Skip to content

Commit

Permalink
Merge pull request #22 from rcui7/master
Browse files Browse the repository at this point in the history 
groupJ submission
  • Loading branch information
@stubbiali
stubbiali authored Sep 15, 2021
2 parents 7b3d579 + 67b45dd commit 020d2b0
Show file tree
Hide file tree
Showing 12 changed files with 1,114 additions and 0 deletions.
Empty file removed projects/2021/groupJ_shockwaves_gt4py/.invisible
View file
Empty file.
38 changes: 38 additions & 0 deletions projects/2021/groupJ_shockwaves_gt4py/README.md
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,38 @@
# HPC4WC: Simulating Shock Waves Using a Domain-Specific Language (DSL) in Python

Implement the 5th-order weighted essentially non-oscillatory (WENO) scheme (WENO5) for spatial approximation and the 3rd-order 3-stage Strong Stability Preserving Runge Kutta (SSP-RK(3,3)) as time integration using GT4Py to solve Burger's equation.

### Root directory
* ```./src/BurgersWENOdriver.py``` Driver for main program
* ```./src/BurgersWENOcalc.py``` Contains functions required for the main program
* ```./src/BurgersWENOperf.py``` Plot performance from log file
* ```./src/BurgersWENO.py``` Main program
* ```./run``` Run performance tests with different settings
* ```./output/log``` Output to validate the performance
* ```./output/WENO2D_*``` Figure outputs for different GT4Py backends
* ```./requirements.txt``` Python requirements
* ```./report.pdf```
* ```./README.md```

### Setup environment
```source HPC4WC_venv/bin/activate ```

### Plot the figure using command line
```python3 BurgersWENOdriver.py --nx=50 --ny=30 --nz=1 --plot=True --backend=gtcuda --time=True --tmp=True```

### Running options
* ```--m``` m(2m-1)th order of ENO(WENO) approximation
* ```--nx``` Number of gridpoints in x-direction
* ```--ny``` Number of gridpoints in y-direction
* ```--nz``` Number of gridpoints in z-direction
* ```--gt4py``` Use of GT4Py?
* ```--tmp``` Use of temporary variables?
* ```--backend``` NumPy or GT4Py backends?
* ```--plot``` Make a plot of the result?
* ```--time``` Write the elapsed time?

### Write the performance
```bash run```


