Mykhailo Moroz
@michael-moroz.bsky.social
Love physics simulations and rendering algorithms. Currently work at
ZibraAI as Lead 3D Research Engineer. Discord: misha.m. michaelmoroz.github.io/about/
ZibraAI as Lead 3D Research Engineer. Discord: misha.m. michaelmoroz.github.io/about/
Reposted by Mykhailo Moroz
I am making clouds for @skipsy.ca's new VRChat world. The clouds render faster than the mountains lol
#VRChat
#VRChat
November 8, 2025 at 8:49 PM
I am making clouds for @skipsy.ca's new VRChat world. The clouds render faster than the mountains lol
#VRChat
#VRChat
Slang can be made extremely close to GLSL/Shadertoy style of shaders by using a set of defines and typealiases, I've made a template in compute.toyts just for that compute.toys/view/1947
I also ported my most popular shadertoy using it:
compute.toys/view/1948
I also ported my most popular shadertoy using it:
compute.toys/view/1948
May 11, 2025 at 2:37 PM
Slang can be made extremely close to GLSL/Shadertoy style of shaders by using a set of defines and typealiases, I've made a template in compute.toyts just for that compute.toys/view/1947
I also ported my most popular shadertoy using it:
compute.toys/view/1948
I also ported my most popular shadertoy using it:
compute.toys/view/1948
Compute.toys finally got the Slang compiler integrated, which means its finally time to ditch WGSL 🎉🎉
I've ported my FFT to it, coding in it is much nicer. compute.toys/view/1922
Should try something with its autodiff next.
I've ported my FFT to it, coding in it is much nicer. compute.toys/view/1922
Should try something with its autodiff next.
May 5, 2025 at 2:24 PM
Compute.toys finally got the Slang compiler integrated, which means its finally time to ditch WGSL 🎉🎉
I've ported my FFT to it, coding in it is much nicer. compute.toys/view/1922
Should try something with its autodiff next.
I've ported my FFT to it, coding in it is much nicer. compute.toys/view/1922
Should try something with its autodiff next.
Made a mixed radix FFT implementation in compute.toys/view/1915
It is neat, but I still wonder how do you properly handle large prime factor size FFT's?
It is neat, but I still wonder how do you properly handle large prime factor size FFT's?
Faster Mixed Radix FFT
Forked from https://compute.toys/view/1914
Faster version
compute.toys
May 3, 2025 at 2:37 PM
Made a mixed radix FFT implementation in compute.toys/view/1915
It is neat, but I still wonder how do you properly handle large prime factor size FFT's?
It is neat, but I still wonder how do you properly handle large prime factor size FFT's?
I've ported my vrchat isosurface ray tracer to shadertoy and added it to my fluid simulation there.
Its probably the fastest multibounce ray tracing for liquids I've made thus far.
shadertoy.com/view/w3sXzs
Its probably the fastest multibounce ray tracing for liquids I've made thus far.
shadertoy.com/view/w3sXzs
April 23, 2025 at 3:45 PM
I've ported my vrchat isosurface ray tracer to shadertoy and added it to my fluid simulation there.
Its probably the fastest multibounce ray tracing for liquids I've made thus far.
shadertoy.com/view/w3sXzs
Its probably the fastest multibounce ray tracing for liquids I've made thus far.
shadertoy.com/view/w3sXzs
I posted my own implementation of the trilinear isosurface intersection on shadertoy
2 methods, one is the one from jcgt.org/published/00... and the other one is my own idea of using 4 trilinear samples (can use HW samplers here) to reconstruct the cubic.
www.shadertoy.com/view/WXS3Rm
2 methods, one is the one from jcgt.org/published/00... and the other one is my own idea of using 4 trilinear samples (can use HW samplers here) to reconstruct the cubic.
www.shadertoy.com/view/WXS3Rm
Shadertoy
www.shadertoy.com
April 6, 2025 at 5:55 PM
I posted my own implementation of the trilinear isosurface intersection on shadertoy
2 methods, one is the one from jcgt.org/published/00... and the other one is my own idea of using 4 trilinear samples (can use HW samplers here) to reconstruct the cubic.
www.shadertoy.com/view/WXS3Rm
2 methods, one is the one from jcgt.org/published/00... and the other one is my own idea of using 4 trilinear samples (can use HW samplers here) to reconstruct the cubic.
www.shadertoy.com/view/WXS3Rm
Reposted by Mykhailo Moroz
Managed to get Nvidia Flow (PhysX) to compile and run, after initially failing the other week. Was curious for a quick peek now that it's been open sourced (github.com/NVIDIA-Omniv...)
