Re: [PATCH][i386] Properly scale vec_construct cost

2015-04-21 Thread Uros Bizjak
On Tue, Apr 21, 2015 at 1:30 PM, Richard Biener wrote: > Well, anyway - below for the "obvoious" cost model patch. > > Boostrapped on x86_64-unknown-linux-gnu, testing in progress. > > Ok for trunk? > > Thanks, > Richard. > > 2015-04-21 Richard Biener > > * config/i386/i386.c (ix86_bui

[PATCH][i386] Properly scale vec_construct cost

2015-04-21 Thread Richard Biener
Hi, currently vec_construct cost is simply TYPE_VECTOR_SUBPARTS / 2 + 1, a reasonable estimate only of other target stmt costs are close to 1. The idea was you need that many vector stmts thus the following patch which should fix skewed costs for bdver2 for example with a vec_stmt_cost of 6. Fix