https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/128506
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/128506
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/137811
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/137183
>From 253bbc8cf673dd1d0ca606058dbd26e15cee651a Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Thu, 24 Apr 2025 11:24:30 +0100
Subject: [PATCH 1/2] [libclc] Support the generic address space
This commit
Author: Fraser Cormack
Date: 2025-04-29T15:48:46+01:00
New Revision: 6ffccea1c2b804d530c0f7bf3cc92db42b4432c5
URL:
https://github.com/llvm/llvm-project/commit/6ffccea1c2b804d530c0f7bf3cc92db42b4432c5
DIFF:
https://github.com/llvm/llvm-project/commit/6ffccea1c2b804d530c0f7bf3cc92db42b4432c5.diff
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/137824
By having these already defined as type 'gentype' we can avoid unnecessary
casting.
>From 662f8881d3da88714231b9a232a6006ae834b3b7 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 29 Apr 2025 16:05
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/137824
>From abb52892c044ca1883cc41d97c3edf08aafc9746 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 29 Apr 2025 16:05:19 +0100
Subject: [PATCH] [libclc] Avoid casting NANs & literals to 'gentype'
By hav
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/137824
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127532
libclc uses llvm-link to link together all of the individually built libclc
builtins files into one module. Some of these builtins files are compiled from
source by clang whilst others are converted from LL
@@ -1,8 +1,4 @@
-#if __clang_major__ >= 7
-target datalayout =
"e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5"
-#else
-target datalayout =
"e-p:32:32-p1:64
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/127515
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127515
In #127378 it was reported that builds without clspv targets enabled were
failing after #124727, as all targets had a dependency on a file that only
clspv targets generated.
A quick fix was merged in #1273
@@ -13,19 +13,25 @@
const UINT_TYPE sign_bit = (UINT_TYPE)1
\
<< (sizeof(INT_TYPE_SCALAR) * 8 - 1);
\
const UINT_TYPE sign_bit_mask = sign_bit - (UINT_TYPE)1;
\
-INT_TYP
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/127469
>From dbe1d8d133a54d0353fdb466cd59f5931d670bbf Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Mon, 17 Feb 2025 10:17:44 +
Subject: [PATCH 1/2] [libclc] Improve nextafter behaviour around zero
This
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/127469
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127783
During a recent change, the build system accidentally dropped the (theoretical)
support for the CLC builtins library to build target-specific builtins from the
'amdgpu' directory, due to a change in variabl
@@ -12411,16 +12411,13 @@ def err_builtin_is_within_lifetime_invalid_arg :
Error<
"%select{non-|function }0pointer argument to '__builtin_is_within_lifetime' "
"is not allowed">;
-def err_builtin_invalid_arg_type: Error <
- "%ordinal0 argument must be "
- "%select{a vec
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127628
This commit is a broad update across libclc to use the CLC conversion builtins
in CLC functions, even those with a '__clc' prefix in the generic folder. This
better prepares them for an official move to the
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/127532
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -1,8 +1,4 @@
-#if __clang_major__ >= 7
-target datalayout =
"e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5"
-#else
-target datalayout =
"e-p:32:32-p1:64
frasercrmck wrote:
Just one failure:
```
Timed Out Tests (1):
MLIR :: Examples/standalone/test.toy
Total Discovered Tests: 131804
Skipped : 48 (0.04%)
Unsupported : 2855 (2.17%)
Passed : 128580 (97.55%)
Expectedly Failed:320 (0.24
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/127532
>From 2004d2a93073273cd7e09fbf6da6de7972412e98 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Mon, 17 Feb 2025 18:04:08 +
Subject: [PATCH] [libclc] Suppress data-layout warnings during linking
libc
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/127532
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/127532
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/127628
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Author: Fraser Cormack
Date: 2025-02-18T14:56:25Z
New Revision: 378c6fbe330e6de2feac8a8e3bc0c93ec6aca80d
URL:
https://github.com/llvm/llvm-project/commit/378c6fbe330e6de2feac8a8e3bc0c93ec6aca80d
DIFF:
https://github.com/llvm/llvm-project/commit/378c6fbe330e6de2feac8a8e3bc0c93ec6aca80d.diff
LOG
@@ -12411,16 +12411,13 @@ def err_builtin_is_within_lifetime_invalid_arg :
Error<
"%select{non-|function }0pointer argument to '__builtin_is_within_lifetime' "
"is not allowed">;
-def err_builtin_invalid_arg_type: Error <
- "%ordinal0 argument must be "
- "%select{a vec
frasercrmck wrote:
ping, thanks
https://github.com/llvm/llvm-project/pull/125673
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -1,8 +1,4 @@
-#if __clang_major__ >= 7
-target datalayout =
"e-p:64:64-p1:64:64-p2:32:32-p3:32:32-p4:64:64-p5:32:32-p6:32:32-i64:64-v16:16-v24:32-v32:32-v48:64-v96:128-v192:256-v256:256-v512:512-v1024:1024-v2048:2048-n32:64-S32-A5"
-#else
-target datalayout =
"e-p:32:32-p1:64
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/127529
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127529
These targets don't include all OpenCL builtins, so there will always be
external calls in the final bytecode module.
