Hi,
    The script works OK for me :)   I really appreciate it.

BTW, there are some error info in the log, does it make sense?
  [exiv2] Failed to convert Xmp.exif.GPSLatitude to Exif.GPSInfo.GPSLatitude
  [exiv2] Exif tag Exif.PentaxDng.0x0405 not encoded

Except getting dev_process_export time, is there another way to check the 
result is OK[1]?

[1]the ppm file is right, and opencl really work 

Thanks,
Meng

-----Original Message-----
From: Christian Kanzian [mailto:christian.kanz...@gmx.at] 
Sent: Wednesday, November 12, 2014 5:44 AM
To: darktable-devel@lists.sourceforge.net
Subject: Re: [darktable-devel] test opencl performance by darktable

Am Dienstag, 11. November 2014 schrieb Meng, Mengmeng:
> Hi,
> In Linux, I want to test opencl performance by darktable 
> automatically. But there are some things confused me, could somebody help me?
> 
> In console, I run "darktable  -d opencl -d perf test.jpg[1]". But 
> after done the testing, the progress can't be closed unless using "Ctrl + c".
> Then I can get 4 kinds time: export, dev_process_thumbnail, 
> dev_process_image and dev_process_preview. Is OpenCL only used in export ?
> I also see some  info as "opencl_profiling" in dev_process_thumbnail.
> 
> 
> [1]test.jpg is huge file ~8 MB.
> 
> 
> 
> Thanks,
> Meng

Hi,

I guess you are looking for darktable-cli? Some time ago Tobias suggested 
something like "darktable-cli RAW  RAW.xmp  output.ppm --width 0 --height 0 
--hq 1-d opencl - d  perf" 
for testing/benchmarking.

I tried to create a small bash script arround that for benchmarking. However 
it's far from being perfect. 
The script processes a single RAW file creating a PPM using darktable-cli and 
calculates the average processing time of three runs. Processing time is taken 
form the darktable log file. Different sidecar files could be defined 
reflecting different image developments. Eg. I'm testing profiled denoise 
on/off. Optional OpenCL parametres could be set. As default OpenCL true/false 
will be passed to darktable-cli. So, theoretically the difference can be seen, 
but there is no error handling and testing if opencl really works. 

The script (dt_benchmark.sh), a raw file and 2 sidecarfiles can be found here:
https://homepage.boku.ac.at/kanzian/darktable
Just copy everthing in one dir, run dt_benchmark.sh from the consol and watch.

I hope this helps.

All the best,
Christian

------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
darktable-devel mailing list
darktable-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-devel

------------------------------------------------------------------------------
Comprehensive Server Monitoring with Site24x7.
Monitor 10 servers for $9/Month.
Get alerted through email, SMS, voice calls or mobile push notifications.
Take corrective actions from your mobile device.
http://pubads.g.doubleclick.net/gampad/clk?id=154624111&iu=/4140/ostg.clktrk
_______________________________________________
darktable-devel mailing list
darktable-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/darktable-devel

Reply via email to