
Member
Join Date: Oct 2022
Posts: 63
Rep Power: 4 |
Quote:
I use an Intel processor that has 24 physical cores. When I use the number of cores for example 14 or 16 or... in order to parallelize and reduce the solution time, my solution goes slowly. What should I do ? Thankyou |
||
|
|
|
||
|
|
|
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Problem running OpenFOAM 2.2.x in parallel in Centos 5 | lvalvare | OpenFOAM Running, Solving & CFD | 33 | December 9, 2020 09:44 |
| Temperature Problem in Parallel Processing | pdp.aero | SU2 | 9 | July 27, 2014 23:11 |
| problem with parallel processing | keng | FLUENT | 4 | June 15, 2013 00:34 |
| problem parallel processing (meshSearch, openMPI) | FabienF | OpenFOAM | 0 | September 10, 2012 05:23 |
| About parallel processing in Linux | tuks | CFX | 10 | August 8, 2005 09:22 |
Linear Mode
