Re: [darktable-user] Chosing the optimal CPU for darktable
Am Sonntag, 27. November 2016, 16:25:59 CET schrieb Rico Heil: > Am 27.11.2016 um 16:17 schrieb Ulrich Pegelow: > > Am 27.11.2016 um 16:04 schrieb Rico Heil: > >> Actually, I can't disable it. > >> Seems like a cosmetical bug to me: The checkmark should not be set > > > > True, that's a minor UI issue. Fixing it requires a bit of thought... > > ... and I bet there are many more beneficial areas to spend your time ;-) > > >> and > >> there should be some visual representation of the fact that the control > >> is disabled (ie. it should be grayed). > > > > It should be grayed out already. > > It's not on my machine: > > > At least I am not able to see any difference. For me it looks the same. It differs from sensitive toggles in the tiny light grey line in the top and bottom of the box. Quite subtle. > Rico Tobias signature.asc Description: This is a digitally signed message part.
Re: [darktable-user] lighttable pixelization
Problem appears to emulate from the cropping is there a way to read the crop values from the sidecar? David On 11/25/2016 01:28 AM, Matthieu Moy wrote: - Original Message - I sometimes find that after processing my images get pixelised on the lighttable. If I zoom in or out the problem remains but it is not creating problems on the final output. anybody else see this? If you look closely, you'll see that it's not pixelization : you do see large squares, but there are details inside these squares (as opposed to a large pixel which obviously doesn't). This looks very much like an interference pattern due to resizing (e.g. subsampling with an "almost integer" factor which results in sampling always the R pixels of your sensor for a while, then only the B pixels, ...). It's probably not a coincidence that the pattern you see on the image is the same as the color pattern on your camera's color filter array. Unfortunately, I don't know the algorithms enough to give a real diagnosis, and even less a solution :-(. If you're looking for a combination of modules, I'd look at the crop module (the crop factor may be the one giving you the "almost integer" factor), and the demosaic module. darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Sharpen and denoise vs. RawTherapee
Le dimanche 27 novembre 2016 à 11:22 -0500, Robert Krawitz a écrit : > It's a very different story with Darktable. I can't get anywhere > near > the combination of smoothness and sharpness whatever I do. With > profiled denoise, some areas get denoised and some really don't; > adjusting the strength slider does have some effect here, but mostly > it just blurs out some areas while leaving some areas untouched for a > very unnatural effect. Have you tried using the color noise reduction on the "color" channel only? Read the section "3.4.4.3. Denoise – profiled" on the darktable manual if not already done. This is almost becoming a FAQ :) Hope this helps! -- Pascal Obry / Magny Les Hameaux (78) The best way to travel is by means of imagination http://www.obry.net gpg --keyserver keys.gnupg.net --recv-key F949BD3B darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
- Original Message - > Am 26.11.2016 um 18:11 schrieb Niccolò Belli: > > You will probably get better performance saving some bucks on the CPU > > and buying a very fast GPU for OpenCL acceleration. Something like the > > RADEON RX 480 would be an optimal solution because of FOSS drivers, > > but you will have to use the AMDGPU-PRO driver until Clover Image > > support gets mainlined (hopefully soon). > > I was planning to use Intel chipset graphics without buying an > additional GPU at all. > Can darktable use OpenCL with those? Essentially: * A good GPU makes dt faster that a CPU (several times faster) * A not-so-good GPU (typically the integrated Intel chipset, or some cheap external graphics card) would make dt *slower* than your CPU. Recent dt version try to disable OpenCL when it notices such situation. * If you have a good GPU, then a faster CPU won't make much difference as most computations can use OpenCL. > If yes: how much performance improvement would be expected if I add an > additional graphics board? I did some benchmarks on my machine with and without GPU, the discussion is still available here: https://www.mail-archive.com/darktable-users@lists.sourceforge.net/msg06504.html -- Matthieu Moy http://www-verimag.imag.fr/~moy/ darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Sorry, it