Been watching your development for a while on Twitter. This is a monumental achievement and I hope it gets the recognition it deserves.
Bend: a high-level language that runs on GPUs (via HVM2)
41–50 of 269 posts
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#42Looks cool but what's one toy problem that it can solve more efficiently than others?
This is unproven (and not a toy problem), but I imagine it's going to do pretty well at compilers. The amount of time I'm waiting at work, hypnotizing the tsc process that sits at 100% CPU, wishing it was parallel...
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#43Looking forward to using this. Curious about how far away WebGPU/WASM support might be, it could provide a single cross-platform backend.
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#44Wow this is very impressive!
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#45Would a compiler be faster by using HVM? Would love to see a fully parallel version of typescript tsc
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#46What's going on with the super-linear speedup going from one thread to all 16?
210 seconds (3.5 minutes) to 10.5 seconds is a 20x speedup, which isn't really expected.
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#47Honestly incredible, and congrats on the release after what looks like an insane amount of work.
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#48What's going on with the super-linear speedup going from one thread to all 16? 210 seconds (3.5 minutes) to 10.5 seconds is a 20x speedup, which isn't really expected.
Its possible to see such scaling if involving any level of cache or I/O.
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#49As a resident of Bend, Oregon... it was kind of funny to read this and I'm curious about the origin of the name.
Re: Bend: a high-level language that runs on GPUs (via HVM2)
#50What's going on with the super-linear speedup going from one thread to all 16? 210 seconds (3.5 minutes) to 10.5 seconds is a 20x speedup, which isn't really expected.
the single-thread case ran a little slower than it should on this live demo due to a mistake on my part: `run` redirected to the Rust interpreter, rather than the C interpreter. the Rust one is a little bit slower. the numbers on the site and on all docs are correct though, and the actual speedup is ~12x, not ~16x.