Assuming a fully pipelined vector unit (and from experience on
AArch64), an u-arch's scalar-to-vector move cost is likely to play a
significant role in whether this will be profitable or not.
--Philipp.
On Wed, 31 May 2023 at 00:10, Jeff Law via Gcc-patches
wrote:
>
>
>
> On 5/30/23 16:01, 钟居哲
On 5/30/23 16:13, 钟居哲 wrote:
Ok. I prefer just keep scalar load + vmv.v.x by default since I believe
most machines prefer this way.
Seems quite reasonable to me.
jeff
Ok. I prefer just keep scalar load + vmv.v.x by default since I believe most
machines
prefer this way.
juzhe.zh...@rivai.ai
From: Jeff Law
Date: 2023-05-31 06:09
To: 钟居哲; andrew; rdapp.gcc
CC: gcc-patches; kito.cheng; palmer
Subject: Re: [PATCH] RISC-V: Synthesize power-of-two constants
On 5/30/23 16:01, 钟居哲 wrote:
I agree with Andrew.
And I don't think this patch is appropriate for following reasons:
1. This patch increases vector workload in machine since
it convert scalar load + vmv.v.x into vmv.v.i + vsll.vi.
This is probably uarch dependent. I can probably
I disagree with this patch.
Thanks.
juzhe.zh...@rivai.ai
From: Andrew Waterman
Date: 2023-05-31 04:18
To: Robin Dapp
CC: gcc-patches; Kito Cheng; palmer; juzhe.zh...@rivai.ai; jeffreyalaw
Subject: Re: [PATCH] RISC-V: Synthesize power-of-two constants.
This turns out to be a de-optimization for imple
This turns out to be a de-optimization for implementations with any
amount of temporal execution (which is most machines with LMUL > 1 and
even some machines with LMUL <= 1). Scalar instructions are generally
cheaper than multi-cycle-occupancy vector operations, so reducing
scalar work by