© Ruoyi Cui, Shuchang Liu, Shihao Zeng
Binary file added projects/2021/groupJ_shockwaves_gt4py/output/WENO2D_gtcuda.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added projects/2021/groupJ_shockwaves_gt4py/output/WENO2D_performance.png
View file
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
245 changes: 245 additions & 0 deletions projects/2021/groupJ_shockwaves_gt4py/output/log
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,245 @@
GT4Py tmp Backend nx ny nz Elapsed
True True numpy 30 30 1 4.455572
True True numpy 30 30 1 4.341864
True True numpy 30 30 1 4.410543
True True numpy 30 30 1 4.341405
True True numpy 30 30 1 4.446870
True True numpy 50 30 1 6.285703
True True numpy 50 30 1 6.351858
True True numpy 50 30 1 6.260638
True True numpy 50 30 1 6.376969
True True numpy 50 30 1 6.293254
True True numpy 100 30 1 10.638374
True True numpy 100 30 1 10.737999
True True numpy 100 30 1 10.737260
True True numpy 100 30 1 10.747593
True True numpy 100 30 1 10.696152
True True numpy 200 30 1 19.878750
True True numpy 200 30 1 19.941572
True True numpy 200 30 1 19.572268
True True numpy 200 30 1 19.824927
True True numpy 200 30 1 19.828471
True True numpy 500 30 1 48.771551
True True numpy 500 30 1 48.471563
True True numpy 500 30 1 48.303117
True True numpy 500 30 1 49.211676
True True numpy 500 30 1 50.279034
True True numpy 1000 30 1 100.797641
True True numpy 1000 30 1 98.628214
True True numpy 1000 30 1 100.702740
True True numpy 1000 30 1 100.544396
True True numpy 1000 30 1 99.998802
True True gtmc 30 30 1 4.522080
True True gtmc 30 30 1 4.500722
True True gtmc 30 30 1 4.488715
True True gtmc 30 30 1 4.495965
True True gtmc 30 30 1 4.507765
True True gtmc 50 30 1 6.687369
True True gtmc 50 30 1 6.485512
True True gtmc 50 30 1 6.468528
True True gtmc 50 30 1 6.447612
True True gtmc 50 30 1 6.409881
True True gtmc 100 30 1 11.208968
True True gtmc 100 30 1 11.059226
True True gtmc 100 30 1 11.169734
True True gtmc 100 30 1 11.128614
True True gtmc 100 30 1 11.165113
True True gtmc 200 30 1 20.266581
True True gtmc 200 30 1 20.173133
True True gtmc 200 30 1 20.374801
True True gtmc 200 30 1 20.275722
True True gtmc 200 30 1 20.337360
True True gtmc 500 30 1 48.882098
True True gtmc 500 30 1 48.671979
True True gtmc 500 30 1 48.671847
True True gtmc 500 30 1 49.128291
True True gtmc 500 30 1 49.421175
True True gtmc 1000 30 1 96.997524
True True gtmc 1000 30 1 96.916647
True True gtmc 1000 30 1 97.571124
True True gtmc 1000 30 1 96.320021
True True gtmc 1000 30 1 97.112095
True True gtx86 30 30 1 4.504176
True True gtx86 30 30 1 4.479908
True True gtx86 30 30 1 4.548739
True True gtx86 30 30 1 4.476457
True True gtx86 30 30 1 4.494388
True True gtx86 50 30 1 6.445041
True True gtx86 50 30 1 6.422671
True True gtx86 50 30 1 6.433626
True True gtx86 50 30 1 6.504877
True True gtx86 50 30 1 6.442340
True True gtx86 100 30 1 11.173429
True True gtx86 100 30 1 11.058379
True True gtx86 100 30 1 11.195030
True True gtx86 100 30 1 11.055428
True True gtx86 100 30 1 11.068404
True True gtx86 200 30 1 20.366818
True True gtx86 200 30 1 20.314615
True True gtx86 200 30 1 20.220284
True True gtx86 200 30 1 20.327784
True True gtx86 200 30 1 20.347377
True True gtx86 500 30 1 48.644161
True True gtx86 500 30 1 48.617223
True True gtx86 500 30 1 48.711633
True True gtx86 500 30 1 48.506912
True True gtx86 500 30 1 48.812679
True True gtx86 1000 30 1 96.542383
True True gtx86 1000 30 1 96.813625
True True gtx86 1000 30 1 96.856360
