Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread juzhe.zh...@rivai.ai
and documents (No complicate vectorizer support). Really appreciate your help! juzhe.zh...@rivai.ai From: Richard Biener Date: 2023-06-27 16:56 To: juzhe.zh...@rivai.ai CC: gcc-patches; richard.sandiford; pan2.li Subject: Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE On Tue

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread Richard Biener via Gcc-patches
write a testcase that exercises it (and verifies correctness). Richard. > Thanks. > > > juzhe.zh...@rivai.ai > > From: Richard Biener > Date: 2023-06-27 16:34 > To: juzhe.zh...@rivai.ai > CC: gcc-patches; richard.sandiford; pan2.li > Subject: Re: Re: [PATCH V4]

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread juzhe.zh...@rivai.ai
lding of dependent stmts it might be nice to have the constant folded masks. I will open a bugreport with this testcase. Richard. > > juzhe.zh...@rivai.ai > > From: Richard Biener > Date: 2023-06-27 15:47 > To: juzhe.zh...@rivai.ai > CC: gcc-patches; richard.

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread Richard Biener via Gcc-patches
nice to have the constant folded masks. I will open a bugreport with this testcase. Richard. > > juzhe.zh...@rivai.ai > > From: Richard Biener > Date: 2023-06-27 15:47 > To: juzhe.zh...@rivai.ai > CC: gcc-patches; richard.sandiford; pan2.li > Subject: Re: Re: [PATCH

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread Richard Biener via Gcc-patches
ar element load CSE from VLA [masked/with-len] stores if we can for example compute a lower bound on the number of elements accessed. Richard. > Thanks. > > > juzhe.zh...@rivai.ai > > From: Richard Biener > Date: 2023-06-27 15:47 > To: juzhe.zh...@rivai.ai > CC: gc

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread juzhe.zh...@rivai.ai
s. > > > juzhe.zh...@rivai.ai > > From: Richard Biener > Date: 2023-06-27 15:33 > To: Ju-Zhe Zhong > CC: gcc-patches; richard.sandiford; pan2.li > Subject: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE > On Tue, 27 Jun 2023, juzhe.zh...@rivai.ai wrot

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread juzhe.zh...@rivai.ai
des, RVV currently can't have VLS modes (V4SI) since LTO issue. I am trying to find a CSE opportunity on VL vectors. Thanks. juzhe.zh...@rivai.ai From: Richard Biener Date: 2023-06-27 15:47 To: juzhe.zh...@rivai.ai CC: gcc-patches; richard.sandiford; pan2.li Subject: Re: Re: [PATCH V4] SCCVN:

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread Richard Biener via Gcc-patches
From: Richard Biener > Date: 2023-06-27 15:33 > To: Ju-Zhe Zhong > CC: gcc-patches; richard.sandiford; pan2.li > Subject: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE > On Tue, 27 Jun 2023, juzhe.zh...@rivai.ai wrote: > > > From: Ju-Zhe Zhong > > > >

Re: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread juzhe.zh...@rivai.ai
ches; richard.sandiford; pan2.li Subject: Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE On Tue, 27 Jun 2023, juzhe.zh...@rivai.ai wrote: > From: Ju-Zhe Zhong > > Hi, Richi. > > I tried to understand your last email and to refactor the do-while loop using > VECTOR_

Re: [PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-27 Thread Richard Biener via Gcc-patches
On Tue, 27 Jun 2023, juzhe.zh...@rivai.ai wrote: > From: Ju-Zhe Zhong > > Hi, Richi. > > I tried to understand your last email and to refactor the do-while loop using > VECTOR_CST_NELTS. > > This patch works fine for LEN_MASK_STORE and compiler can CSE redundant store. > I have appended testc

[PATCH V4] SCCVN: Add LEN_MASK_STORE and fix LEN_STORE

2023-06-26 Thread juzhe . zhong
From: Ju-Zhe Zhong Hi, Richi. I tried to understand your last email and to refactor the do-while loop using VECTOR_CST_NELTS. This patch works fine for LEN_MASK_STORE and compiler can CSE redundant store. I have appended testcase in this patch to test VN for LEN_MASK_STORE. I am not sure whet