Hmm. I could’ve just hardcoded all the variables of course, but I decided to intentionally generate all the boilerplate constants to make the generated code more easy to read.
I use simple reflection to analyze the template class, extracting the fields for it.
Oh, and execute(int fbo) (or whatever args the generator puts in there) is supposed to be private. x___x Gonna go fix that. =P That should be enough to make it clear that the private execute() method is meant to do work.
Added CSS to the SilenceEngine GWT Tests page.
http://i.imgur.com/u3EMRS0.png
Click for full size. Night for today!
I drew my first Vulkan triangle! ;D ;D ;D
Added move-to-mouse-click to Vangard, no more watching the paint dry as the sprite slowly rotates onto a new bearing 
Also added a tranlucent radial popup menu to Vangard… although currently only the menu is displayed and not the menu items yet.
Good job! Doesn’t look like a Rubiks, maybe a shengshou? That’s the most common speed cube 4x4 that non cubers buy usually. You should consider getting a 5x5 or some cuboids (3x3x2 etc). Lots of fun to be had
No idea what it is. I got it at a xmass market in Vienna.
My cubes. The 4x4 is harder than a 5x5 of course. Since you get parity. The 5x5 is a rubics and is crap.
Awesome! Yeah so what you’ve got there is a gear cube and a mirror blocks cube. The mirror blocks is solvable just like a 3x3 like I’m sure you know, but the gear cube is a bit harder. I wouldn’t even try to design commutators or anything for the gear cube, all you really need to solve it is maybe a couple re-orientation algorithms. It’s mostly an intuitive solve.
I received this two days ago and finally finished the solve today google image because I’m too lazy to upload a picture.
http://thecubicle.us/images/500x544xlimcubedreidelb2.jpg.pagespeed.ic.P4kjtkDeCk.webp
So the last couple of days were interesting. I implemented fluid/gas simulation using the Lattice Boltzmann Method.
Unfortunately I dont get the math behind all of this. (Never had a PDE course yet, second university semester starts next monday). That means that i derived the code from the most “concrete” equations in papers i could find. This is the result:
Tvoj15jq_f0
Interestingly, the motivation for this was creating a game (as always). I wanted to make an Element-bending game in which you could control air, water, etc. like in the Avatar: The last airbender tv series. Lets see how far this will get. I don’t even care whether I finish someting playable. This was plenty of fun already
So after much mucking around and playing and saving and… Anyway now my pipeline loads texture data directly from the blend files correctly. Still some funk with normal’s but that was fixed by re generating them, and i need to manually create binormals. The result however is now we have all the textures!
There is plenty to do.
As for fluids, i added fluid simulation to Art Of Illusion way back. AoI was the java version of blender. However it was mostly one guy who is now rather busy. I didn’t do a great job and abandoned the project. I used Smoothed Particle Hydrodynamics (SPH). I find it worked better for free surfaces. I would do things very differently if i did them again.
Lattice Boltzmann methods are pretty cool. The math even cooler. In the old days they even had integer latices gases. That was for back when floating point units were slow and expensive. These days i would look for a lib that uses the gpu. Hard to make it the same everywhere however.
Blender’s blend file or FBX? o:
[quote=“delt0r,post:4129,topic:49634”]
Can’t this be done in the shader? I’m assuming you’re using them to calculate a TBN matrix for your normalmapping. Though, I guess I’ve you’ve already went ahead and done it on the CPU then there’s no reason to switch 
Today I come among you, a great community.
These days I watch the complete code of SilenceEngine, and now I can not say that I get to make my own game engine without copying.
SHC did it so well, with such a good architecture… so I’ll never can said “I’ve made my own GE !” :s
This is with jMonkey engine. So directly from blender files.
I am just using the default jME system. I need to call the create binormals thing. Since default textures (spec & normal maps) is all we need right now.
This week i am adding a proper IK system to jME that will also fix some of the other blender import issues. going to be all FABRIK, much like unity. But i get to match constraints to the blender system.
Hey Kevin,
Welcome on to this forum, And thanks for your comment. I already spent nearly two (one and a half, November 16th is it’s birthday) years on the engine, and I got my hands into low level things, and now it’s time to make some games with it.
Doing LudumDare this time?
Yay! I got a spot as a summer intern software dev. There was only one position available and at least 10 other people applied, so I’m very happy about this. Plus it’s good money too.
Hopefully I’ll learn a bunch
Also got a new one handed best (solving the 3x3 one hand): 20.44! My first sub 20 OH solve will be any day now. Nationals is looking good.
It’s mega cool for Texturing algoritm (in future ^^)
SadFace…
up:
Link share 
Estimated delivery: Nov 2016
X86 - Win ^^
IMO: Little overpriced (you can buy for same price cheap smartphone with onboard screen battery etc)
but its intel supports x86
Nice project, glad to see so much italy in it, they will offer back an icecream as a reward if you visit them ;D
But when I look at the gpu here, Braswell HD Graphics 400, it has 4.3 on Windows but only 3.3 on Linux… wtf Intel ???
It is known the linux driver team of Intel is much better than the windows’ one, but they should also keep it up…
Well, at least ES 3.1 on Android, not bad at all
Interesting board, thanks for sharing.
I see that the power consumption is said to be between 5 and 6 Watts which is only double or triple the ARM boards.
Here are the board’s processor specs for the Intel® Celeron® Processor N3160 (2M Cache, up to 2.24 GHz).
Interesting to see that it’s a Celeron which as a desktop has a bad reputation for being slow but comparable to the ARM boards it must be relatively faster. I notice that this incarnation of the celeron was released in 2016Q1.
I hope this UDOOx86 board takes off and gives rise to lots more competition. One of the pains with the ARM boards is that few programs work with them. Even java is a pain to run on ARM since Oracle attempt to charge a licensing fee for the embedded Oracle JDK if you distribute a commercial program, unlike the regular JVM.
Also, I haven’t been able to find an ARM64 headful (graphics capable) Oracle JVM, embedded or not. This is not a big problem since the openJDK arm64 JVM works headful but I suspect that it’s slower than the oracle one.
Having a fast and power-efficient x86 board would be great from a software point of view since ordinary java will work on it. Also, the linux operating systems on x86 are fully functional and stable, unlike arm64 which is a big headache. For example, Chromium browser doesn’t work on Ubuntu arm64 at the moment, among other driver and OS problems:
http://forum.odroid.com/viewtopic.php?f=29&t=20085
One thing that seems strange is this quote from the UDOOx86 kickstarter:
“Quad Core 64-bit new-generation x86 processors”
I thought x86 meant 32 bit?
Given that the processor is actually 64 bit, shouldn’t the whole project be called the UDOOx64 rather than UDOOx86 ???
Can I ask where you’re interning and what you’re working on? That’s great though! Have fun, ask questions if you need to and impress everyone