Fixes #127316.
>From 3b172c687e3c276eea4489a39de5ed285c8ddd0e Mon Sep 17 00:00:00 2001
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/126052
>From b09ec4e569a2659eca66dc38e35a02a87134c209 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 28 Jan 2025 12:19:23 +
Subject: [PATCH 1/2] [libclc] Move fma to the CLC library
This builtin is a
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/127828
>From 34c217f46e81cbceda12ffa47c1825a57706fdc8 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Wed, 19 Feb 2025 16:25:22 +
Subject: [PATCH] [libclc] Move modf to the CLC library
The "generic" unary_
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127828
The "generic" unary_(def|decl)_with_ptr files are intended to be re-used by the
sincos and fract builtins in the future as they share an identical type
signature.
>From 09835c009430464426dced5598a5e8a9cae1
Author: Fraser Cormack
Date: 2025-02-19T15:18:05Z
New Revision: fb5a87e1a6febb2a81fd85f800e78c2e6dff5715
URL:
https://github.com/llvm/llvm-project/commit/fb5a87e1a6febb2a81fd85f800e78c2e6dff5715
DIFF:
https://github.com/llvm/llvm-project/commit/fb5a87e1a6febb2a81fd85f800e78c2e6dff5715.diff
LOG
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127823
Doing so provides stability when compiling the builtins in a mode in which
unqualified pointers may be interpreted as being in the generic address space,
such as in OpenCL 3.0.
We eventually want to provid
frasercrmck wrote:
ping, thanks
https://github.com/llvm/llvm-project/pull/125673
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -46,9 +46,7 @@ _CLC_DEF _CLC_OVERLOAD __CLC_GENTYPE
__clc_hypot(__CLC_GENTYPE x,
__CLC_GENTYPE retval = __clc_sqrt(__clc_mad(fx, fx, fy * fy)) * fx_exp;
retval = (ux > PINFBITPATT_SP32 || uy == 0) ? __CLC_AS_GENTYPE(ux) : retval;
- retval = (ux == PINFBITPATT_SP32 ||
https://github.com/frasercrmck milestoned
https://github.com/llvm/llvm-project/pull/126908
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/126908
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
frasercrmck wrote:
/cherry-pick a2b05761724e5243056988d9d6bf1a5a94715b74
https://github.com/llvm/llvm-project/pull/126908
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/129045
>From 4efd607ca71b91847c9271740df7aab4534fe01b Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Thu, 27 Feb 2025 12:48:25 +
Subject: [PATCH 1/3] [libclc] Move rsqrt to the CLC library
This also adds
@@ -0,0 +1,4 @@
+__attribute__((weak)) _CLC_OVERLOAD _CLC_DEF __CLC_GENTYPE
+__clc_rsqrt(__CLC_GENTYPE val) {
+ return __CLC_FP_LIT(1.0) / __clc_sqrt(val);
frasercrmck wrote:
Good point. Perhaps we just use the elementwise sqrt builtin. I'm not sure I
want to t
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/129045
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/129871
Also replace some magic constants with named ones.
Checking against FP zero and using isnan and isinf functions allows the
optimizer to create one unified @llvm.is.fpclass intrinsic. This results in
fewer
@@ -86,13 +84,10 @@ _CLC_DEF _CLC_OVERLOAD __CLC_GENTYPE
__clc_hypot(__CLC_GENTYPE x,
r = c ? s : r;
// Check for NaN
- c = __clc_isnan(x) || __clc_isnan(y);
- r = c ? __CLC_AS_GENTYPE((__CLC_ULONGN)QNANBITPATT_DP64) : r;
+ r = __clc_isnan(x) || __clc_isnan(y) ? __CLC_
@@ -46,9 +46,7 @@ _CLC_DEF _CLC_OVERLOAD __CLC_GENTYPE
__clc_hypot(__CLC_GENTYPE x,
__CLC_GENTYPE retval = __clc_sqrt(__clc_mad(fx, fx, fy * fy)) * fx_exp;
retval = (ux > PINFBITPATT_SP32 || uy == 0) ? __CLC_AS_GENTYPE(ux) : retval;
- retval = (ux == PINFBITPATT_SP32 ||
frasercrmck wrote:
It's worth noting that alive2 considers this a correct transformation for the
before/after IR, for the scalar float/double versions of __clc_frexp for amdgcn.
https://github.com/llvm/llvm-project/pull/129871
___
cfe-commits mailing
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/129871
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
frasercrmck wrote:
It looks that with
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33805 mesa is no
longer using the libclc headers.
https://github.com/llvm/llvm-project/pull/126908
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127469
This commit improves the behaviour of (__clc_)nextafter around zero.