True True gtx86 1000 30 1 96.994937
True True gtx86 1000 30 1 97.225100
True True gtc:dace 30 30 1 4.618316
True True gtc:dace 30 30 1 4.630934
True True gtc:dace 30 30 1 4.613844
True True gtc:dace 30 30 1 4.618588
True True gtc:dace 30 30 1 4.614735
True True gtc:dace 50 30 1 6.639204
True True gtc:dace 50 30 1 6.630151
True True gtc:dace 50 30 1 6.684541
True True gtc:dace 50 30 1 6.613182
True True gtc:dace 50 30 1 6.686515
True True gtc:dace 100 30 1 11.578210
True True gtc:dace 100 30 1 11.545269
True True gtc:dace 100 30 1 11.539422
True True gtc:dace 100 30 1 11.553653
True True gtc:dace 100 30 1 11.625582
True True gtc:dace 200 30 1 21.279351
True True gtc:dace 200 30 1 21.252109
True True gtc:dace 200 30 1 21.167343
True True gtc:dace 200 30 1 21.051675
True True gtc:dace 200 30 1 21.322242
True True gtc:dace 500 30 1 51.563408
True True gtc:dace 500 30 1 51.677147
True True gtc:dace 500 30 1 51.606193
True True gtc:dace 500 30 1 52.054671
True True gtc:dace 500 30 1 51.622771
True True gtc:dace 1000 30 1 105.696969
True True gtc:dace 1000 30 1 105.479794
True True gtc:dace 1000 30 1 105.057695
True True gtc:dace 1000 30 1 105.677828
True True gtc:dace 1000 30 1 105.766253
True True gtcuda 30 30 1 4.437476
True True gtcuda 30 30 1 4.855096
True True gtcuda 30 30 1 4.521071
True True gtcuda 30 30 1 4.416198
True True gtcuda 30 30 1 5.220069
True True gtcuda 50 30 1 6.086745
True True gtcuda 50 30 1 6.643351
True True gtcuda 50 30 1 6.095497
True True gtcuda 50 30 1 6.397161
True True gtcuda 50 30 1 6.013193
True True gtcuda 100 30 1 10.133513
True True gtcuda 100 30 1 10.330877
True True gtcuda 100 30 1 10.070746
True True gtcuda 100 30 1 10.251322
True True gtcuda 100 30 1 10.204558
True True gtcuda 200 30 1 18.329628
True True gtcuda 200 30 1 18.232420
True True gtcuda 200 30 1 18.463790
True True gtcuda 200 30 1 17.975838
True True gtcuda 200 30 1 18.797593
True True gtcuda 500 30 1 43.170795
True True gtcuda 500 30 1 42.590678
True True gtcuda 500 30 1 43.538645
True True gtcuda 500 30 1 42.531381
True True gtcuda 500 30 1 42.348237
True True gtcuda 1000 30 1 85.135280
True True gtcuda 1000 30 1 83.293649
True True gtcuda 1000 30 1 84.018219
True True gtcuda 1000 30 1 85.144900
True True gtcuda 1000 30 1 83.924436
False True numpy 30 30 1 2.159891
False True numpy 30 30 1 2.177704
False True numpy 30 30 1 2.173837
False True numpy 30 30 1 2.191250
False True numpy 30 30 1 2.156557
False True numpy 50 30 1 5.837689
False True numpy 50 30 1 5.891549
False True numpy 50 30 1 5.988059
False True numpy 50 30 1 5.865530
False True numpy 50 30 1 5.856378
False True numpy 100 30 1 22.659357
False True numpy 100 30 1 22.836402
False True numpy 100 30 1 22.865970
False True numpy 100 30 1 22.791305
False True numpy 100 30 1 22.722815
False True numpy 200 30 1 87.828047
False True numpy 200 30 1 88.072068
False True numpy 200 30 1 87.871486
False True numpy 200 30 1 87.981900
False True numpy 200 30 1 87.879858
False True numpy 500 30 1 543.737415
False True numpy 500 30 1 545.126304
False True numpy 500 30 1 542.427482
False True numpy 500 30 1 543.686740
False True numpy 500 30 1 544.940447
False True numpy 1000 30 1 2170.769349
False True numpy 1000 30 1 2165.737256
False True numpy 1000 30 1 2157.595848
False True numpy 1000 30 1 2184.731343
False True numpy 1000 30 1 2160.047810
True False numpy 30 30 1 4.467573
True False numpy 30 30 1 4.477871
True False numpy 30 30 1 4.508997
True False numpy 30 30 1 4.422922
True False numpy 30 30 1 4.407830
