yus3710-fj wrote:
Please note that there is a (non-functional) change from the previous patch:
2e6180ed4799230c5e25f85a7c633ff42f3d78ae
https://github.com/llvm/llvm-project/pull/118933
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://li
yus3710-fj wrote:
Thank you for the informantion, @tblah. I also confirmed that the regression
doesn't happen after fbd89bcc6647ed611e579d8f9c38c97b8e6f7936.
I'm a little concerned that the regression still occurs when LTO is diabled.
However, I'd like to reland this patch because this isn't t
yus3710-fj wrote:
I confirmed the performance regression on both Graviton3 and Grace when the
value of `copies` is set to 1.
The result of the measurement on Graviton3:
||max|min|avg|
|---|---|---|---|
|before this patch|52.56|52.44|52.518|
|after this patch|48.90 (-7.48%)|48.79 (-7.48%)|48.86
yus3710-fj wrote:
> @yus3710-fj after this patch we get a 7-10% slowdown for 503.bwaves_r from
> SPEC2017. I built with `-Ofast -mcpu=native -fuse-ld=lld -flto`
@tblah I'm attempting to reproduce the performance issue on my Grace machine,
but have not yet been successful. Could you please prov
yus3710-fj wrote:
Thank you for the information, @tblah. I'll revert this.
> As I understood it, the previous patch was supposed to enable your
> transformation by default, and this one was only supposed to clean up old
> flags. Did I understand wrong?
This patch also enables some transformat
https://github.com/yus3710-fj created
https://github.com/llvm/llvm-project/pull/113901
Reverts llvm/llvm-project#110063 due to the performance regression on
503.bwaves_r in SPEC2017.
>From 2f315eb090ec08e193acbc59127b958efa93fd35 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Mon, 28 Oct
https://github.com/yus3710-fj closed
https://github.com/llvm/llvm-project/pull/110063
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj edited
https://github.com/llvm/llvm-project/pull/110063
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj edited
https://github.com/llvm/llvm-project/pull/110063
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj ready_for_review
https://github.com/llvm/llvm-project/pull/110063
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj edited
https://github.com/llvm/llvm-project/pull/110063
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110063
>From 934a116155c63fdad76096faa4edd3a9c19ff5d8 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Thu, 19 Sep 2024 14:44:49 +0900
Subject: [PATCH] [flang] integrate the option
-flang-experimental-integer-ove
https://github.com/yus3710-fj closed
https://github.com/llvm/llvm-project/pull/110061
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
yus3710-fj wrote:
CI failed again, but the test mentioned above was passed this time. So, I think
I can merge this.
https://github.com/llvm/llvm-project/pull/110061
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin
yus3710-fj wrote:
CI failed due to `ClangdTests/25/79` (`DynamicIndexIncludeInsertion` in
CodeCompleteTests.cpp?). It's weird because there is no functional change on
clang (and clangd). In addition, I can't reproduce this on my machine.
I'll try re-running CI with an empty commit.
https://git
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110061
>From 6215099ca7da009e98eca138590f78a5e524a1e9 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 18 Sep 2024 21:12:43 +0900
Subject: [PATCH 1/2] [flang][Driver] Add support for -f[no-]wrapv and
-f[no]-
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110061
>From 6215099ca7da009e98eca138590f78a5e524a1e9 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 18 Sep 2024 21:12:43 +0900
Subject: [PATCH] [flang][Driver] Add support for -f[no-]wrapv and
-f[no]-stri
yus3710-fj wrote:
Thank you all for the review! I'll merge this after rebasing.
https://github.com/llvm/llvm-project/pull/110061
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
yus3710-fj wrote:
Gentle ping @MaskRay.
https://github.com/llvm/llvm-project/pull/110061
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -27,6 +27,14 @@ namespace Fortran::common {
class LangOptionsBase {
public:
+ enum SignedOverflowBehaviorTy {
+// -fno-wrapv (default behavior in Flang)
+SOB_Undefined,
yus3710-fj wrote:
This is used as the default value of a LangOption `SignedOve
@@ -6921,16 +6921,9 @@ void Clang::ConstructJob(Compilation &C, const JobAction
&JA,
Args.AddLastArg(CmdArgs, options::OPT_ftrap_function_EQ);
- // -fno-strict-overflow implies -fwrapv if it isn't disabled, but
- // -fstrict-overflow won't turn off an explicitly enabled
@@ -866,6 +866,17 @@ void Flang::ConstructJob(Compilation &C, const JobAction
&JA,
}
}
+ // -fno-strict-overflow implies -fwrapv if it isn't disabled, but
+ // -fstrict-overflow won't turn off an explicitly enabled -fwrapv.
