I need to use Platform MPI 9.1.3 with my application being run in -IBAL mode.
Unfortuantlly no driver released later than WinOF v2.1.2 ( 3.xx , 4.xx , 5.xx ) won’t be detected by MPI and it fires error like this :
Rank 0:1: MPI_Init: didn’t find active interface/port
Rank 0:1: MPI_Init: Can’t initialize RDMA device
Rank 0:1: MPI_Init: Internal Error: Cannot initialize RDMA protocol
if -IBAL switch is removed, TCP will be utilized and the program works but infiniband is not implemented.
Please let me know what’s the reason beyond this. Why Mellanox released so many drivers so far which are not compatible with Platform MPI ?
Am I missed a tweak ?
On Windows 10 Build 1607 with MS-MPI 8 and WinOF 5.35 and Firmware 2.42.5, everything was fine on two ConnectX-3 cards.
After applying 1709 upgrade for Windows 10, no more Network Direct connection could be made.
MS-MPI only uses Sockets as communication method and if Sockets are being disabled in MS-MPI, no more connection could be made.
I have also upgrade MS-MPI 8 to 9 but no change.
I have installed WinOF 5.40 but no change.
Is there any compatibility issue with Mellanox and Windows 10 1709 ?
Problem solved by a fresh installation of Windows 10 Pro Build 1803 and Mellanox 5.50 Driver and MS-MPI 9.0.1
Correct. It is only MS-MPI and not Platform-MPI that that Mellanox WinOF supports since v2.1 and the reason is stemming from Microsoft compatibility requirements.