was using default processing, here are the correct results: GPU: [dev_process_export] pixel pipeline processing took 64,495 secs (14,497 CPU) CPU: [dev_process_export] pixel pipeline processing took 43,410 secs (168,420 CPU) On domenica 27 novembre 2016 17:08:48 CET, Niccolò Belli wrote: GPU: [dev_process_export] pixel pipeline processing took 1,253 secs (0,240 CPU) CPU: [dev_process_export] pixel pipeline processing took 0,875 secs (3,210 CPU) model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz totalusedfree shared buff/cache available Mem: 7,7G1,6G2,3G359M 3,8G5,4G Swap: 8,0G167M7,8G 00:02.0 VGA compatible controller: Intel Corporation HD Graphics 5500 (rev 09) Linux arch-laptop 4.8.10-1-ARCH #1 SMP PREEMPT Mon Nov 21 11:55:43 CET 2016 x86_64 GNU/Linux As you can see the cpu is ~40% faster than my Broadwell GPU. darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
[darktable-user] Sharpen and denoise vs. RawTherapee
I'm trying to sharpen and denoise photographs from my Canon 7DmkII at high ISO. When I shoot basketball games, I'm forced to ISO 1 or thereabouts by our lighting (it was probably 1/3 stop too high as you'll see, but that doesn't really change matters very much). For various reasons I'd prefer to use Darktable (for the most part it's more powerful, it's much more stable, and I'm liking the interface), but the noise reduction simply isn't acceptable. Last night I was forced to use both (initial processing with RawTherapee, including exposure/lightness/saturation, and that since it interacts with NR, and then cropping and watermarking with darktable because RT crashes endlessly). Reference https://rlk.smugmug.com/Photography/DarktableRawTherapee/ (which is a crop from the original). The 7DmkII has a fairly aggressive optical low pass filter, so it's well-advertised that shots from said camera need to be sharpened out of the camera. It's quite noisy at ISO 1 (first shot in the linked gallery), not too surprisingly, but I don't have much trouble cleaning it up very nicely with RawTherapee (third shot), using sharpen with Sharpen Only Edges enabled and straightforward noise reduction (defaults for chroma, which isn't much of a problem here anyway, and a good bit of Luminance and Luminance Detail). It's a very different story with Darktable. I can't get anywhere near the combination of smoothness and sharpness whatever I do. With profiled denoise, some areas get denoised and some really don't; adjusting the strength slider does have some effect here, but mostly it just blurs out some areas while leaving some areas untouched for a very unnatural effect. Sharpening sharpens the residual noise as much as it sharpens the edges. I've attached the metadata files from DT and RT for illustration. The actual shot, fully processed and cropped (very differently) is at https://rlk.smugmug.com/Sports/Basketball/MIT-SarLaw-mbb-20161126/i-chq2xjG/A -- Robert Krawitz *** MIT Engineers A Proud Tradition http://mitathletics.com *** Member of the League for Programming Freedom -- http://ProgFree.org Project lead for Gutenprint --http://gimp-print.sourceforge.net "Linux doesn't dictate how I work, I dictate how Linux works." --Eric Crampton darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org test.jpg.pp3 Description: Binary data test.jpg.xmp Description: Binary data
Re: [darktable-user] Chosing the optimal CPU for darktable
On domenica 27 novembre 2016 16:24:19 CET, Patrick Rudin wrote: My old Phenom II 840 needs 68 seconds, a new and expensive i7 6700k still 36 seconds. The i7 with activated GTX1060 needs only 5 seconds. What? Are we talking about the same bench.SRW.xmp? It takes less than a second with my Ultra Low Voltage CPU. Niccolò Belli darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
GPU: [dev_process_export] pixel pipeline processing took 1,253 secs (0,240 CPU) CPU: [dev_process_export] pixel pipeline processing took 0,875 secs (3,210 CPU) model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz model name : Intel(R) Core(TM) i7-5600U CPU @ 2.60GHz totalusedfree shared buff/cache available Mem: 7,7G1,6G2,3G359M3,8G 5,4G Swap: 8,0G167M7,8G 00:02.0 VGA compatible controller: Intel Corporation HD Graphics 5500 (rev 09) Linux arch-laptop 4.8.10-1-ARCH #1 SMP PREEMPT Mon Nov 21 11:55:43 CET 2016 x86_64 GNU/Linux As you can see the cpu is ~40% faster than my Broadwell GPU. darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 27.11.2016 um 16:25 schrieb Rico Heil: It should be grayed out already. It's not on my machine: At least I am not able to see any difference. Actually not really visible. See the attached file how it looks here. The difference is not very big, though. That's a style issue. Not even sure if the relevant settings are within darktable. Ulrich darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