Flow itself isn't CUDA based, but instead uses compute shaders
Some higher level observations in the replies 1/?
Flow itself isn't CUDA based, but instead uses compute shaders
Some higher level observations in the replies 1/?
April 6, 2025 at 12:40 PM
Managed to get Nvidia Flow (PhysX) to compile and run, after initially failing the other week. Was curious for a quick peek now that it's been open sourced (github.com/NVIDIA-Omniv...)
Flow itself isn't CUDA based, but instead uses compute shaders
Some higher level observations in the replies 1/?
Flow itself isn't CUDA based, but instead uses compute shaders
Some higher level observations in the replies 1/?
I've been improving my VRChat liquid simulation for the last few weeks. Of the most notable things that I've implemented is a ray tracing algorithm over a 3d texture isosurface (using some ideas from jcgt.org/published/00...)
Its surprisingly performant for VR resolutions!
Its surprisingly performant for VR resolutions!
April 4, 2025 at 7:32 PM
I've been improving my VRChat liquid simulation for the last few weeks. Of the most notable things that I've implemented is a ray tracing algorithm over a 3d texture isosurface (using some ideas from jcgt.org/published/00...)
Its surprisingly performant for VR resolutions!
Its surprisingly performant for VR resolutions!
Is it me or is bluesky video quality absolute trash
March 20, 2025 at 9:16 PM
Is it me or is bluesky video quality absolute trash
I'm slowly porting my approximate grid SPH simulation to VRChat. It did require few improvements to be stable enough, now its a hybrid neighbor index/particle average grid.
Still need to implement
@d4rkpl4y3r.bsky.social 's mip counting sorting method for improved performance.
Still need to implement
@d4rkpl4y3r.bsky.social 's mip counting sorting method for improved performance.
March 20, 2025 at 7:51 PM
I'm slowly porting my approximate grid SPH simulation to VRChat. It did require few improvements to be stable enough, now its a hybrid neighbor index/particle average grid.
Still need to implement
@d4rkpl4y3r.bsky.social 's mip counting sorting method for improved performance.
Still need to implement
@d4rkpl4y3r.bsky.social 's mip counting sorting method for improved performance.
I usually render particle fluids with particles, as it gives the most details, but here I tried using ray marching for the fluid. It certainly is much more liquid like, but at the same time its slower and loses some detail. I wonder if I can improve this. compute.toys/view/1809
March 9, 2025 at 6:02 PM
I usually render particle fluids with particles, as it gives the most details, but here I tried using ray marching for the fluid. It certainly is much more liquid like, but at the same time its slower and loses some detail. I wonder if I can improve this. compute.toys/view/1809
I tried to apply my previous simulation approach to model soft bodies. By keeping the initial positions and using those to create springs between particles with spring length being the distance between initial positions its easy to make a soft body sim.
compute.toys/view/1807
compute.toys/view/1807
March 9, 2025 at 12:01 AM
I tried to apply my previous simulation approach to model soft bodies. By keeping the initial positions and using those to create springs between particles with spring length being the distance between initial positions its easy to make a soft body sim.
compute.toys/view/1807
compute.toys/view/1807
Been recently looking into fluid simulations again. This time I've testing if I can make a good SPH simulation but only having "average" cell particles. It seems to work surprisingly well, and it doesn't need sorting either.
compute.toys/view/1782
Explanation thread👇🧵
compute.toys/view/1782
Explanation thread👇🧵
March 6, 2025 at 2:38 AM
Been recently looking into fluid simulations again. This time I've testing if I can make a good SPH simulation but only having "average" cell particles. It seems to work surprisingly well, and it doesn't need sorting either.
compute.toys/view/1782
Explanation thread👇🧵
compute.toys/view/1782
Explanation thread👇🧵
Reposted by Mykhailo Moroz
On a day like this I recommend donating to Ukraine's military units. @serhii-rieznik.bsky.social has a drive going. A fellow Berliner, collecting for Kraken unit to grab them another drone. I can personally vouch for Serhii as an aid organiser who has done great work.
www.paypal.com/paypalme/ser...
www.paypal.com/paypalme/ser...
