I just replied to you on Flickr as well.

One additional point to note is that a lot of the processing is being handled by Gegl. While I have optimized it a little, it may appear slower than the 8-bit counterpart.

Thanks Partha. As I said at flickr, I did a reinstall because the 2.9 install messed up my 2.8.2 install.. 2.9 still gives me a resolution error on some tifs when I open them. (The ones I've tested are dng's converted using Irfanview)..

I assumed it would be slower with gegl and 16bit, but this is definitely too slow.. Hopefully still room for improvement. Could you help me understand why it's this slow in gimp, yet when in delab for instance it's not slow at all? Even when I load jpegs in 2.9 gimp, it's painfully slow for every function.

I'll try to hold out before moving back to 2.8.2 just in case there is anything you want me to check/test.. I'd like to have them both installed? Any recommendations for that? Installing 2.9 definetely hurt my Partha 2.8.2 install. I was getting errors (lib something dll errors). I wish I would have taken a snap of it but I didn't.

I responded to this on Flickr.

This should not happen. Jpegs are 8-bit and so loading should not be an issue. Definitely room for improvement. 2.9 is work in progress and will remain so for awhile.

Delaboratory uses a completely different mechanism for loading files. It's written from the ground up to deal with 8/16/32 bit color. Gimp is loaded with a lot of legacy baggage which will take time to fix.

Sorry about that. Definitely urge you to use 2.8.2 for production work and 2.9 for playing. Please use the portable version of 2.9 so that you won't have to mess your 2.8.2 installation. 2.9 does not touch your personal space (.gimp-2.8) and so will not have messed your personal stuff. During first run 2.9 simply copies over your settings.

cool thanks.

Just installed the 64 bit portable build to look at. Might convert one of my RAW files tomorrow to a 16 bit TIFF to see how well it handles it. I was converting them to 8 bit TIFFs after seeing the obvious failure of 2.6.x with 16 bits but these days typically just open the RAW file in photoshop and save to PSD for the archives and JPG for posting.
All of the images I have used in the GIMP lately were JPGs to start with; mainly the original images on FARK contests.

I just downloaded Partha Gimp 2.9 portable, some of the filters have a G in front of them, I'm not sure why? I guess I need more coffee

G means those plugins are operated by GEGL.

GIMP 2.10 will be fully ported to GEGL libraries, it won't be released until then. Right now GIMP porting to GEGL is about 20% accomplished. GEGL is what is going to allow GIMP to work with 16-bit and 32-bit floating point depth precision.
The problem is that all C plugins have to be re-coded to be able to work with GEGL.
Old, not-maintained plugins like UfRAW, GAP, Mathmap, Separate+ etc. will not work in 2.10.

Not if you use PAE

I have updated all builds (Mac, Windows 32 bit, Windows 64 bit, Windows regular and portable versions) with the latest source.

Now the builds do not interfere with your existing 2.8.2 builds and so you can have these builds side by side or use the portable version which includes the plugins. The plugins include the latest G'mic 1.5.3.0.

Thanks for all your support!

Partha

Does this mean I can download to my macbook and it won't interfere with my mcgimp/gmic?