Making your ASM intrinsics typed is a terrible, terrible, terrible, terrible, terrible, terrible, terrible, terrible, terrible, idea.
@cmuratori @jcalsbeek @won3d (ie., I do not have to issue an instruction to convert 16xu8 to 8xs16 before doing a 16-bit signed add)
-
-
@cmuratori@won3d Yeah, you're right. Even the differently performant load/store ops should probably have the same signature. -
@jcalsbeek@won3d Mmm, on SSE that's not possible because there are aligned and unaligned ones, so you have to say which one you're doing. - Show replies
New conversation -
-
-
@cmuratori@jcalsbeek the issue is between precomputed float stuff. So int/float, but not due to different pipelines.Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
@cmuratori@jcalsbeek yeah, your example is poppy. - Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.