Pay Serhii Rieznik using PayPal.Me
Go to paypal.me/serhiirieznik and type in the amount. Since it’s PayPal, it's easy and secure. Don’t have a PayPal account? No worries.
www.paypal.com
March 4, 2025 at 10:50 AM
On a day like this I recommend donating to Ukraine's military units. @serhii-rieznik.bsky.social has a drive going. A fellow Berliner, collecting for Kraken unit to grab them another drone. I can personally vouch for Serhii as an aid organiser who has done great work.
www.paypal.com/paypalme/ser...
www.paypal.com/paypalme/ser...
There was a severe lack of good volumetrics in VRC, so I decided to make my first VRChat world👀
vrchat.com/home/world/w...
The world is based on Room -02- world from どんぐりのお店's BOOTH and the volumetric fog is a modification of this Shadertoy shadertoy.com/view/Wc23Rm
vrchat.com/home/world/w...
The world is based on Room -02- world from どんぐりのお店's BOOTH and the volumetric fog is a modification of this Shadertoy shadertoy.com/view/Wc23Rm
February 24, 2025 at 7:16 PM
There was a severe lack of good volumetrics in VRC, so I decided to make my first VRChat world👀
vrchat.com/home/world/w...
The world is based on Room -02- world from どんぐりのお店's BOOTH and the volumetric fog is a modification of this Shadertoy shadertoy.com/view/Wc23Rm
vrchat.com/home/world/w...
The world is based on Room -02- world from どんぐりのお店's BOOTH and the volumetric fog is a modification of this Shadertoy shadertoy.com/view/Wc23Rm
Reposted by Mykhailo Moroz
Successfully smashing together two neutron stars, to make a #blackhole. The amount of extra resolution you need to simulate a black hole compared to a neutron star is kind of nuts
February 24, 2025 at 12:25 AM
Successfully smashing together two neutron stars, to make a #blackhole. The amount of extra resolution you need to simulate a black hole compared to a neutron star is kind of nuts
Reposted by Mykhailo Moroz
Met with @michael-moroz.bsky.social
February 21, 2025 at 1:09 AM
Met with @michael-moroz.bsky.social
Reposted by Mykhailo Moroz
Today I uploaded my Powder Toy in #VRChat creation as a world, so people who don't know me can play with it: vrchat.com/home/launch?...
I also added 10 new elements, including a bunch of explosives which can push around particles when detonated, via a rudimentary wind simulation.
I also added 10 new elements, including a bunch of explosives which can push around particles when detonated, via a rudimentary wind simulation.
February 17, 2025 at 9:52 PM
Today I uploaded my Powder Toy in #VRChat creation as a world, so people who don't know me can play with it: vrchat.com/home/launch?...
I also added 10 new elements, including a bunch of explosives which can push around particles when detonated, via a rudimentary wind simulation.
I also added 10 new elements, including a bunch of explosives which can push around particles when detonated, via a rudimentary wind simulation.
Reposted by Mykhailo Moroz
I made Powder Toy in VRChat! It supports several kinds of liquids and solids, including chemical reactions between them. It's done entirely with fragment shaders and cameras, so I can put it on my avatar.
February 14, 2025 at 3:03 PM
I made Powder Toy in VRChat! It supports several kinds of liquids and solids, including chemical reactions between them. It's done entirely with fragment shaders and cameras, so I can put it on my avatar.
Almost forgor about this place huh
In any case, here is a video of a superfluid simulation
www.youtube.com/watch?v=uuGR...
