PaulAffinity Posted July 4, 2017 Share Posted July 4, 2017 Problem is an INPAINTING that takes around 12 seconds in Live, takes nearly 30 seconds in Beta.75. Identical areas selected. Is this extra Beta code or an issue? cheers, Paul malyKeri 1 Link to comment Share on other sites More sharing options...
Staff Sean P Posted July 6, 2017 Staff Share Posted July 6, 2017 Hi Paul,Just tried this on a file myself and the different was about 4 seconds in Release and 7-8 in Beta so didn't seem too bad, but 20seconds extra is a bit much! Do you have an image you could attach (along with the area you were trying to remove)? I can then take a look and check with development. Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 6, 2017 Author Share Posted July 6, 2017 Hi Sean P.... While 4s is not a big deal, that is 100% time increase, which is massive. I will send later. Cheers P Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 7, 2017 Author Share Posted July 7, 2017 Attached image - I used INPAINTING on Just the guy at the centre. Single select with a few pixels around. Let me know, cheers, Paul Link to comment Share on other sites More sharing options...
AlainP Posted July 7, 2017 Share Posted July 7, 2017 I juste mad a test using PaulAffinity's file. Inpainting is faster (and does a better job) in beta 1.6.0.75 than in the official release .69, approx 5 seconds vs 8 seconds on a quite basic and not very fast i3 computer. It's probably a graphic card/drivers issue.... -- Window 11 - 32 gb - Intel I7 - 8700 - NVIDIA GeForce GTX 1060 -- iPad Pro 2020 - 12,9 - 256 gb - Apple Pencil 2 -- iPad 9th gen 256 gb - Apple Pencil 1 -- Macbook Air 15" Link to comment Share on other sites More sharing options...
Staff Sean P Posted July 7, 2017 Staff Share Posted July 7, 2017 Same for me! The results for me were very quick on both and about the same. Paul,When painting with the inpainting brush are you using a tablet device? If so does using the mouse make a difference to the time? Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 7, 2017 Author Share Posted July 7, 2017 25 seconds using the mouse... P Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 7, 2017 Author Share Posted July 7, 2017 ! 26 seconds using 'WARP'..... Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 7, 2017 Author Share Posted July 7, 2017 Also, just as an addendum; If I use Ctrl+0 fit the image and do INPAINTING - 27s now using AMD Radeon in .75. If I ZOOM and fill my screen with area to INPAINT using same settings takes 21s. Big question is for me - what has change in the Code Base between LIVE and .75 (or prior betas - which I have not used for a while) regarding either GRAPHICS DRAW or INPAINTING. Cheers, Paul Link to comment Share on other sites More sharing options...
tkane18 Posted July 12, 2017 Share Posted July 12, 2017 I downloaded your picture and painted out the guy on the rocks in 3 seconds on beta .75. I tried various zoom levels while doing it and the results were the same. I'm running an HP laptop with an i7-4810MQ CPU @2.8GHz with 16GB of RAM with Windows 7 SP1 64-bit. The GPU is an AMD FirePro M6100 FireGL V. Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 13, 2017 Author Share Posted July 13, 2017 While I appreciate your testing posts, I have a fundamental issue. Let me explain thoroughly (as an IT specialist of 35+ years); 1. I use Win 7 sp1. I do not update this with ANY Windows updates. 2. The hardware has not been changed in 2 years (SSD replaced an HD) 3. All drivers unchanged for 4 years. 4. No software addition or change since AP and AP .75 beta. So, I am using identical machine configurations as I have for years. IF beta .75 is x00% slower than Live AP on my machine I can only assume that the CODE base for .75 (or a prior beta - as it is age since I tested an AP beta) has fundamentally changed. My simple question is has it? If so is that change to inpainting or a called routine from it that either computes, calculates or draws? Serif software guys please? Cheers Paul. PS. Remember Wacom or mouse no discernible difference and Warp or Built in Radeon likewise. Link to comment Share on other sites More sharing options...
harrym Posted July 14, 2017 Share Posted July 14, 2017 Hi Paul, Sorry I'm obviously not staff and it's them you really want to hear from but I'm curious about something. Assuming with that set up your machine is not connected to the internet, do you have AV on that machine? if so what is it and does switching it off then re-starting AP change the timings any? Regards Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 14, 2017 Author Share Posted July 14, 2017 Hi Harrym, No AV. As it is not web connected I have 'Malwarebytes' but freemium, so no real time malware checking just ad-hoc. As ever if the 2 versions running on identical machine configurations have significantly differences in performance I can only extrapolate that the code base has changed. Intrigues me what level of machine (specification) based regression testing is undertaken on betas? Several questions to answer please Serif? Cheers Paul Link to comment Share on other sites More sharing options...
harrym Posted July 14, 2017 Share Posted July 14, 2017 Hi Paul, The reason I mention AV is that when I use DxO Optics Pro to process RAWs I have to switch my AV off (and internet obviously!) AV kills it, no matter what exclusions I try so I'm always suspicious of it. In fairness AP does play well with it though. Looking through this forum the recurring theme is speed/performance in the Win environment. I have no doubt that Serif are working real hard on this aspect. Now total speculation on my part but I've noticed on the latest beta more screen flashing for want of a better description between certain processes almost like they are trying different ways to manipulate/use machine resources. Up to .75 I've found a marked difference in performance (to betas advantage) but if I'm honest I'm not sure about .75 it just feels less stable. I'd image that beta and live are using different code base I'd also image that in dev they have a large number of code branches too! I'd really like them to release a new live version that's stable and no debug with the same document structure as the current beta so there is no potential compatability issues. They could then move forward with beta developement with 'less' concern over something unstable or underperforming getting released, it is beta after all. Of course caviat to that would be that the document structure is getting changed for performance reasons so it's pointless releasing a live update right now! Who knows, I obviously don't! :) Regards Link to comment Share on other sites More sharing options...
PaulAffinity Posted July 14, 2017 Author Share Posted July 14, 2017 Hi Harrym, IMHO AV software causes a significant amount of spurious issues with software applications. Never use the stuff! My query over code base is linked to Regression Testing. As a 'legacy hardware' user I am concerned that; 1. Serif are prioritising performance improvements based upon Intel and i3/5/7 architecture. 2. AMD GPU/APU will suffer or worse become unsupported. Your comment reply 'screen redraw' or flicker implies some performance based recode of that component. I switched to AP to save £s and support a company I have long ties with, but if the cost is a new machine to do what PS already does then it is either PS or stick with the current Live AP release and never upgrade.... Cheers, Paul Link to comment Share on other sites More sharing options...
Recommended Posts