Re:Bend: a high-level language that runs on GPUs (via HVM2)
by carlhjerpe on Saturday May 18, 2024 @11:33PM
(#40404044)
If you wanna be really honest you write one short example where HVM is weak and one where it's strong, use the opportunity to explain why if you want
Re:Bend: a high-level language that runs on GPUs (via HVM2)
by LightMachine on Monday May 20, 2024 @06:14PM
(#40421088)
That is actually an amazing idea. I'll adopt it.
Just a note: we are NOT 10x slower than Python. I think a lot of people got the wrong message from this thread. HVM is actually quite fast already. It is just that, on this specific program, Python was doing no allocations, while HVM was allocating a lot.
If you compare programs that do the same allocation, HVM already outperforms not just Python but even compiled languages like Haskell/GHC, due to using all cores. See the Bitonic Sort example. And that's extremely relevant, because real world programs in high-level languages allocate a lot!
I think I made a huge mistake of using a "sum" example on the GUIDE, since it is actually one of the few specific programs where HVM will do poorly today.