Michael Mara @mikemx7f

Computer Graphics Researcher (Just finished PhD Stanford, previously Oculus Research, NVIDIA Research and Casual Effects) illumination-codified.com La Jolla, CA Joined September 2011

Tweets

458
Followers

894
Following

395
Likes

7K

Michael Mara @mikemx7f

4 years ago

My friend's home burned down in the Boulder county fires on Thursday, and he lost basically all of his possessions. If you can spare some change, please give him some to help get back on his feet. gofund.me/92ca0c56

0 1 1 0 0

Friendly Capybara 🇵🇸 🇺🇦 @capybaranorth

5 years ago

I'm working with @msbernst and epidemiologists at @UVA_BI to enable volunteers to fill a gap in scientists’ ability to model COVID-19. Help predict the spread of COVID-19 by adding data on what social distancing rules are in place in your own county at socialdistancing.stanford.edu!

5 30 57 0 3

Patrick Walton @pcwalton

7 years ago

BTW, is there a preferred name for the “integer divide rounding up” function?

4 1 4 0 0

morgan @CasualEffects

7 years ago

Blast from the past. My first collaboration with NVIDIA, ten years ago. Real-time global illumination using image space photon mapping. Fast forward to today and you can run this with 1000x more complicated scenes on RTX. youtube.com/watch?v=OKldiJ…

4 13 83 0 2

Alberto Jaspe @albertojaspe

7 years ago

Realtime #raytracing a 1M^3 voxelated Power Plant, compressed and encoded with an Symmetry-aware DAG. 1.57T non-empty #voxels, into 3.2GBytes, so 0.017 bits/voxel. #GLSL renderer with shadow rays, screen-space normals and AO. youtube.com/watch?v=UG6cTS…

3 16 57 0 7

Mauricio @pixnblox

7 years ago

Well, this is unexpected. The book "Physically Based Rendering" is now available for free online: pbr-book.org. And, I can't overemphasize this, it has LINKS within the book, so browsing the inline code is now trivial (credit to @marcosalvi for the notice)

1 280 620 0 42

Download Image

morgan @CasualEffects

7 years ago

The bibliography (references) is excluded from the page count limitation this year at @I3DCONF. You no longer need to compress those bibtex entries or trim the number of citations to make room for more result figures.

0 3 6 0 0

morgan @CasualEffects

7 years ago

A thread of notes on the ACM ToG Production Rendering special edition from @Peter_shirley & me. dl.acm.org/citation.cfm?i…

1 39 111 0 25

Download Image

Jason Peng @xbpeng4

7 years ago

Video imitation paper accepted into @SIGGRAPHAsia! Blog: bair.berkeley.edu/blog/2018/10/0… Project page: xbpeng.github.io/projects/SFV/i… Big thanks to all my collaborators for making this work possible, @akanazawa, Jitendra Malik, @pabbeel, @svlevine

16 295 1K 0 50

Download Video

StanfordGWU-UE @StanfordGWU

7 years ago

Stanford Solidarity Network stands with Dr. Christine Blasey Ford and all survivors of gender-based violence. Stanford students/affiliates -- express your support for Dr. Ford, who is a member of our campus community, through this letter of solidarity! bit.ly/2xZ9u6F

0 4 10 0 0

morgan @CasualEffects

7 years ago

The @mikemx7f 2013 photon mapping paper lighting algorithm was designed in anticipation of Turing (that's why he assumed that photon tracing is fast)...you can now run the whole photon mapping pipeline for 60 fps graphics! casual-effects.com/research/Mara2…

2 28 77 0 5

Download Image

Michael Mara @mikemx7f

7 years ago

Co-presenting optlang.org, our nonlinear least squares optimization DSL with Zach DeVito, at the first technical papers session of #SIGGRAPH, starting now in room 109-110 West.

0 3 13 0 0

morgan @CasualEffects

7 years ago

Looking forward to @mikemx7f's presentation of Opt tomorrow in the 10:45am ToG session. optlang.org s2018.siggraph.org/presentation/?…

0 4 8 0 0

Download Image

morgan @CasualEffects

7 years ago

Since a few people asked afterward: the "screen space ray tracing" I referred to in my HPG AO presentation yesterday isn't "SSAO", it is this kind of technique jcgt.org/published/0003…, which gives a correct result wherever the depth buffer accurately represents the scene content.

1 7 33 0 6

Michael Mara @mikemx7f

7 years ago

Sad that (like me) you are missing this years @HPG_Conf presentations? The conference website is being updated with talk slides, almost all of yesterday’s talks are up for free: highperformancegraphics.org/2018/program/

1 45 97 0 4

morgan @CasualEffects

7 years ago

A primary challenge in modern graphics is that the power of stochastic methods is in gathering maximum unique and uncorrelated information per sample, while the performance of parallel processing (GPUs) is from exploiting redundancy in both computation and data.

6 18 92 0 4

Kostas Anagnostou @KostasAAA

7 years ago

This is a prime example of why one should do algorithmic improvements first instead of micro-optimisations: Shadow pass started at 16ms, using a custom buffer layout to reduce mem bandwidth took it to 13ms, using SAH to reduce traversal steps took it down to 3.6ms (GTX 970).