I wasn’t planning to apply until about 3 weeks before , really was quite pressured. Have to sit the MAT in a few weeks aswell , that is going to be interesting…
Hope tomorrow sort all my HDD Data and finally move it To VHD
Because Window f**k up Disk with 500 gb all over the time (especially git folder XD)
first I even think its problem with my HDD – but no, I test it on bad sectors in DOS and
it find only 1 bad sector
I have 70 gb free space NTFS but once i fill 499 GB and it was start of all problems
now Windows can’t defragment it in background
(manual I even don’t want try - it’s take 5 or even 10+ hrs)
today hi finally kill my patience when I try copy 1 gb file on same disk(500 gb-70 gb free)
and hi give me 500 kb/s write speed
Why VHD? because its easy to backup – copy 1 file, faster read write, even may have dynamic grow Size, but need separate important data to different VHDs to not copy all, one big piece.
but VHD have 1 problem if Win f**k up this 1 big file - chances that hi restore it almost 0 ^^
but i use VHD for backup - it easy to do copy, so even if this happen data can be restored from copy
or maybe google how restore broken VHD =)
Logical copy better then hope in win chkdsk:
that maybe hi restore all broken files
- no he don’t – I check many time in 50% cases hi simple lost files
p.s Reminder for self – Never-ever do Logical disk with 100 gb’s+ on old file system like NTFS
except maybe Windows Sys – because it writes all crap without asking - at same disk
I watched a presentation about JEP295 and the plans they’ve got for AOT compilation a while ago. I can’t help but think that they’re just sort of barking up the wrong tree and they should have just bought Excelsior, which does exactly what everyone wants.
Cas 
Is anyone studying at the Technical University of Cologne, Germany and applying for the Computer Graphics course? (I didn’t)
Found this video tutorial (in German) about basic linear algebra using JOML, and… log4j. (wonder how they go together)
I always had extremely bad marks in German classes. I understood (most of) the 5 minutes of that video that I watched. I feel like a genius.
Microsoft: 'Do you really think you can run away so easily - ;)”
VHD – dynamic have 25% performance and creating static takes eternityyyy 2 Gb created in 2 min or so
Yes it definitely faster to copy-paste - especially full Git rep – 1 file
But this not save you from nightmare of defragmentation
(because inside VHD you still have same Logic disk =))
P.s IMO Best way to copy Git is: -bundle or git -archive for only data – but VHD Faster XD
pp.s try Deferment disk 8 gb it take 2 hrs 
no I event don’t want try it on 500 gb
IMHO: Best way defragment disk is – copy all important data to another disk – and let it BURN
Recreate Disk Volume – with full format XDDDDD
Up:
Recreate 500 gb partition - separate it
- all backs to normal Write speed 20-50 mb/s on big file =)
not bad for 7 year old HDD in everyday Use XD
Went back to programming after a long break, got my old 80’s radio working, and I’m going to get my gramophone repaired. Yay!
Updated my fractal generator to use color schemes! (original post: here)
It’s also much faster and utilizes OpenGL.
Or1sjXI9708
Next steps??? (I’m thinking of trying to get past the zoom level inhibited by the precision of a double).
Are you sure? That’s quite a trip down the rabbit hole :persecutioncomplex:
The last of the art we need is rolling in and the game is starting to look “real.” It’s pretty exciting.
I think the especially cool thing about this game is that it runs on the same engine that We Shall Wake does, and performs exceedingly well at both styles of development.
I’m sure it is quite a lot to consider. I still have to research more about a few methods I am considering and whether it is really feasible with my current approach. Problem is, I am restricted to whatever features glsl has to offer…
Edit: The main reason I proposed this idea is because I am at a loss for what to work on next ???
Ahh, it feels good to have completed BSP tree collision
Also spent the past few days abstracting the whole process:
Currently supporting AABB, spheres, and ellipsoids.
This collision includes sliding along the collided surface, stepping up stairs, jumping, and gravity
Updated my lighting system to produce a nice inverse square type spread , came up with this equation to model it
where k is the range of the light and p is the power. This functions so that it doesn’t require a square root calculation and that y<1 when x = k.
Can someone explain me this?
http://pastebin.java-gaming.org/baee259814e1d
Up: Caution MAY freeze System on couple min’s (it create only 2 gb)
but it may lags on not clean Disk!!! - because of fragmentation or other background IO operations
(sorry for code it’s very raw example)
It create and write 2 GB Data
for 3.1679208 sec its 631 mb/s
p.s if you have same results and it not some trick (that i don’t understand at this point)
pp.s maybe i made some mistakes in code that can explain such result)
Up Solved: Trying create 5 files for 2 gb with 4 Gb Ram - And its freezes after Write 3.5 Gb =)
and write next very slow == it’s some system(Java) cache that give only illusion of such cool speed
(Yes-Yes MappedByteBuffer.force() - but it freezes whole system for me on couple min’s 
)
Most interesting is: this Write IO is in - some low Core Sys
- because none of Sys profilers(and external profilers) not show any IO activity at all but all lags as Hell untill All IO ends
Submitted my very first college application - Early Admission to Georgia Tech. 15 minutes before the deadline. Here’s to a smoother and more timely process for future apps. Cheers.
Been trying to start learning libGDX. Setting everything up for the first time is… definitely not easy.
It gets better, when you’re on your second project it starts to make sense.
Updated my algorithm a little bit , gives a nicer result with light regions getting larger on the darker colours.
with range being defined as plnp + p.
This produces the nice result shown below.
It would be nice to make this faster , as lighting calculations don’t need to be exact is there any decent and rapid approximation to ln p , of course p^2 ln p can be approximated but the only implementations I have found don’t have much explanation.
I don’t know if this makes sense but I always try to keep as little divisions as possible, so I’d multiply everything by p/p, to give something like [i]p/i² / (px² + (plnp)² )
As for a fast way to deal with ln, I guess you could try the usual series lnx ~ x-1 - 0.5(x-1)(x-1) + 0.33(x-1)(x-1)(x-1) - … and shorten it to the smallest number of terms that would still produce acceptable results…
Ps: not gonna use latex sorry not sorry