1,710 karma · joined February 13, 2023
(Ex-CG/VFX renderer dev)
Hyphens should be used to link compound/suffix words, not em dashes.
e.g. I see a general lack of reserve()s on strings and vectors (a few tests do use them a bit, like `graph` and `particles`), and there are several other things which I wouldn't personally call "optimized", so I assume those ones are the 'idiomatic' ones?
Same goes for the Rust code: a lot of the code seems to be using Vec::new() and String::new(), instead of ::with_capacity() which would optimize the memory re-allocation (reducing it as the containers/strings grow).
If I reduced the res a bit, then it worked, so there did seem to be a limit there somewhere. height and width are stored as u32s internally though, so yes, the format itself technically supports it...
I never looked into it more, so might have been a libavif issue, or maybe the way I was using it (maybe it has defaults or profiles?)
The only two lights (they're not 'houses') on Stewart Island are actually on this map.
The Snares don't have any.
Looks like a bug with the view origin point being mis-matched with what the view's actually focused on - maybe high-DPI scale mis-match?
Although this one has done a reasonable job at culling most of those out, at least on the south coast of England.
Some of that was due to just human time resource constraints (and assets being modeled for director validation at full screen turn-table resolution, but then just used at 1/8 size in the film), but other times it was just people losing context as to what was "needed" or would even make any difference (unfortunately, often without considering the "cost" - in terms of memory usage and CPU-time for rendering / compositing) - vs look difference.
Then there are questions from the artists like "Why's this rendering so slow, it's 'only' a load of skyscrapers in a city with the camera 0.5 km away?"
You then have a look at what's actually going on, and all the rooms in the skyscrapers have modeled (and lookdev-ed) furniture and things like lamp shades around lights. Which A: is expensive memory-wise and just silly, but also makes light-transport to sample all the lights and light the scene (even with clever NEE light-sampling techniques) stupidly expensive.
Then again, that same building, this time being rendered from about 300m away was very slow to render, and we looked at the BSDF material stack, and there was a homogeneous volumetric scattering layer on the glass windows. Like - WTF - that's insane for a uni-directional path-tracer (and physically wrong). "What's this for, this is why it's so slow.?"
"Oh, it's to represent the algae/dirt on the glass".
"I'm sorry!?"
The website mentions:
> FLUX 3 Image is available under a commercial weights license for companies running image generation at scale. Fine-tune and deploy it on your own infrastructure. Reach out to us to learn more.
I guess the open ones would be non-commercial?
Because a lot of work was done to prepare and fix potential issues.
"Carter Leffer only directed GPT–6 Astra to see if it could break any of the unbroken Enigma messages published on the Crypto Cellar Research web page."
Seriously? Categories of types maybe, but literal types it's more than that.
Even closures allocate if they need to capture their environment.
Sometimes using a global one isn't the best thing (it's often a good idea to specialise base on allocation size, reuse and lifetimes), but I've used them quite a bit in C++ over the past 16 years doing HPC for graphics, rendering and simulation, so calling them only useful for "toy problems" likely shows you just haven't found a need for them in what you've been doing.
What does that mean?
The only thing previously stopping them holding the strait "hostage" was the threat from the US of doing an "Operation Praying Mantis v2" and destroying their navy and attacking them.
And when the US did that anyway, they had nothing to lose.
> The data model is defined using a custom format .rmodel based on the ddkit > framework (rmodel-api).
> Please note that the .rmodel data model is for documentation purposes only and > does not reflect the actual interface used on the communication layer. Network > communication is only possible through the provided libraries.
What other protocols and systems were you thinking of that can provide (depending on broker) hard-realtime delivery that can be used in safety-critical systems?
Weird as well, it's clearly pulled out some 1884 patent on clock designs, and a quick ddg/google doesn't show it as anything to do with grandfather clocks.
Mathematics is being used as a benchmark because there are some high-profile awards in this area I guess, and possibly because 2/3 years ago LLMs were pretty atrocious at it so the level of improvement has been significant.