Learning Forum

[David Pomeroy]

Is it possible to specify the memory mode from the command line?

I am new to Mechanical APDL and am using it to run workstation and cluster performance benchmarks (MAPDL_2022R1_Benchmark) on Linux virtual machine clusters using the command line. When I run the benchmarks in distributed memory mode, I often get a warning that there is insufficient in-core memory allocated even though the hardware has enough memory to run fully in-core. I found some suggestions to use the command "DSPOPTION,,INCORE". Is it possible to set this option at the command line without editing the benchmark .dat file, or are there other options I can use to force it to use more physical memory?

The command I am using is like this:

ansys -b -i {path}/V22direct-1.dat -np 88 -dis -mpi INTELMPI -machines=$(cat machine_file)

[mrife]

Hi David

Warning are often ok - MAPDL has dynamic memory allocation so the warning about in-core might have been just a warning before MAPDL allocated itself more memory.  Or it may have actually run out-of-core.  Check the output file.

See the MAPDL Help Operations Guide chapter 4.1 on the command line options.  There are two for memory that work hand-in-hand.  -m and -db

Don't try to allocation all the available memory to MAPDL if the model does not require it - the OS will use memory for things like caching files for I/O etc.  

Mike

[David Pomeroy]

No, it definitely didn't fix the issue itself because I'm getting huge amounts of I/O and elapsed time exceeds CPU time by a significant amount.

I've tried using the -m option but it is not working as expected. To test this out, I'm running on one node with 88 cores, 4GB physical memory per core for a total of 352GB. I tried -m 256000 which should allocate a total of 250GB of memory, much less than the physical amount in the node, but what actually happened is Total memory allocated for solver = 490.634 GB and then the run failed because it exceeded physical memory. So when the manual says that -m is the total workspace allocation, it doesn't seem to be functioning that way. I tried adjusting the value downward:

-m 128000 still allocated the same 490.6GB

-m 64000 allocated 477GB

-m 32000 allocated 373GB

-m 16000 allocated 166GB 

So it does seem to affect the allocation, but in a nonlinear way. I haven't determined what multiplier or algorithm it's using. I tried -m 3072 to see if it allocates memory per core, but then it only allocated 163GB and was no better than if I didn't set -m.

Any other ideas?

[mrife]

 

David

Well remember the command line -m is the initial allocation; MAPDL still can do dynamic allocation from there.  Are you a commercial customer (I'd like to see the output file)?  Also the value can be set to a negative number to force using a specific amount (and so turning off dynamic allocation).  The -db size, which can also grow dynamically, is taken out of the -m amount.