2016-11-27 16:24 GMT+01:00 Patrick Rudin : > I would be interested in the numbers of a RX460/70/80 with working > openCL... In ~10 days I should install the RX480 that I have bought some days ago. I will let you know about benchmarks darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 27.11.2016 um 16:17 schrieb Ulrich Pegelow: > Am 27.11.2016 um 16:04 schrieb Rico Heil: >> Actually, I can't disable it. >> Seems like a cosmetical bug to me: The checkmark should not be set > > True, that's a minor UI issue. Fixing it requires a bit of thought... ... and I bet there are many more beneficial areas to spend your time ;-) > >> and >> there should be some visual representation of the fact that the control >> is disabled (ie. it should be grayed). > > It should be grayed out already. It's not on my machine: At least I am not able to see any difference. Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Rico Heil worte: > If yes: how much performance improvement would be expected if I add an > additional graphics board? Very much. I did run a test yesterday in the german newsgroup dcoulh [1] with a simple testcase (including denoise and equalizer for heavy GPU-use) and the difference was huge: My old Phenom II 840 needs 68 seconds, a new and expensive i7 6700k still 36 seconds. The i7 with activated GTX1060 needs only 5 seconds. So: Forget the CPU, take a cheap one, the GPU matters. If someone wants to compare, take these two files: http://www.mirada.ch/bench.SRW http://www.mirada.ch/bench.SRW.xmp and then run: darktable-cli bench.SRW test.jpg --core -d perf -d opencl for testing the CPU only run: darktable-cli bench.SRW test.jpg --core --disable-opencl -d perf If you want to post your results here, maybe add: grep "model name" /proc/cpuinfo && free -h && lspci | grep "VGA" && uname -a I would be interested in the numbers of a RX460/70/80 with working openCL... regards Patrick [1]| https://groups.google.com/forum/#!topic/de.comp.os.unix.linux.hardware/K2mXwGyrdrM darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 27.11.2016 um 16:04 schrieb Rico Heil: Am 27.11.2016 um 15:38 schrieb Christian Kanzian: Am Sonntag, 27. November 2016, 15:32:24 schrieb Rico Heil: This discussion made me check the OpenCL paramters in my current darktable installation. "activate opencl support" is checked and I cannot uncheck it. Does this mean I am forced to use OpenCL or does it mean my GPU does not support OpenCL at all and the disabled control shows a default value that's irrelevant for me? It has been disabled. Newer dt versions do a quick test at startup for OpenCL. If the test fails, you can't enable OpenCL. Actually, I can't disable it. Seems like a cosmetical bug to me: The checkmark should not be set True, that's a minor UI issue. Fixing it requires a bit of thought... and there should be some visual representation of the fact that the control is disabled (ie. it should be grayed). It should be grayed out already. Ulrich At least I noticed the tooltip now, which states quite clearly "auf diesem System nicht verfügbar" - "not available on this system". Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am So 27 Nov 2016 15:32:24 CET schrieb Rico Heil : > This discussion made me check the OpenCL paramters in my current > darktable installation. > "activate opencl support" is checked and I cannot uncheck it. > Does this mean I am forced to use OpenCL or does it mean my GPU does > not support OpenCL at all and the disabled control shows a default > value that's irrelevant for me? > > I am using an nVidia GeForce 9800 GT with the video-nouveau driver, > but wouldn't mind switching to another driver if that would speed up > darktable. To find out what darktable is doing start it from the command line, "darktable -d opencl". And try switching to the nonfree nvidia drivers. As far as I know, OpenCL support in nouveau is not yet there. The feature matrix at https://nouveau.freedesktop.org/wiki/FeatureMatrix/ says "compute" is not available for most cards, work in progress for some. The nvidia drivers