On Mon, Aug 02, 2021 at 08:31:43AM -0500, Bill Schmidt wrote: > Interestingly, when the quadword compares are expanded at GIMPLE time, > we generate worse code involving individual 64-bit compares. For the > time being, I will not expand these at GIMPLE time; independently, this > bears looking at to see why expressions like (uint128_1 < uint128_2) > will generate poor code.
Details like this should probably not be exposed before RTL anyway? Everything else is at a more abstracted level as well before expand? It will be interesting to see what causes the worse code though :-) Segher