Specifically, the nextafter value of very small negative numbers in the
positive direction is now negative zero. Previously we'd return p
frasercrmck wrote:
With this PR and with #126905 I achieved a fully passing OpenCL-CTS conversions
run.
https://github.com/llvm/llvm-project/pull/127469
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/lis
frasercrmck wrote:
See #127469 for the fix for the other conversion builtins.
https://github.com/llvm/llvm-project/pull/126905
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/127836
This commit moves the frexp builtin to the CLC library.
It simultaneously optimizes the code generated for half vectors, which was
previously scalarizing and casting up to float. With this commit it still c
@@ -0,0 +1,73 @@
+#if __CLC_FPSIZE == 32
+
+_CLC_OVERLOAD _CLC_DEF __CLC_INTN __clc_ilogb(__CLC_GENTYPE x) {
+ __CLC_UINTN ux = __CLC_AS_UINTN(x);
frasercrmck wrote:
Yeah, I'm not sure using `frexp` is currently better given that there's no
native path for it.
https://github.com/frasercrmck approved this pull request.
https://github.com/llvm/llvm-project/pull/128692
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
frasercrmck wrote:
ping. Are the discussions above blocking the PR?
If clang's internal headers are unsuitable must we fix those before landing
this, or temporarily fix the installation of libclc's headers instead?
https://github.com/llvm/llvm-project/pull/126908
__
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/128540
>From 0417fefc8b1d6b65941e48a1c2c4c700f601b38b Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Mon, 24 Feb 2025 17:27:59 +
Subject: [PATCH 1/2] [libclc] Move log/log2/log10 to CLC library
No changes
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/128540
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -0,0 +1,339 @@
+/*
+ * Copyright (c) 2014,2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
deal
+ * in the Software without restri
@@ -0,0 +1,339 @@
+/*
+ * Copyright (c) 2014,2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
deal
+ * in the Software without restri
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/126905
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/126052
>From 18d67915671a081540abde18b1b6c009ac094a27 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 28 Jan 2025 12:19:23 +
Subject: [PATCH 1/2] [libclc] Move fma to the CLC library
This builtin is a
frasercrmck wrote:
> > FYI @rjodinchr this breaks downstream clspv testing.
> > ```
> > # | declare !kernel_arg_name !7 float @llvm.fma.f32(float, float, float) #2
> > # | Unsupported llvm intrinsic
> > # | UNREACHABLE executed at /clspv/lib/SPIRVProducerPass.cpp:2390!
> > ```
> >
> >
> >
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/126052
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128500
The clc and clc+clspv modes produced the same conversions code, so this patch
simplifies the process. It further simplifies the internal checks the script
makes by assuming the mutual exclusivity.
>From ce
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128500
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128506
Note the CLC versions of these builtins don't offer the vector/scalar forms,
for simplicity. The OpenCL layer converts the vector/scalar form to
vector/vector.
The CLC builtins use clang's __builtin_elemen
frasercrmck wrote:
Note @arsenm I didn't touch [amdgcn's
fmin/fmax](https://github.com/llvm/llvm-project/blob/main/libclc/amdgcn/lib/math/fmin.cl)
or
[r600's](https://github.com/llvm/llvm-project/blob/main/libclc/r600/lib/math/fmin.cl)
as I wasn't sure if any of that could be updated or at le
frasercrmck wrote:
> These should use the regular builtin fmin / fmax.
Do you mean the AMD implementations, or the CLC ones too? Note there's no
vector support for `__builtin_fmin` which is why I chose
`__builtin_elementwise_min`. They appear to generate the same code so maybe I'm
misundersta
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128521
None
>From 49a2cb5fc5a098dc16f9e348eba8a93b3433a379 Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Mon, 24 Feb 2025 15:11:07 +
Subject: [PATCH] [libclc] Move nan to the CLC library
---
libclc/clc
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128521
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128540
No changes to codegen for AMDGPU/Nvidia targets.
Note that for simplicity this commit doesn't try to refactor or optimize the
implementations. Notably, each log is only implementated for scalar types;
vect
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128012
Several users of (mostly math/) gentype.inc rely on types other than the
'gentype'. This is commonly intN as several maths builtins expose this as a
return or paramter type. We were previously explicitly de
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128012
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128028
This commit moves the logb and ilogb builtins to the CLC library.