+ if (Arg *A = Args.getLastArg(options::OPT
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110061
>From aea2cfa4b1d812dc84cb1609f93cc2ec2bcd33b4 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 18 Sep 2024 21:12:43 +0900
Subject: [PATCH 1/3] [flang][Driver] Add support for -f[no-]wrapv and
-f[no]-
@@ -0,0 +1,10 @@
+! Test for correct forwarding of integer overflow flags from the compiler
driver
+! to the frontend driver
+
+! RUN: %flang -### -fno-strict-overflow %s 2>&1 | FileCheck %s
--check-prefixes CHECK,INDUCED
yus3710-fj wrote:
As you mentioned, the
@@ -866,6 +866,17 @@ void Flang::ConstructJob(Compilation &C, const JobAction
&JA,
}
}
+ // -fno-strict-overflow implies -fwrapv if it isn't disabled, but
+ // -fstrict-overflow won't turn off an explicitly enabled -fwrapv.
+ if (Arg *A = Args.getLastArg(options::OPT
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110061
>From aea2cfa4b1d812dc84cb1609f93cc2ec2bcd33b4 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 18 Sep 2024 21:12:43 +0900
Subject: [PATCH 1/2] [flang][Driver] Add support for -f[no-]wrapv and
-f[no]-
https://github.com/yus3710-fj ready_for_review
https://github.com/llvm/llvm-project/pull/110061
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/110061
>From aea2cfa4b1d812dc84cb1609f93cc2ec2bcd33b4 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 18 Sep 2024 21:12:43 +0900
Subject: [PATCH] [flang][Driver] Add support for -f[no-]wrapv and
-f[no]-stri
https://github.com/yus3710-fj created
https://github.com/llvm/llvm-project/pull/110061
This patch introduces the options for integer overflow flags into Flang.
The behavior is similar to that of Clang.
>From d3a02bcea0cda96694b79c8b22812c29150c00b1 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
https://github.com/yus3710-fj closed
https://github.com/llvm/llvm-project/pull/91579
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/91579
>From 262824a13615b700bbfc672c3a2d06e63892c6e3 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 24 Apr 2024 14:42:21 +0900
Subject: [PATCH] [flang] Add nsw flag to do-variable increment with a new
opti
@@ -34,5 +34,9 @@ ENUM_LOWERINGOPT(NoPPCNativeVecElemOrder, unsigned, 1, 0)
/// On by default.
ENUM_LOWERINGOPT(Underscoring, unsigned, 1, 1)
+/// If true, add nsw flags to arithmetic operations for integer.
+/// Off by default.
+ENUM_LOWERINGOPT(NoSignedWrap, unsigned, 1, 0)
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/91579
>From f51cfbe1e50c7a1aa902c684f12a20d0fac39c21 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 24 Apr 2024 14:42:21 +0900
Subject: [PATCH 1/3] [flang] Add nsw flag to do-variable increment with a new
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/91579
>From f51cfbe1e50c7a1aa902c684f12a20d0fac39c21 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 24 Apr 2024 14:42:21 +0900
Subject: [PATCH 1/2] [flang] Add nsw flag to do-variable increment with a new
yus3710-fj wrote:
Windows build failed with the following message.
> fatal error C1060: compiler is out of heap space
I'll try re-running CI with an empty commit because it might have been happend
by accident.
https://github.com/llvm/llvm-project/pull/91579
__
https://github.com/yus3710-fj updated
https://github.com/llvm/llvm-project/pull/91579
>From f51cfbe1e50c7a1aa902c684f12a20d0fac39c21 Mon Sep 17 00:00:00 2001
From: Yusuke MINATO
Date: Wed, 24 Apr 2024 14:42:21 +0900
Subject: [PATCH] [flang] Add nsw flag to do-variable increment with a new
opti
https://github.com/yus3710-fj created
https://github.com/llvm/llvm-project/pull/91579
This patch adds nsw flag to the increment of do-variables when a new option is
enabled.
NOTE 11.10 in the Fortran 2018 standard says they never overflow.
See also the discussion in #74709 and the following di
37 matches
Mail list logo
