forked from sherholz/mitsuba
-
Notifications
You must be signed in to change notification settings - Fork 5
/
reproduce_results.sh
executable file
·45 lines (40 loc) · 2.35 KB
/
reproduce_results.sh
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
#!/bin/bash
if [ ! -f "dist/mitsuba" ]
then
echo You first need to compile Mitsuba.
echo On Linux, you can just run compile.sh
exit 1
fi
# fetch scenes
if [ ! -d "scenes" ]
then
# these scenes are only provided for reproduction purposes - not all scenes' licences allow for redistribution
[ -f "scenes-reproduction.zip" ] || wget http://www.graphics-data.uni-tuebingen.de/ruppert/robust_fitting_parallax_guiding_2020/scenes-reproduction.zip
if [ $? -eq 0 -a "$(md5sum scenes-reproduction.zip | cut -f1 -d' ')" = "f736a16efcae3df39c0af3cf3e67914d" ]
then
unzip scenes-reproduction.zip -d scenes
else
echo Failed to fetch the scene files - please report back to the authors.
exit 1
fi
fi
echo
echo Experiments will be run shortly.
echo Note that we cleaned up the code and made some small optimizations.
echo The code should now run slightly faster than the reported numbers when run on the same hardware.
echo Our reference hardware is using dual Intel\(R\) Xeon\(R\) Gold 5115 CPUs @ 2.40GHz \(40 threads\).
echo Depending on your hardware, execution times will differ - we fixed the sample counts used in the experiments based on the reported numbers.
echo
echo Note that the glossy_cbox scene changed due to a fixed bug in Mitsuba. The results will differ slightly as a result of that.
echo Finally, keep in mind that the random nature of path tracing will introduce some variance to the results, especially in terms of the reported relMSE.
echo
echo Running experiments for table 3 / figure 9 \(Li cos only\) ... \(about 2 hours total / 10 minutes per scene on similar hardware\)
echo Results will be created in the scenes folder as scene_table3.exr and scene_table3.log
echo The resulting table will be stored in table3.csv \(speedup = relMSE baseline / relMSE of the result\)
sh reproduction/run_table_3.sh > table3.csv
echo
echo Running experiments for figure 8 \(about 2.5 hours total / 10 minutes per scene and configuration on similar hardware\)
echo Results will be created in the scene folders as scene_pt.exr, scene_pg_li.exr, scene_pg_li_parallax.exr,
echo scene_pg_li_parallax_ssmem_cos.exr, scene_pg_li_parallax_ssmem_cos_bsdf.exr, scene_pg_li_parallax_ssmem_cos_bsdf_nee.exr.
echo The resulting table will be stored in figure8.csv \(speedup = relMSE for Li / relMSE of the result\)
sh reproduction/run_figure_8.sh > figure8.csv