I just ran a Calibrate (using the GAI standalone vers,). It changed my Prefs to GPU = Off, OpenVino = On. With GPU off I didn’t get any crashing events. I tried a couple times. Hopefully the logs above will help figure out why GPU crashes post AMD Radeon driver update & didn’t prior to that using the same GAI rel.
Here are the current Prefs:
With GPU turned off, so not crashing, Standalone vers processes relatively quickly. Plugin (Ps 2021) still lags and processes only in 7% increment hops.
Also - Question:Why has GAI started squishing my visible Orig & Preview up to a collapsed format while the upscaling is in progress and completed? I’m working with a single image and I’d prefer it remain visible. I could understand if the default was to collapse if a batch of 10+ images is being run. But I’m only working with a single pic and would like it to remain open and for me to be able to go back to it if desired.
I have unistalled and re-installed a couple of times, When I try to open I get this system error message.
“The code execution cannot proceed because CuteLogger.dll and lensfun.dll was not found. Reinstalling the program may fix the problem.”
In Widows Explorer you can copy a photo and paste it back into the same directory several times. It will automatically add the word copy and a number to the file name. You then open all the files in Gigapixel and process each one as you like and save.
I checked my install on my Windows 10 machine and both of those files are there. I would download the program again and delete the existing install before re-installing. Those files may have been corrupted.
Hi. Are you talking about the crashing or processor bug alert? Wasn’t sure what you’re experiencing too. It might help the devs to know common user results. Ciao.
Okay, I created a recovery pt., but then decided to do another set of runs b4 going thru the hassle of the driver, etc. reinstall
Here’s what happened (using this rel. on Win 10, standalone vers only):
GPU = On, OpenVino = On Resource Util = Medium GAI crashed/AMD Bug Msg Gen’d
GPU = On, OpenVino = Off Ditto Ditto
GPU = On, OpenVino = Off Resource Util = High Ditto
GPU = On, OpenVino = Off Resource Util = Low Image processed 15-20 mins.
(avg GPU % util varied from 60-100% throughout that run)
GPU = Off, OpenVino = On N/A Image processed faster than GPU
but still not fast.
I’m just going to leave the GAI prefs set to CPU and OpenVino & not mess around w/driver re-install based on the test/perf results. Typically the processing runs in the GAI Ps plugin are slower than the standalone. And I only work with the plugin b/c my entire image processing workflow is in Ps - I don’t want to be hopping in & out of misc external programs…
It’s not - it’s a design trick to make the program’s UI feel more responsive. It’s hard to explain without going in-depth on UX design principles.
This’ll most likely go away late January / early February when the new AI engine gets migrated into Gigapixel. It’s being migrated into DeNoise right now, and processing times are in the seconds and not minutes, so that processing panel will no longer be needed.
I’ve loosely followed the development of Gigapixel AI for some time and it seemed like a good time to finally buy it on this Holiday Promo. And so far so good
A small detail I would have altered is the File function Open Recent (list of a few recent files). It is a rather short list to be of that much use, and an option to turn it off altogether for those who do not wish to have it would be good idea.
Out of curiosity I digged into the Windows registry and found the key “recentFilesList”, but modifying its value can make Gigapixel AI unable to load more than one file at a time. So it’s best not tampered with.
Another option I’d like to see is to have the Preview function turned off by default, instead of enabling itself every time a file is loaded. For those who rarely use the previews.
WHAT! WHAT! WHAT! Are you serious!? You mean that I no longer have to take a 1 hour coffee break when processing 600 photos? “Processing times of seconds not minutes” is something we wouldn’t be able to achieve no matter if we throw lots of money at buying new hardware. And you’re saying the new software update will deliver this? This is the biggest news of 2021 if I understood this right!
I feel like there’s way too much people complaining in all the release threads. I just wanna say that you guys are AWESOME!
Will the AI models need re-training for the new engine?
Also, I noticed that Gigapixel currently has tensorflow internally, so you’ve got something better than Tensorflow in the new engine?
“Processing times of seconds not minutes” is something we wouldn’t be able to achieve no matter if we throw lots of money at buying new hardware. And you’re saying the new software update will deliver this? This is the biggest news of 2021 if I understood this right!