Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-14 Thread Maxim Kuvyrkov
On 14/08/2012, at 7:08 PM, Richard Sandiford wrote: > OK with those changes, thanks. Checked in with the noted changes and a fixed bug. It turns out that mips_linked_madd_p is also called via mips_macc_chains_reorder, which may pass a (use ...) instruction, which causes get_attr_* to blow up.

Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-14 Thread Richard Sandiford
Maxim Kuvyrkov writes: > I thought I'll butt in since I did a very similar thing for > sync_memmodel a couple of months ago. Thanks. > + /* We take care in instruction definitions to make sure accum_in operand is > + a register_operand or [a more restrictive] muldiv_target_operand. */ > +

Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-13 Thread Maxim Kuvyrkov
On 9/08/2012, at 7:10 AM, Richard Sandiford wrote: > Hmm, yeah, it does look like they should be using mips_linked_madd_p > instead, except that mips_linked_madd_p isn't yet wired up to handle > DSP macs. Rather than pattern-match them all, the easiest thing would > probably be to define a new at

Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-08 Thread Richard Sandiford
Sandra Loosemore writes: >> +;; Before reload, all multiplier is registered as imul3 (which has a long >> +;; latency). We temporary jig the latency such that the macc groups >> +;; are scheduled closely together during the first scheduler pass. >> +(define_bypass 1 "r74k_int_mul3" "r74k_dsp_ma

Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-04 Thread Sandra Loosemore
On 08/04/2012 07:48 AM, Richard Sandiford wrote: Sandra Loosemore writes: The existing scheduler bypass information for madd on the 74k uses some bits copied from the 24k, and is not quite correct. This patch is based on one originally sent to us by MIPS and has been present in our local sourc

Re: [PATCH, MIPS] 74k madd scheduler tweaks

2012-08-04 Thread Richard Sandiford
Sandra Loosemore writes: > The existing scheduler bypass information for madd on the 74k uses some > bits copied from the 24k, and is not quite correct. This patch is based > on one originally sent to us by MIPS and has been present in our local > source base for years. I've confirmed that we

[PATCH, MIPS] 74k madd scheduler tweaks

2012-08-01 Thread Sandra Loosemore
The existing scheduler bypass information for madd on the 74k uses some bits copied from the 24k, and is not quite correct. This patch is based on one originally sent to us by MIPS and has been present in our local source base for years. I've confirmed that we are legally allowed to contribut