On 02/16/2017 11:48 AM, Bin Cheng wrote: > BTW, it may also help PR78116? Hi Pat, could you please help verify this? > Thanks!
The first testcase in pr78116 no longer contains loads from spill in the loop even before your patch. When built with your patch, there are four additional register copies in the loop (vmovaps %zmm2, %zmm14). As for the second testcase, your patch gets rid of the 12 loads from spill in the loop. -Pat