Re: [PATCH][i386] Update Zen tuning for vector load cost

On Mon, 8 Oct 2018, Jan Hubicka wrote: > > > > This adjusts Zen AVX256 vector load cost to be twice as expensive > > than AVX128 vector load cost (twice via ix86_vec_cost keying on > > TARGET_AVX128_OPTIMAL - should rather use some > > TARGET_VECTOR_IMPL_WIDTH or so). > > > > Likely the current

Re: [PATCH][i386] Update Zen tuning for vector load cost

> > This adjusts Zen AVX256 vector load cost to be twice as expensive > than AVX128 vector load cost (twice via ix86_vec_cost keying on > TARGET_AVX128_OPTIMAL - should rather use some > TARGET_VECTOR_IMPL_WIDTH or so). > > Likely the current cost value was meant to make AVX256 loads _cheaper_ >

[PATCH][i386] Update Zen tuning for vector load cost

This adjusts Zen AVX256 vector load cost to be twice as expensive than AVX128 vector load cost (twice via ix86_vec_cost keying on TARGET_AVX128_OPTIMAL - should rather use some TARGET_VECTOR_IMPL_WIDTH or so). Likely the current cost value was meant to make AVX256 loads _cheaper_ than two AVX128