from nvidia support OpenCL well. On most systems, OpenCL should speed up darktable. If your system has a recent CPU and an old graphics card, things may be different. Cheers Michael darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 27.11.2016 um 15:38 schrieb Christian Kanzian: > Am Sonntag, 27. November 2016, 15:32:24 schrieb Rico Heil: >> This discussion made me check the OpenCL paramters in my current >> darktable installation. >> "activate opencl support" is checked and I cannot uncheck it. >> Does this mean I am forced to use OpenCL or does it mean my GPU does not >> support OpenCL at all and the disabled control shows a default value >> that's irrelevant for me? > It has been disabled. Newer dt versions do a quick test at startup for > OpenCL. > If the test fails, you can't enable OpenCL. Actually, I can't disable it. Seems like a cosmetical bug to me: The checkmark should not be set and there should be some visual representation of the fact that the control is disabled (ie. it should be grayed). At least I noticed the tooltip now, which states quite clearly "auf diesem System nicht verfügbar" - "not available on this system". Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
On Sun, Nov 27, 2016 at 5:38 PM, Christian Kanzian wrote: > Hi Hi. > Am Sonntag, 27. November 2016, 15:32:24 schrieb Rico Heil: >> This discussion made me check the OpenCL paramters in my current >> darktable installation. >> "activate opencl support" is checked and I cannot uncheck it. >> Does this mean I am forced to use OpenCL or does it mean my GPU does not >> support OpenCL at all and the disabled control shows a default value >> that's irrelevant for me? > It has been disabled. Newer dt versions do a quick test at startup for OpenCL. > If the test fails, you can't enable OpenCL. O rly? There are two tests - availability of opencl, and opencl speed. If the first one fails, then you have no opencl, and can not enable/disable opencl. If the second-one fails, then opencl is simply disabled by default, and you can enable/disable it at any time. So if the opencl toggle is untoggleable, that means you have no opencl at all. Be it due to no opencl drivers, or if opencl driver is cpu based, or if your hardware does not support all the required functionality. See darktable-cltest. >> >> I am using an nVidia GeForce 9800 GT with the video-nouveau driver, but >> wouldn't mind switching to another driver if that would speed up darktable. >> >> Rico >> > Christian Roman. >> >> >> darktable user mailing list >> to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org > > > darktable user mailing list > to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org > darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Hi Am Sonntag, 27. November 2016, 15:32:24 schrieb Rico Heil: > This discussion made me check the OpenCL paramters in my current > darktable installation. > "activate opencl support" is checked and I cannot uncheck it. > Does this mean I am forced to use OpenCL or does it mean my GPU does not > support OpenCL at all and the disabled control shows a default value > that's irrelevant for me? It has been disabled. Newer dt versions do a quick test at startup for OpenCL. If the test fails, you can't enable OpenCL. > > I am using an nVidia GeForce 9800 GT with the video-nouveau driver, but > wouldn't mind switching to another driver if that would speed up darktable. > > Rico > Christian > > > darktable user mailing list > to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
This discussion made me check the OpenCL paramters in my current darktable installation. "activate opencl support" is checked and I cannot uncheck it. Does this mean I am forced to use OpenCL or does it mean my GPU does not support OpenCL at all and the disabled control shows a default value that's irrelevant for me? I am using an nVidia GeForce 9800 GT with the video-nouveau driver, but wouldn't mind switching to another driver if that would speed up darktable. Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
