|
[Sponsors] |
Running Fluent & CFX directly using MS MPI on Windows? |
|
LinkBack | Thread Tools | Search this Thread | Display Modes |
January 5, 2020, 16:14 |
Running Fluent & CFX directly using MS MPI on Windows?
|
#1 |
Member
Join Date: Jul 2011
Posts: 53
Rep Power: 15 |
Hi,
Anyone using MS MPI to run CFX or Fluent in distributed parallel in a Windows environment *without* using the Microsoft HPC Pack job scheduler? I'd like to use MS MPI as a drop down option in Fluent launcher and/or CFX Solver Manager. Is there a way to edit the properties of the MPI hook that gets called by the launcher/solver manager at the beginning of the run? MS MPI seems to be the only way to run via native infiniband on current versions of Windows as its the only MPI to currently support RDMA/native infiniband with Mellanox WinOF-2 driver. |
|
January 7, 2020, 14:39 |
|
#2 |
Senior Member
Erik
Join Date: Feb 2011
Location: Earth (Land portion)
Posts: 1,188
Rep Power: 23 |
Which Driver are you using? This one?
MLNX_WinOF_VPI_2_1_2_win7_x64 I use this driver with Windows 7 and IBM Platform MPI, using Infinihost III 20Gbps cards. Every driver after this one was released looks like it only supports MS-MPI for windows, like you say. I have not yet tested it on Windows 10. I will have to in a week here though, as we are being forced to switch to Win10. When I spoke with ANSYS a long time ago, I was informed they used these same drivers with windows 10, so I have a some hope it should work. |
|
January 8, 2020, 03:02 |
|
#3 |
Member
Join Date: Jul 2011
Posts: 53
Rep Power: 15 |
I'm using ConnectX-5 and ConnectX-6 cards, so I'm on the newer WinOF-2 software.
Ansys Support is adamanent that Fluent can only be run in distributed mode using MSMPI in conjunction with the Microsoft HPC scheduler. Still trying to get CFX to work. |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Running UDF with Supercomputer | roi247 | FLUENT | 4 | October 15, 2015 14:41 |
CFX overwhelming Fluent in mass convergence of boundary layer separation case | Pierre1 | FLUENT | 7 | March 26, 2015 22:43 |
Convergence problem of CFX when comparing with FLUENT with same mesh | guxin7005 | CFX | 8 | May 22, 2014 16:13 |
Different result in CFX and fluent for mass trans.? is segregated better? | ftab | CFX | 7 | September 27, 2012 08:57 |
High Resolution (CFX) vs 2nd Order Upwind (Fluent) | gravis | ANSYS | 3 | March 24, 2011 03:43 |