Am 18.11.2014 um 20:41 schrieb Christian Kanzian:
> Fine. I guess that's all? 
> 
> BTW, my times for a GeForce GT 640 and an i7-2600 CPU @ 3.40GHz are:
> .. average processing time in sec
> sidecarfile: profiled denoise=off
> opencl=false:5.38
> opencl=true:3.79
> 
> sidecarfile: profiled denoise=on
> opencl=false:19.75
> opencl=true:27.40
> 
> So OpenCL will be faster, but not if I use profiled denoise (nlm) on my 
> graphics card. Next time I will take another one ;-).

Hi,

I just want to add my opencl performance data because it is significant. 
Especially with denoised profile. Five times faster with opencl.
Running on a i7-2600K CPU @ 3.40GHz with GeForce GTX 750 TI:

Without opencl I get:
=====================
[dev_pixelpipe] took 0.000 secs (-0.000 CPU) initing base buffer [export]
[dev_pixelpipe] took 0.029 secs (0.112 CPU) processed `let's go float!' on CPU, 
blended on CPU [export]
[dev_pixelpipe] took 0.026 secs (0.124 CPU) processed `white balance' on CPU, 
blended on CPU [export]
[dev_pixelpipe] took 0.019 secs (0.112 CPU) processed `highlight 
reconstruction' on CPU, blended on CPU [export]
[dev_pixelpipe] took 0.215 secs (1.144 CPU) processed `demosaic' on CPU, 
blended on CPU [export]
[dev_pixelpipe] took 32.670 secs (222.948 CPU) processed `denoise (profiled)' 
on CPU, blended on CPU [export]
[dev_pixelpipe] took 0.113 secs (0.680 CPU) processed `base curve' on CPU, 
blended on CPU [export]
[dev_pixelpipe] took 0.139 secs (0.952 CPU) processed `input color profile' on 
CPU, blended on CPU [export]
[dev_pixelpipe] took 0.256 secs (1.624 CPU) processed `sharpen' on CPU, blended 
on CPU [export]
[dev_pixelpipe] took 0.165 secs (1.076 CPU) processed `output color profile' on 
CPU, blended on CPU [export]
[dev_pixelpipe] took 0.056 secs (0.344 CPU) processed `gamma' on CPU, blended 
on CPU [export]
[dev_process_export] pixel pipeline processing took 33.689 secs (229.116 CPU)

But with opencl I get:
======================
[dev_pixelpipe] took 0.000 secs (0.000 CPU) initing base buffer [export]
[dev_pixelpipe] took 0.024 secs (0.020 CPU) processed `let's go float!' on GPU, 
blended on GPU [export]
[dev_pixelpipe] took 0.008 secs (0.008 CPU) processed `white balance' on GPU, 
blended on GPU [export]
[dev_pixelpipe] took 0.008 secs (0.008 CPU) processed `highlight 
reconstruction' on GPU, blended on GPU [export]
[dev_pixelpipe] took 0.045 secs (0.036 CPU) processed `demosaic' on GPU, 
blended on GPU [export]
[dev_pixelpipe] took 5.990 secs (4.888 CPU) processed `denoise (profiled)' on 
GPU, blended on GPU [export]
[dev_pixelpipe] took 0.020 secs (0.020 CPU) processed `base curve' on GPU, 
blended on GPU [export]
[dev_pixelpipe] took 0.025 secs (0.024 CPU) processed `input color profile' on 
GPU, blended on GPU [export]
[dev_pixelpipe] took 0.053 secs (0.044 CPU) processed `sharpen' on GPU, blended 
on GPU [export]
[dev_pixelpipe] took 0.027 secs (0.016 CPU) processed `output color profile' on 
GPU, blended on GPU [export]
[dev_pixelpipe] took 0.189 secs (0.400 CPU) processed `gamma' on CPU, blended 
on CPU [export]
[opencl_profiling] spent  0.0146 seconds in [Write Image (from host to device)]
[opencl_profiling] spent  0.0071 seconds in letsgofloat_1ui
[opencl_profiling] spent  0.0064 seconds in whitebalance_1f
[opencl_profiling] spent  0.0063 seconds in highlights_1f
[opencl_profiling] spent  0.0158 seconds in ppg_demosaic_green
[opencl_profiling] spent  0.0250 seconds in ppg_demosaic_redblue
[opencl_profiling] spent  0.0028 seconds in border_interpolate
[opencl_profiling] spent  0.0119 seconds in denoiseprofile_precondition
[opencl_profiling] spent  0.0053 seconds in denoiseprofile_init
[opencl_profiling] spent  0.8217 seconds in denoiseprofile_dist
[opencl_profiling] spent  0.8358 seconds in denoiseprofile_horiz
[opencl_profiling] spent  1.6055 seconds in denoiseprofile_vert
[opencl_profiling] spent  2.5643 seconds in denoiseprofile_accu
[opencl_profiling] spent  0.0138 seconds in denoiseprofile_finish
[opencl_profiling] spent  0.0179 seconds in basecurve
[opencl_profiling] spent  0.0228 seconds in colorin_unbound
[opencl_profiling] spent  0.0185 seconds in sharpen_hblur
[opencl_profiling] spent  0.0164 seconds in sharpen_vblur
[opencl_profiling] spent  0.0170 seconds in sharpen_mix
[opencl_profiling] spent  0.0246 seconds in colorout
[opencl_profiling] spent  0.1420 seconds in [Read Image (from device to host)]
[opencl_profiling] spent  6.1955 seconds totally in command queue (with 0 
events missing)
[dev_process_export] pixel pipeline processing took 6.389 secs (5.464 CPU)

Matthias



------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=157005751&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