As well as the simulation itself compute.toys/view/1640
In any case, here is a video of a superfluid simulation
www.youtube.com/watch?v=uuGR...
As well as the simulation itself compute.toys/view/1640
Superfluid 3D simulation
YouTube video by Mykhailo Moroz
www.youtube.com
January 8, 2025 at 11:31 PM
Almost forgor about this place huh
In any case, here is a video of a superfluid simulation
www.youtube.com/watch?v=uuGR...
As well as the simulation itself compute.toys/view/1640
In any case, here is a video of a superfluid simulation
www.youtube.com/watch?v=uuGR...
As well as the simulation itself compute.toys/view/1640
Reposted by Mykhailo Moroz
Hello everyone! I've been experimenting with some different ways of rendering a fluid simulation, and made a video about the journey over here: youtu.be/kOkfC5fLfgE
Coding Adventure: Rendering Fluids
YouTube video by Sebastian Lague
youtu.be
December 6, 2024 at 11:50 AM
Hello everyone! I've been experimenting with some different ways of rendering a fluid simulation, and made a video about the journey over here: youtu.be/kOkfC5fLfgE
Been trying to figure out a way to get random sequence shuffles cheaply in-place.
Seem to have gotten something good enough
Perhaps its not the best way(and to be fair idk if there is anything better), but seems to work.
Seem to have gotten something good enough
Perhaps its not the best way(and to be fair idk if there is anything better), but seems to work.
December 2, 2024 at 2:13 AM
Been trying to figure out a way to get random sequence shuffles cheaply in-place.
Seem to have gotten something good enough
Perhaps its not the best way(and to be fair idk if there is anything better), but seems to work.
Seem to have gotten something good enough
Perhaps its not the best way(and to be fair idk if there is anything better), but seems to work.
Very cool
I guess the TLDR here would be:
1) Properly utilize memory hierarchy (groupshared cache, more register util by computing multiple elements per thread)
2) Less thread divergence = better
3) I guess using something like sum tables to reduce blur compute amount?
I guess the TLDR here would be:
1) Properly utilize memory hierarchy (groupshared cache, more register util by computing multiple elements per thread)
2) Less thread divergence = better
3) I guess using something like sum tables to reduce blur compute amount?
Let's wrap up this lovely week with a nice technical post
This is the "case study" from my Masterclass at GPC, where I apply a series of optimizations to improve the effective bandwidth of a 3x3x3 blur (a proxy for a huge set of operations on volumetric data)
Check ALT text for (a lot of) context.
This is the "case study" from my Masterclass at GPC, where I apply a series of optimizations to improve the effective bandwidth of a 3x3x3 blur (a proxy for a huge set of operations on volumetric data)
Check ALT text for (a lot of) context.
November 30, 2024 at 6:18 PM
Very cool
I guess the TLDR here would be:
1) Properly utilize memory hierarchy (groupshared cache, more register util by computing multiple elements per thread)
2) Less thread divergence = better
3) I guess using something like sum tables to reduce blur compute amount?
I guess the TLDR here would be:
1) Properly utilize memory hierarchy (groupshared cache, more register util by computing multiple elements per thread)
2) Less thread divergence = better
3) I guess using something like sum tables to reduce blur compute amount?
Reposted by Mykhailo Moroz
Interesting set of optimizations on SVO (size and traversal) dubiousconst282.github.io/2024/10/03/v...
A guide to fast voxel ray tracing using sparse 64-trees
dubiousconst282.github.io
November 30, 2024 at 5:01 PM
Interesting set of optimizations on SVO (size and traversal) dubiousconst282.github.io/2024/10/03/v...
A week ago was trying to replicate the look of the lightbulb from Technology Connections video www.youtube.com/watch?v=dHsk...
I didn't get it very similar, but nonetheless it still is a very interesting result. compute.toys/view/1556
November 30, 2024 at 4:39 PM
A week ago was trying to replicate the look of the lightbulb from Technology Connections video www.youtube.com/watch?v=dHsk...
I didn't get it very similar, but nonetheless it still is a very interesting result. compute.toys/view/1556