Brad Roepstorff

3K posts

Brad Roepstorff banner
Brad Roepstorff

Brad Roepstorff

@broepstorff

Graphics programmer. Previously: Amazon, Avalanche Studios, 38 Studios. Square-Enix, Pandemic. Opinions do not represent my employer

Salt Lake City, UT Katılım Şubat 2010
794 Takip Edilen1.2K Takipçiler
Brad Roepstorff retweetledi
Ryan Fleury
Ryan Fleury@rfleury·
New post: "Multi-Core By Default" On multi-core programming, not as a special-case technique, but as a new dimension in all code.
Ryan Fleury tweet media
English
37
71
610
72.2K
Brad Roepstorff
Brad Roepstorff@broepstorff·
@TerriblePosture Ok that trailer kinda rules. Also: my 1.5 year old watching over my shoulder just demanded to watch “bata choo choo” a bunch of times in a row
English
1
0
0
21
Brad Roepstorff retweetledi
Jarkko Lempiäinen
Jarkko Lempiäinen@JarkkoPFC·
Local lights are now live in HypeHype! 🎉 I developed a new stochastic lighting algorithm — from concept to prototype to production — in just 6 months. Super proud of this one! [Details in thread 🧵] Here's a walkthrough of a game I relit with local lights
English
12
59
410
33K
Brad Roepstorff retweetledi
rob cupisz
rob cupisz@robcupisz·
For the love of our eyeballs, do not align your ember particles to velocity! Don’t give them that comet texture either, ugh. Let’s make ✨physically based flying embers✨ A Unity vfx graph implementation and explanation below. 🧵
English
10
82
1K
68.4K
Brad Roepstorff retweetledi
NOTimothyLottes
NOTimothyLottes@NOTimothyLottes·
[0] What a horrible term 'printf-debugging' If you are debugger-free might as well at least be libc-free Here is how I work CPU C code without any debugger in my engines, and how it works better than a debugger for me ...
English
4
9
132
19.3K
Brad Roepstorff retweetledi
Walaber
Walaber@walaber·
yep, definitely giving this a try.
つーき/つーき制作所@divideby_zero

#Highking をついに本日11月13日(=良い登山)に Steam にてリリースしました! ローンチセールで30%OFFの490円です。 Steamは初でつまづく事も沢山ありましたが、なんとかここまでこぎ着けました。 今後アップデートも予定してますので何卒応援よろしくお願いいたします! store.steampowered.com/app/2883890/Hi…

English
0
1
20
2.2K
Brad Roepstorff
Brad Roepstorff@broepstorff·
@yiningkarlli yeah, I was scratching my head at that one too. I guess maybe it counts because Feynman did some AI research back in the day? hah
English
0
0
1
93
Yining Karl Li
Yining Karl Li@yiningkarlli·
The Nobel Prize for Physics this year goes to.... not physics???
English
3
0
7
1.2K
Brad Roepstorff retweetledi
Jade Law
Jade Law@jadel4w·
The reason Japanaese studios are making consistently great games as an industry hub is quite simple. Many of their teams have been making games together for a long time, sometimes decades. They build upon their pipelines, they know the strengths and weaknesses of their teams. Theres a healthy mix of new staff with new ideas and seasoned devs with retained knowledge. Even when new companies form, they often emerge from teams who had worked together prior. The average employee stays in a job for 12 years in Japan. Up to 3x longer than western hubs. The west needs to do more to retain staff, stop treating them as disposable. You cant simply throw money at building a team for every game, there's a cadence that comes with familiarity and group knowledge. Not saying Japan's perfect but there's definitely positive development outcomes from this culture. I find the number one reason for job hopping is lack of clear role progression and lack of support in improving. Not money. Growth. Afterall money comes with growth. The problem is too many studios use money as a way to bribe staff into staying in poor situations rather than solving root issues. I've fallen for this myself. Such a simple thing, and yet when I think back throughout my career it was never clear on how to attain my next promotion, and was I getting any training? nope. We need to stop seeing "investing" in staff as throwing money at poaching staff, and instead see it as putting money into growing and retaining the staff we have.
English
32
142
927
46.5K
Brad Roepstorff
Brad Roepstorff@broepstorff·
@rfleury Yeah, agreed...except for artists type users, who mostly seem to be allergic to typing code into open text files. I think it's because the possibility space of text is too large and compilers are necessarily pedantic, so only certain engineer types can put up with it daily
English
0
0
0
45
Ryan Fleury
Ryan Fleury@rfleury·
@broepstorff You can get this more or less with a good graphical debugger. I think current paradigms would improve a lot, even holding the source code format constant, with just a better build model, and much faster builds, to allow for real time feedback, hot reloading, and looped editing.
English
1
0
3
160
Ryan Fleury
Ryan Fleury@rfleury·
As one who suspects there’s a high ceiling beyond text as the visual representation for code, I think text has a design advantage in its ability to implicitly/invisibly denote references (via identifiers) between two locations. I’d say that future work should/could include that.
Тsфdiиg@tsoding

People do not realize the level of complexity software developers have to deal with. We have pushed the limits of the textual languages to contain this complexity long time ago, but any other representation is just somehow worse at that.

English
2
1
58
7.1K
Brad Roepstorff
Brad Roepstorff@broepstorff·
@rfleury I mostly still just want text but with the two main advantages of a noodle graph: improved look up of functions/variables and zero syntax errors Inline display of variables during execution is also a nice advantage of noodle graphs, but is pretty domain specific
English
1
0
0
178
Brad Roepstorff
Brad Roepstorff@broepstorff·
@rfleury Yup. Blows me away that most noodle graph languages don't have named variables Text also has a (vertical) timeline of statement execution which is a big plus (but which falls apart in async/parallel code)
English
1
0
2
330
Brad Roepstorff retweetledi
Forrest Smith
Forrest Smith@ForrestTheWoods·
I made a VSCode extension that uses PDBs to provide GotoDefinition intellisense support. Verified to work with Jai, C++, Rust, Zig, Odin, and Nim. Should work with *any* compiled language that makes PDBs. Here's a little video. Links to blog post / marketplace in reply.
English
6
30
214
16.2K
Brad Roepstorff retweetledi
Keenan Crane
Keenan Crane@keenanisalive·
Signed distance functions (SDFs) are an important surface representation, which can be directly visualized via the “sphere tracing” algorithm. At #SIGGRAPH2024 we showed how to sphere trace a whole new class of surfaces, based on *harmonic functions* rather than SDFs. [1/n]
Keenan Crane tweet media
English
4
149
729
73.7K
Brad Roepstorff retweetledi
Lucky Iyinbor
Lucky Iyinbor@Luckyballa·
A very interesting technique for mesh simplification and skinning weight optimization. While it may seem trivial, extracting a single layer from a multi-layered object while avoiding collisions is far from easy. kuiwuchn.github.io/proxycloth.html
English
2
34
161
9.5K
Brad Roepstorff
Brad Roepstorff@broepstorff·
@Donzanoid Heh, the game was targeting xbox 360 and wii (but not PS3 for some reason), and that was the only viable choice at the time.
English
0
0
0
76
Brad Roepstorff
Brad Roepstorff@broepstorff·
@Donzanoid You haven't lived unless you programmed tev stages in Gambryo using xml lol
English
0
0
0
125