Ben Vanik
Ben Vanik
Closing stale PR.
There's a difference between a new ukernel (very interesting) and a new arch variant for an existing in-tree ukernel - I thought that this issue would be about the former...
(if you do want to continue down the path of splicing in an arch variant like this, the best way for you to do so is maintain a fork with...
I forget how things fit together and what our default flag state is but we're currently real bad on several metrics (peak memory consumption and utilization) with data tiling/mmt4d -...
Most of the func.func->util.func changes landed as part of #16411. There's still work to do to make use of the new util.func tied operand feature and enable IPO to work...
Good to see we're at the point where the way we lower to pack/unpack today before dispatch region formation is unable to satisfy the requirements. We've known about it forever...
Do we know how terrible of code this generates? This is a lot of stuff to do per element and I suspect it may break vectorization and such - we'd...
sounds good - @MaheshRavishankar may be able to speak to the formation questions - I mostly am just curious if we're sending all scatters down a scalar path with multiple...
yeah, that reduction is likely to be an issue - we don't really want to serialize everything like that - is there a reason to reduce? (I'd expect this to...
@rsuderman still valid?