You can print several reports of memory and processor usage from the Parallel ribbon tab.
You can print out a table summarizing the processor usage on each machine that has compute processes spawned for the current session by clicking CPU Info in the Parallel ribbon tab (System group box). For example, if processes have been spawned on three machines, a table similar to the following is displayed:
and selecting--------------------------------------------------------------------------------------- | CPU | System Mem (GB) Hostname | Sock x Core Clock (MHz) Load | Total Available --------------------------------------------------------------------------------------- host23 | 8 x 8, HT 2399.85 0.07 | 129.039 89.565 host24 | 8 x 8, HT 2399.95 1.08 | 129.039 32.516 host25 | 8 x 8, HT 2400.4 3.14 | 129.039 62.531 --------------------------------------------------------------------------------------- Total | 192 - - | 387.117 184.612 ---------------------------------------------------------------------------------------
Under CPU
:
- Sock x Core
displays the number of processor sockets, number of cores per socket, and whether hyper-threading is used.
- Clock (MHz)
is the processor speed.
- Load
is the work load on the machine.
Under System Mem (GB)
:
- Total
is the total system memory on the machine.
- Available
is the available system memory on the machine.
You can print out information about the GPUs available on the machine by clicking GPU Info in the Parallel ribbon tab (System group box). If your machine has one or more suitable GPUs present, you can use these to accelerate parallel processing computations as described in Using General Purpose Graphics Processing Units (GPGPUs) With the Algebraic Multigrid (AMG) Solver.
and selectingCUDA visible GPUs on host01 CUDA runtime version 5000 Driver version 6000 Number of GPUs 1 0. Quadro K2100M 3 SMs 0.6665 GHz 2.14748 GBytes
You can print out tables summarizing memory usage by node and by host by clicking Memory Usage in the Parallel ribbon tab (System group box).
and selecting--------------------------------------------- | Virtual Mem Usage (GB)| ID | Current Peak | Page Faults --------------------------------------------- host | 0.0680117 0.0695 | 2.667e+04 n0 | 0.0855195 0.114609 | 7.229e+04 n1 | 0.0869922 0.115367 | 6.988e+04 n2 | 0.0881836 0.117613 | 7.318e+04 n3 | 0.0840781 0.11384 | 6.999e+04 --------------------------------------------- Total | 0.412785 0.53093 | 3.12e+05 --------------------------------------------- ----------------------------------------------------------------- | Virtual Mem Usage (GB) | System Mem (GB) Hostname | Current Peak | ----------------------------------------------------------------- host01 | 0.412785 0.53093 | 32.673 ----------------------------------------------------------------- Total | 0.412785 0.53093 | -----------------------------------------------------------------
Under Virtual Mem Usage (GB)
:
- Current
is the virtual memory usage at the time the report is generated.
- Peak
is the peak virtual memory usage.
(Linux only) Under Resident Mem Usage (GB)
:
- Current
is the resident memory usage at the time the report is generated.
- Peak
is the peak resident memory usage.
- Page Faults
is the number of page faults that have occurred.
You can use these commands to plan Ansys Fluent jobs and machines accordingly. They may also be useful to diagnose performance problems.
You can print out detailed information about the CPU timings for the current session as well as solver timings by clicking Time Usage in the Parallel ribbon tab (System group box). The solver timings presented are described in Checking Parallel Performance.
and selecting--------------------------------------------- | CPU Time Usage (Seconds) ID | User Kernel Elapsed --------------------------------------------- host | 1.34161 0.452403 2328.63 n0 | 3.47882 0.702005 2325.88 n1 | 19.9525 1.15441 2325.87 n2 | 19.8433 0.873606 2325.87 n3 | 19.8589 0.764405 2325.86 --------------------------------------------- Total | 64.4752 3.94683 - --------------------------------------------- Model Timers (Host) Other Models Time: 2.858 sec Total Time: 2.858 sec Model Timers Other Models Time: 2.871 sec Total Time: 2.871 sec Performance Timer for 472 iterations on 4 compute nodes Average wall-clock time per iteration: 0.823 sec Global reductions per iteration: 93 ops Global reductions time per iteration: 0.000 sec (0.0%) Message count per iteration: 3842 messages Data transfer per iteration: 17.226 MB LE solves per iteration: 7 solves LE wall-clock time per iteration: 0.242 sec (29.4%) LE global solves per iteration: 3 solves LE global wall-clock time per iteration: 0.001 sec (0.1%) LE global matrix maximum size: 24 AMG cycles per iteration: 8.017 cycles Relaxation sweeps per iteration: 374 sweeps Relaxation exchanges per iteration: 0 exchanges LE early protections (stall) per iteration: 0.000 times LE early protections (divergence) per iteration: 0.000 times Total SVARS touched: 385 Time-step updates per iteration: 0.11 updates Time-step wall-clock time per iteration: 0.003 sec (0.4%) Total wall-clock time: 388.669 sec Simulation wall-clock time for 472 iterations 426 sec