I installed the dev release using this link Release Automated Builds · Beep6581/RawTherapee · GitHub and now I have a file (it’s not a folder, I can’t open it) in my computer and I have no idea what to do with this file or what I’m supposed to do next. Actually, I don’t know how technical stuff works at all.
Hi, good to see you try even if it’s outside your comfort zone!
If you’re on Windows, did you download the zip or exe? You need the zip, extract it to a reasonable folder, and just run the RawTherapee.exe file from there. Perhaps create a shortcut on your desktop to facilitate this.
If you’re on Linux, download the AppImage. Open a terminal, navigate to the file, make it executable, chmod +x yourfile and then just run it.
For macOS, I’m not really up to speed on the exact instructions.
In the thread linked above I was linked to another place for nightly builds. You have to unzip it into your program files and make a desktop shortcut. There are different builds, to figure out which one find out your CPU’s microarchitecture. If you don’t know, download CPU-Z and run it, look where it says “Code name” once its started. That’s the microarchitecture. IF none match, select the generic build.
IMHO RawTherapee is very easy to handle. I just download the ZIP. Expand it to a folder. Click on the EXE. That’s all. No installation required. With that method you can have several version parallel, and ART if you want.
I don’t know. I have 5.7 installed, but I almost like never use it, I always use one of the dev version. When I click the dev exe the dev version starts. You can try and rename the installed RawTherapee folder.
@marckquz
If you are on W10, you really should download and install 7zip. It’s a kind of standard for compression.
I use it as the compression is much better than using basic zip.
Then you can extract the directory where you want.
Regretfully, the automatic builds has some hiccups that prevent to build the last dev version.
The last dev version is 29 commits ahead of the official nightly build.
That really matters for me as first, I can keep a few builds (5XN) on my machine, second my upload times are less and third, for those who have a slow connection it will be more comfortable.
GCC can build for different target microarchitectures. The available instructions set (for instance vector operations) are thus used accordingly, permitting a specific optimization for the considered architecture and thus improving the speed. Don’t ask how much, as I cannot make benchmarks as some builds cannot work on my machine.
What I can say is that on my i7 6700K which has a Skylake micro architecture, the skylake build is significantly faster than the generic.
Now, there is a drawback as Intel tried to mess with microarchitecture names (which define the instruction set) and microprocessors family names. Even the GCC dev seem at loss.
For Ryzen 2 architecture it’s simpler, take znver2.
For Ryzen 3 there exist a znver3 option that I can build if requested.
Yes, but not so simple, as in a generation you can have different architecture I think. Icelake client is not applicable for instance to all 10th generation.
10th gen was a disaster, a bunch of them were just rebadged/mildly tweaked 9th-gen due to all of the problems Intel had with 7nm rollout (or was it 10 that they faceplanted on? I can’t remember…) - for the most part 10th is an anomaly in that regard