On domenica 27 novembre 2016 11:20:59 CET, Ulrich Pegelow wrote: Intel graphics does not have a working OpenCL implementation for Linux systems. There is the free beignet driver around since quite some time but that is still buggy and so far nobody has been able to get it working with darktable. I did and it works quite well. Unfortunately it's actually slower than the CPU, so I always keep it disabled. Niccolò Belli darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 27.11.2016 um 11:00 schrieb Rico Heil: Am 26.11.2016 um 18:11 schrieb Niccolò Belli: You will probably get better performance saving some bucks on the CPU and buying a very fast GPU for OpenCL acceleration. Something like the RADEON RX 480 would be an optimal solution because of FOSS drivers, but you will have to use the AMDGPU-PRO driver until Clover Image support gets mainlined (hopefully soon). I was planning to use Intel chipset graphics without buying an additional GPU at all. Can darktable use OpenCL with those? Intel graphics does not have a working OpenCL implementation for Linux systems. There is the free beignet driver around since quite some time but that is still buggy and so far nobody has been able to get it working with darktable. If yes: how much performance improvement would be expected if I add an additional graphics board? I doubt that integrated graphics chips will come close to the performance win you can expect from a decent NVIDIA or AMD graphics card. A modern graphics card in the cost range of 200 to 300€ will probably give you a speed-up during export of 3 to 10 times during export, compared to a modern CPU. Of course this depends on the modules you are using. The more computational expensive the bigger the gap. IMHO the most important point is the latency you experience during interactive work when you change module settings. YMMV, but here it makes a quite noticable difference; between waiting a bit with each parameter change and a mostly fluent way of working. Ulrich darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 26.11.2016 um 18:22 schrieb Robert Krawitz: > The PassMark (http://www.cpubenchmark.net) number for the i7-6700K is > 11,044; for the i5-6500, it's 7072. My experience has been that this > provides a pretty good measure for heavily multi-threaded > CPU-intensive workloads, which darktable is in large part. For > operations that don't parallelize as well, the single thread PassMark > number is also of interest; for the 6700 it's 2341 while for the 6500 > it's 1945. The upshot is that you should expect somewhere between 20% > and 55% improvement with the 6700. Thanks a lot for those numbers! This makes an objective choice much easier :-) > There are faster CPUs. The i7-5820K, for example, has a PassMark > number of 12980 and a single thread rating of 2010. That's because it > has more cores (6 vs. 4). The i7-4790K is 11187 while its single > thread rating is 2527 (which is the fastest available). Yes, that one's really nice - but I don't think this will be the kind of computer I am buying ;-) > Offhand, I'd say that the i7-6700K would be a very good affordable > choice if you're doing a lot of photos. Make sure you get plenty of > RAM (at least 16GB). An SSD for your root+home directory would be a > good idea, but for your bulk image storage, it won't help you as much. I am planning to buy 32 GB of RAM and of course the system will reside on an SSD. > The RAW processing is not especially I/O-intensive; 10x20MP images is > only about 250MB in and 100MB out, so call it 350 MB/minute, or 6 > MB/sec. This I/O streams well, so you're nowhere near limited by disk > I/O (typically around 100 MB/sec for conventional spinning media). That's a very good point! I was thinking about buying a larger SSD, so the images I am currently working on could be saved there, while the large masses of images would always have to be on an HDD. But those numbers make me think I should save some money on the SSD and invest it into computing power (whether CPU or GPU still to be decided). Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org
Re: [darktable-user] Chosing the optimal CPU for darktable
Am 26.11.2016 um 18:11 schrieb Niccolò Belli: > You will probably get better performance saving some bucks on the CPU > and buying a very fast GPU for OpenCL acceleration. Something like the > RADEON RX 480 would be an optimal solution because of FOSS drivers, > but you will have to use the AMDGPU-PRO driver until Clover Image > support gets mainlined (hopefully soon). I was planning to use Intel chipset graphics without buying an additional GPU at all. Can darktable use OpenCL with those? If yes: how much performance improvement would be expected if I add an additional graphics board? Rico darktable user mailing list to unsubscribe send a mail to darktable-user+unsubscr...@lists.darktable.org