Ben Vanik
Ben Vanik
Oh the other thing we need to investigate is the best approach to multithreading in the various engines. Ideally we'd be able to load a module and then call in...
I realized I never sketched it out, but here's what we want to do:  * dynamically load one or more wasm modules containing small kernel functions * load/unload of...
Hitting this too. Haven't yet found a workaround.
Cool, this is probably the first complex usage down to this layer. It doesn't look like there's a transform that does a complex tensor -> flattened 2xf32 tensor yet, just...
Interesting - want to take a look or should I? (I thought jakub fixed a lot of that, so maybe something got missed or it's a different issue, would be...
(I don't have anything running yet but can try to make a test for this - thanks for implementing it :)
I haven't tested it yet but one quirk of `hsa_amd_signal_wait_any` is that it does not allow 0/NULL signal handles. This is inconsistent with the AQL `hsa_barrier_or_packet_t` that allows any signal...
Cool! Filed at #252!
maybe you have it the other way around? the error you have says `C:\V\iree\experimental\rocm\rocm_buffer.c` which is ROCM, not HIP that kind of error will happen if the driver is casting...
if you're in a release LTO build it's possible the two functions are identical and got folded, but usually asserts and stuff prevent that - either way, good to test...