« first day (816 days earlier)      last day (410 days later) » 

5:10 PM
@kimholder ha! I've had a similar experience once (it was calculating diffraction patterns) so I know just what it's like. Suddenly I'm sitting there and I... can't... do any of the things I normally do!
Luckily I had found some old SF paperbacks at an English used book store earlier, so I took a little vacation.
Can you use the GPUs of your computer to do the rendering (assuming that's what you are doing)? It's a lot faster than the CPU but only some computers and some programs can do it.
I just saw this article. It talks about high temperature tubes heated by sunlight and I remember we talked a bit about that a while ago after one of your videos.
 
5:32 PM
@uhoh i was using image magick, as the images are too big to even open on my computer. I don't know if it used my cpu or my gpu, but the bottleneck in most cases anyhow was that it was using a special format that memory mapped the image to the hard disk, because otherwise the operations were aborting.
that's where this odyssey is at right now. A moon with exaggerated topology that is as detailed as my computer can take in real time, but not really very satisfying. The image used to generate the topology has twice the resolution as anything else available, but my computer has trouble handling it if i smooth it out and give it the vertices it needs to display well.
@uhoh Tubes based on nickel alloys? I wonder if they would hold up at the 1300 K temperatures the plant has to use to melt the regolith. At some point good work on materials needs doing, alright.
 
5:50 PM
@kimholder Oh! This is the file you mentioned that you were downloading a while ago!
Gigabytes or something
 
yeah. 92,160 px by 46,080 px, 8.5 GB
and that was the consolation prize, after i couldn't do anything at all with the one that is 22 GB
 
Yikes! My mind is suitably boggled by that
 
the processed images are now available for download on moonwards - moonwards.com/project.html#assets
but it's the tiles that are mostly what's practical.
i am just in the process of trying to get a 8192 x 8192 tile of the region around Lalande Crater working on a section of a sphere that maps to the area it covers
 
Geez you are having fun!
 
it'll feel a lot more like that when it works...
 
6:02 PM
I think I see. These are heigh maps where the height is encoded in the pixel brightness?
It's 2AM and I'm signing off soon, so let's not go into detail right now.
I'm just thinking that it would loose vertical resolution if it becomes only 8 bits or so.
Oh, the USGS website says 16 bits astrogeology.usgs.gov/search/map/Moon/LRO/LOLA/…
okay gotta go... chat later
 
yeah, the originals were 16 bit tiffs, but interestingly, they only actually had 157 colors
i suppose they encoded them that way so they would be compatible with GIS software? i don't really know.
but, the histograms don't lie. not many colors in them.
so i do have versions that were resizes as 16 bit PNGs, and in the process, a large number of new shades of gray were created.
but the 65536 x 32768 version of that is 1.5 GB. I had it posted for download for a few days, and then i figured i was offering a huge file that was barely more accurate, if at all, than the one that is 8 bit and still has all the original colors.
so i replaced it with the 8 bit one that is only 148 MB
 

« first day (816 days earlier)      last day (410 days later) »