True False numpy 50 30 1 6.326806
True False numpy 50 30 1 6.336152
True False numpy 50 30 1 6.314680
True False numpy 50 30 1 6.265125
True False numpy 50 30 1 6.234338
True False numpy 100 30 1 10.550262
True False numpy 100 30 1 10.674019
True False numpy 100 30 1 10.403131
True False numpy 100 30 1 10.573890
True False numpy 100 30 1 10.566756
True False numpy 200 30 1 19.617156
True False numpy 200 30 1 19.570396
True False numpy 200 30 1 19.515646
True False numpy 200 30 1 19.388751
True False numpy 200 30 1 19.389820
True False numpy 500 30 1 48.024278
True False numpy 500 30 1 47.187096
True False numpy 500 30 1 48.272399
True False numpy 500 30 1 47.786161
True False numpy 500 30 1 48.090288
True False numpy 1000 30 1 101.959332
True False numpy 1000 30 1 97.855774
True False numpy 1000 30 1 97.549154
True False numpy 1000 30 1 95.719771
True False numpy 1000 30 1 96.736551
True True numpy 1500 30 1 157.745750
True True numpy 1500 30 1 157.644647
True True numpy 1500 30 1 155.070689
True True numpy 1500 30 1 157.502035
True True numpy 1500 30 1 156.251705
True True gtmc 1500 30 1 145.953752
True True gtmc 1500 30 1 146.282009
True True gtmc 1500 30 1 146.179155
True True gtmc 1500 30 1 146.052309
True True gtmc 1500 30 1 146.330244
True True gtx86 1500 30 1 146.003944
True True gtx86 1500 30 1 146.219023
True True gtx86 1500 30 1 146.361376
True True gtx86 1500 30 1 146.792965
True True gtx86 1500 30 1 145.871367
True True gtc:dace 1500 30 1 163.516285
True True gtc:dace 1500 30 1 163.066534
True True gtc:dace 1500 30 1 164.721022
True True gtc:dace 1500 30 1 162.723360
True True gtc:dace 1500 30 1 163.409103
True True gtcuda 1500 30 1 127.510200
True True gtcuda 1500 30 1 127.487468
True True gtcuda 1500 30 1 128.492820
True True gtcuda 1500 30 1 127.119187
True True gtcuda 1500 30 1 126.772917
False True numpy 1500 30 1 4889.370258
False True numpy 1500 30 1 4880.716677
False True numpy 1500 30 1 4854.967964
False True numpy 1500 30 1 4867.788195
True False numpy 1500 30 1 147.806941
True False numpy 1500 30 1 149.534964
True False numpy 1500 30 1 146.775961
True False numpy 1500 30 1 144.989492
True False numpy 1500 30 1 145.897569
Binary file added projects/2021/groupJ_shockwaves_gt4py/report.pdf
View file
Binary file not shown.
10 changes: 10 additions & 0 deletions projects/2021/groupJ_shockwaves_gt4py/requirements.txt
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
Pygments==2.6.1
Pillow==8.3.2
black==21.8b0
autopep8==1.5.7
click==8.0.1
matplotlib==3.4.2
numpy==1.20.3
scipy=1.7.1
pandas=1.3.2
gt4py @ git+https://github.com/gridtools/gt4py.git
27 changes: 27 additions & 0 deletions projects/2021/groupJ_shockwaves_gt4py/run
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
#! /bin/bash

module load daint-gpu

# Use GT4Py
for backend in numpy gtmc gtx86 gtc:dace gtcuda; do
for nx in 30 50 100 200 500 1000 1500; do
# Multiple trials
for i in 1 2 3 4 5; do
python3 BurgersWENOdriver.py --nx=${nx} --ny=30 --nz=1 --gt4py=True --tmp=True --backend=${backend} --time=True
done
done
done

# Use NumPy
for nx in 30 50 100 200 500 1000 1500; do
for i in 1 2 3 4 5; do
python3 BurgersWENOdriver.py --nx=${nx} --ny=30 --nz=1 --gt4py=False --time=True
done
done

# Use GT4Py(numpy) without temporary variables
for nx in 30 50 100 200 500 1000 1500; do
for i in 1 2 3 4 5; do
python3 BurgersWENOdriver.py --nx=${nx} --ny=30 --nz=1 --gt4py=True --backend=numpy --time=True --tmp=False
done
done
Loading

0 comments on commit 020d2b0

Please sign in to comment.