This shows you the differences between two versions of the page.
Both sides previous revision Previous revision Next revision | Previous revision Next revision Both sides next revision | ||
cluster:175 [2018/09/25 12:35] hmeij07 |
cluster:175 [2018/10/09 13:38] hmeij07 |
||
---|---|---|---|
Line 2: | Line 2: | ||
\\ | \\ | ||
**[[cluster: | **[[cluster: | ||
+ | |||
+ | As of | ||
+ | --- // | ||
+ | The P100 with 12 GB is end of life, replaced by the P100 16 GB and \\ | ||
+ | The GTX 1080Ti will be replaced by the GTX2060Ti (no specs yet and not certified for Amber18, yet) | ||
+ | |||
==== P100 vs GTX & K20 ==== | ==== P100 vs GTX & K20 ==== | ||
Line 70: | Line 76: | ||
</ | </ | ||
- | === WZ === | + | ==== Lammps (PMMA) ==== |
- | a) 1GPU with X CPUs | + | Using a material called PMMA (https:// |
- | #cpus | + | ^ gpu ^ |
- | 4 | + | | 1 P100 | |
- | 2 61.6 | + | | 1 GTX | |
- | 1 34.3 | + | | |
- | b) 4 GPUs with 4 CPUs | + | That means the P100 works as well as the GTX. The K20 works at 50% the performance level of the others which is impressive for this old gpu. |
- | 92 ns/day | + | |
- | ---------- | ||
- | compare these results to | ||
- | 1 K20 GPU + 4 CPUs 37 ns/day | ||
- | 1 K20 GPU + 6 CPUs 47 ns/day | ||
- | 1 GTX GPU + 4 CPUs 73 ns/day | ||
- | 1 GTX GPU + 6 CPUs 90 ns/day | ||
- | ---- | ||
- | Test dir location | ||
- | / | ||
- | |||
- | |||
- | -------------- | ||
- | I am running a material called PMMA (https:// | ||
- | |||
- | The reason for the different benchmark than yours is that the PMMA simulations require the calculation of molecular bonds, which is not implemented in GPU. | ||
==== Gromacs ==== | ==== Gromacs ==== | ||
Line 134: | Line 124: | ||
</ | </ | ||
+ | |||
+ | ==== Gromacs 2018.3 ==== | ||
+ | |||
+ | The multidir not running in Gromacs 2018 is a bug in the code clashing with the call MPI_Barrier (communication timing error). | ||
+ | |||
+ | < | ||
+ | |||
+ | # multidir -gpu_id 0123 with four simultaneous gromacs processes | ||
+ | |||
+ | -np 8 -ntomp | ||
+ | 01/ | ||
+ | 02/ | ||
+ | 03/ | ||
+ | 04/ | ||
+ | |||
+ | -np 16 -ntomp | ||
+ | 01/ | ||
+ | 02/ | ||
+ | 03/ | ||
+ | 04/ | ||
+ | |||
+ | multidir -gpu_id 00112233 with eight simultaneous gromacs processes | ||
+ | sharing the gpus, 2 processes per gpu | ||
+ | |||
+ | -np 8 -ntomp | ||
+ | Error in user input: | ||
+ | The string of available GPU device IDs ' | ||
+ | device IDs | ||
+ | |||
+ | </ | ||
+ | |||
+ | That last error when loading multiple processes per gpu is *not* according to their documentation. So the multidir performance is similar to previous single dir performance but still lags GTX performance by quite a bit. Albeit, there is room in utilization rate of the gpus. | ||
+ | |||
==== What to Buy ==== | ==== What to Buy ==== | ||
* Amber folks: does not matter | * Amber folks: does not matter | ||
- | * Lammps folks: P100 nodes please | + | * Lammps folks: P100 nodes please, although in hybrid runs does not matter |
* Gromacs folks: GTX nodes please | * Gromacs folks: GTX nodes please | ||