It simultaneously optimizes them both for vector types and for half types.
Vector types were being scalarized in some cases. Half types were
frasercrmck wrote:
> This commit is causing regressions on the OpenCL-CTS with `clvk/clspv`. I
> need more time to understand what the issue is exactly.
Interesting, I'm sorry about that. I diffed `clc-convert.cl` and
`clc-clspv-convert.cl` before this change and they were identical, so the
c
@@ -21,19 +21,17 @@
*/
#if __CLC_FPSIZE == 64
-#define __CLC_NAN __builtin_nan("")
-#define ZERO 0.0
+#define __CLC_NAN DBL_NAN
#elif __CLC_FPSIZE == 32
-#define __CLC_NAN NAN
-#define ZERO 0.0f
+#define __CLC_NAN FLT_NAN
#elif __CLC_FPSIZE == 16
-#define __CLC_NAN (half)NA
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/126078
>From 27f4e94de5cb6f1220ddbd4b0197a44cb831d0fe Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Wed, 5 Feb 2025 10:38:46 +
Subject: [PATCH] [libclc] Move __clc_ldexp to CLC library
This function was
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/126078
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128864
There is a long-standing workaround in the libclc build system that silences a
warning about the use of parentheses in bitwise conditional operations.
In an effort to remove this workaround, this commit re-
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128864
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128721
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128748
This is fairly straightforward for most targets.
AMDGPU provides its own implementation of sqrt for double types. This commit
moves this into the implementation of CLC sqrt. It uses weak linkage on the
'de
frasercrmck wrote:
An update on this: in adding a clang elementwise builtin for ldexp (and for
clz/ctz) I found it necessary to first refactor the clang diagnostics. I opened
a PR for it (https://github.com/llvm/llvm-project/pull/125673) which isn't
getting much traction so that's slowing thin
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128708
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128721
None
>From f8a607fc8842d101e54648e3970cfa9aff81e6aa Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 25 Feb 2025 13:13:09 +
Subject: [PATCH] [libclc] Move round to CLC library
---
libclc/clc/i
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/128708
We want to move away from using asm declarations to define builtins.
>From 76d6da267c5991374579031d402be33da88ed43f Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 25 Feb 2025 12:48:27 +
Subjec
@@ -2266,8 +2266,10 @@ bool SemaHLSL::CheckBuiltinFunctionCall(unsigned
BuiltinID, CallExpr *TheCall) {
if (CheckVectorElementCallArgs(&SemaRef, TheCall))
return true;
if (SemaRef.BuiltinElementwiseTernaryMath(
-TheCall, /*CheckForFloatArgs*/
-
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128540
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck updated
https://github.com/llvm/llvm-project/pull/128748
>From f827260918c81053d833db6fe81039b7c0990a2f Mon Sep 17 00:00:00 2001
From: Fraser Cormack
Date: Tue, 25 Feb 2025 12:27:06 +
Subject: [PATCH 1/3] [libclc] Move sqrt to CLC library
This is fairly strai
@@ -21,19 +21,17 @@
*/
#if __CLC_FPSIZE == 64
-#define __CLC_NAN __builtin_nan("")
-#define ZERO 0.0
+#define __CLC_NAN DBL_NAN
#elif __CLC_FPSIZE == 32
-#define __CLC_NAN NAN
-#define ZERO 0.0f
+#define __CLC_NAN FLT_NAN
#elif __CLC_FPSIZE == 16
-#define __CLC_NAN (half)NA
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/128748
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/128748
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck closed
https://github.com/llvm/llvm-project/pull/128748
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/129045
This also adds missing half variants to certain targets.
It also optimizes some targets' implementations to perform the operation
directly in vector types, as opposed to scalarizing.
>From 4efd607ca71b9184
frasercrmck wrote:
Just realised I missed the brief - `native_exp2` should use the builtin,
`native_exp` should continue to call `native_exp2`.
https://github.com/llvm/llvm-project/pull/133696
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
ht
https://github.com/frasercrmck edited
https://github.com/llvm/llvm-project/pull/133696
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
frasercrmck wrote:
> Just realised I missed the brief - `native_exp2` should use the builtin,
> `native_exp` should continue to call `native_exp2`.
Done
https://github.com/llvm/llvm-project/pull/133696
___
cfe-commits mailing list
cfe-commits@lists.l
https://github.com/frasercrmck created
https://github.com/llvm/llvm-project/pull/133696
This came up during a discussion on #129679, which has been split out as a
preparatory commit.
An example of the AMDGPU codegen is:
define <2 x float> @_Z10native_expDv2_f(<2 x float> %x) {
%0 =
501 - 600 of 724 matches
Mail list logo
