On 03.04.2018 19:02, Alexander Monakov wrote: > On Tue, 3 Apr 2018, Andrey Belevantsev wrote: > >> Hello, >> >> This issue ended up being fixed the way different from described in the PR. >> We do not want to walk away from the invariant "zero SCHED_TIMES -- insn >> is not scheduled" even for bookkeeping copies (testing showed it trips over >> asserts designed to catch this). Rather we choose merging exprs in the way >> the larger sched-times wins. > > ... but the Changelog and the actual patch take the average rather than the > maximum sched-time? :) I believe either way would be acceptable, but please > clarify the intent.
Sorry, the average is the intent. Just to have a bit more of pipelining chances. Andrey > > Alexander >
