| ... | ... | @@ -1029,6 +1029,80 @@ You can see from the output that we have **3 GPU's: Cuda devices: 0,1,2** |
|
|
|
|
|
|
|
#### Compiling and running GPU programs
|
|
|
|
|
|
|
|
```
|
|
|
|
module purge
|
|
|
|
module add library/cuda/10.1/gcc.8.3.1
|
|
|
|
|
|
|
|
cd
|
|
|
|
git clone https://github.com/NVIDIA/cuda-samples.git
|
|
|
|
cd cuda-samples/Samples/UnifiedMemoryPerf/
|
|
|
|
make
|
|
|
|
|
|
|
|
cat gpu-test.slurm
|
|
|
|
#!/bin/bash
|
|
|
|
#
|
|
|
|
#SBATCH --partition=gpu
|
|
|
|
#SBATCH --gres=gpu:1
|
|
|
|
#SBATCH --ntasks=1
|
|
|
|
#SBATCH --cpus-per-task=1
|
|
|
|
#SBATCH --time=3:00
|
|
|
|
|
|
|
|
module purge
|
|
|
|
module add library/cuda/10.2/gcc.8.3.1
|
|
|
|
|
|
|
|
hostname
|
|
|
|
echo "Cuda devices: $CUDA_VISIBLE_DEVICES"
|
|
|
|
$HOME/cuda-samples/Samples/UnifiedMemoryPerf/UnifiedMemoryPerf
|
|
|
|
|
|
|
|
|
|
|
|
[user@res-hpc-gpu01 GPU]$ nvidia-smi
|
|
|
|
Tue Apr 14 16:06:06 2020
|
|
|
|
+-----------------------------------------------------------------------------+
|
|
|
|
| NVIDIA-SMI 440.33.01 Driver Version: 440.33.01 CUDA Version: 10.2 |
|
|
|
|
|-------------------------------+----------------------+----------------------+
|
|
|
|
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
|
|
|
|
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|
|
|
|
|===============================+======================+======================|
|
|
|
|
| 0 TITAN Xp Off | 00000000:3B:00.0 Off | N/A |
|
|
|
|
| 17% 31C P0 61W / 250W | 0MiB / 12196MiB | 0% Default |
|
|
|
|
+-------------------------------+----------------------+----------------------+
|
|
|
|
| 1 TITAN Xp Off | 00000000:AF:00.0 Off | N/A |
|
|
|
|
| 23% 34C P2 69W / 250W | 259MiB / 12196MiB | 2% Default |
|
|
|
|
+-------------------------------+----------------------+----------------------+
|
|
|
|
| 2 TITAN Xp Off | 00000000:D8:00.0 Off | N/A |
|
|
|
|
| 18% 31C P0 61W / 250W | 0MiB / 12196MiB | 0% Default |
|
|
|
|
+-------------------------------+----------------------+----------------------+
|
|
|
|
|
|
|
|
+-----------------------------------------------------------------------------+
|
|
|
|
| Processes: GPU Memory |
|
|
|
|
| GPU PID Type Process name Usage |
|
|
|
|
|=============================================================================|
|
|
|
|
| 1 29726 C ...les/UnifiedMemoryPerf/UnifiedMemoryPerf 145MiB |
|
|
|
|
+-----------------------------------------------------------------------------+
|
|
|
|
|
|
|
|
|
|
|
|
cat slurm-206625.out
|
|
|
|
res-hpc-gpu01.researchlumc.nl
|
|
|
|
Cuda devices: 0
|
|
|
|
GPU Device 0: "Pascal" with compute capability 6.1
|
|
|
|
|
|
|
|
Running ........................................................
|
|
|
|
|
|
|
|
Overall Time For matrixMultiplyPerf
|
|
|
|
|
|
|
|
Printing Average of 100 measurements in (ms)
|
|
|
|
Size_KB UMhint UMhntAs UMeasy 0Copy MemCopy CpAsync CpHpglk CpPglAs
|
|
|
|
4 10.879 23.178 0.222 0.014 0.031 0.026 0.035 0.026
|
|
|
|
16 10.657 25.849 0.580 0.030 0.051 0.046 0.052 0.039
|
|
|
|
64 21.117 37.351 0.852 0.103 0.124 0.116 0.095 0.081
|
|
|
|
256 21.184 38.074 1.387 0.587 0.450 0.415 0.313 0.302
|
|
|
|
1024 24.174 33.124 3.032 3.650 1.741 1.649 1.211 1.199
|
|
|
|
4096 21.668 35.167 11.067 25.803 7.119 7.104 5.329 5.333
|
|
|
|
16384 51.674 62.263 49.300 191.051 34.179 34.632 28.582 28.054
|
|
|
|
|
|
|
|
NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.
|
|
|
|
```
|
|
|
|
|
|
|
|
### Slurm Environment Variables
|
|
|
|
|
|
|
|
Available environment variables include:
|
| ... | ... | |
