h-vetinari wrote:
@meinersbur, with the patches from this PR (but unchanged recipe otherwise),
I'm running into:
```
-- The Fortran compiler identification is unknown
error: unknown integrated tool '-cc1'. Valid tools include '-fc1'.
-- Detecting Fortran compiler ABI info
CMake Error: Error requ
https://github.com/aartbik commented:
+1 on the simplification for the sparse 1:N conversion!
https://github.com/llvm/llvm-project/pull/116524
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailma
Author: Vy Nguyen
Date: 2024-12-20T11:26:33-05:00
New Revision: 9176779749c91cf708dd6a927e52630523d72ee9
URL:
https://github.com/llvm/llvm-project/commit/9176779749c91cf708dd6a927e52630523d72ee9
DIFF:
https://github.com/llvm/llvm-project/commit/9176779749c91cf708dd6a927e52630523d72ee9.diff
LOG
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/6] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/6] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/matthias-springer edited
https://github.com/llvm/llvm-project/pull/120746
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
llvmbot wrote:
@llvm/pr-subscribers-mlir-core
Author: Matthias Springer (matthias-springer)
Changes
In the dialect conversion driver, use `DominanceInfo` to compute a suitable
insertion point for N:1 source materializations.
Note: This PR is related to #114940, but I could't reopen it.
llvmbot wrote:
@llvm/pr-subscribers-mlir
Author: Matthias Springer (matthias-springer)
Changes
In the dialect conversion driver, use `DominanceInfo` to compute a suitable
insertion point for N:1 source materializations.
Note: This PR is related to #114940, but I could't reopen it.
---
https://github.com/matthias-springer created
https://github.com/llvm/llvm-project/pull/120746
In the dialect conversion driver, use `DominanceInfo` to compute a suitable
insertion point for N:1 source materializations.
Note: This PR is related to #114940, but I could't reopen it.
>From 61b51
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/6] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/5] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/matthias-springer updated
https://github.com/llvm/llvm-project/pull/115587
>From b484c72cf7a79c7203b853725761e08c2233241c Mon Sep 17 00:00:00 2001
From: Matthias Springer
Date: Sat, 9 Nov 2024 07:13:07 +0100
Subject: [PATCH] [mlir][IR][NFC] `DominanceInfo`: Share same impl fo
https://github.com/matthias-springer edited
https://github.com/llvm/llvm-project/pull/115587
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -1986,10 +1986,12 @@ void
VPReverseVectorPointerRecipe::execute(VPTransformState &State) {
// LastLane = 1 - RunTimeVF
Value *LastLane = Builder.CreateSub(ConstantInt::get(IndexTy, 1), RunTimeVF);
Value *Ptr = State.get(getOperand(0), VPLane(0));
- Value *ResultPtr =
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/3] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
@@ -53,6 +54,55 @@ static void logFailure(llvm::ScopedPrinter &os, StringRef
fmt, Args &&...args) {
});
}
+/// Given two insertion points in the same block, choose the later one.
+static OpBuilder::InsertPoint
+chooseLaterInsertPointInBlock(OpBuilder::InsertPoint a,
+
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/2] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/2] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
llvmbot wrote:
@llvm/pr-subscribers-mlir-func
Author: Matthias Springer (matthias-springer)
Changes
This commit updates the internal `ConversionValueMapping` data structure in the
dialect conversion driver to support 1:N replacements. This is the last major
commit for adding 1:N support
llvmbot wrote:
@llvm/pr-subscribers-mlir-linalg
Author: Matthias Springer (matthias-springer)
Changes
This commit updates the internal `ConversionValueMapping` data structure in the
dialect conversion driver to support 1:N replacements. This is the last major
commit for adding 1:N suppor
llvmbot wrote:
@llvm/pr-subscribers-mlir-quant
@llvm/pr-subscribers-mlir-emitc
@llvm/pr-subscribers-flang-codegen
Author: Matthias Springer (matthias-springer)
Changes
This commit updates the internal `ConversionValueMapping` data structure in the
dialect conversion driver to support 1:N
https://github.com/matthias-springer ready_for_review
https://github.com/llvm/llvm-project/pull/116524
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/matthias-springer edited
https://github.com/llvm/llvm-project/pull/116524
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/matthias-springer edited
https://github.com/llvm/llvm-project/pull/116524
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/matthias-springer edited
https://github.com/llvm/llvm-project/pull/116524
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/matthias-springer updated
https://github.com/llvm/llvm-project/pull/116524
>From 53f97f5a68033bd46ffd5a982435d64afe9048dd Mon Sep 17 00:00:00 2001
From: Matthias Springer
Date: Sun, 15 Dec 2024 17:36:49 +0100
Subject: [PATCH] ex
---
.../lib/Optimizer/CodeGen/BoxedProcedure.
https://github.com/fhahn edited https://github.com/llvm/llvm-project/pull/120730
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
https://github.com/fhahn commented:
To clarify, this is only an issue, if we are folding the tail, as then we may
compute an address that we don't in the original scalar loop, so may not be
inbounds? If we don't fold the tail, it should be fine to keep the `inbounds`
flag?
https://github.com/
@@ -1986,10 +1986,10 @@ void
VPReverseVectorPointerRecipe::execute(VPTransformState &State) {
// LastLane = 1 - RunTimeVF
Value *LastLane = Builder.CreateSub(ConstantInt::get(IndexTy, 1), RunTimeVF);
Value *Ptr = State.get(getOperand(0), VPLane(0));
- Value *ResultPtr =
@@ -1986,10 +1986,12 @@ void
VPReverseVectorPointerRecipe::execute(VPTransformState &State) {
// LastLane = 1 - RunTimeVF
Value *LastLane = Builder.CreateSub(ConstantInt::get(IndexTy, 1), RunTimeVF);
Value *Ptr = State.get(getOperand(0), VPLane(0));
- Value *ResultPtr =
@@ -1986,10 +1986,12 @@ void
VPReverseVectorPointerRecipe::execute(VPTransformState &State) {
// LastLane = 1 - RunTimeVF
Value *LastLane = Builder.CreateSub(ConstantInt::get(IndexTy, 1), RunTimeVF);
Value *Ptr = State.get(getOperand(0), VPLane(0));
- Value *ResultPtr =
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH 1/2] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Ve
https://github.com/fmayer updated
https://github.com/llvm/llvm-project/pull/120730
>From 5987219575feabd0eefba5932c21b0eba8ae4fb7 Mon Sep 17 00:00:00 2001
From: Florian Mayer
Date: Fri, 20 Dec 2024 05:35:56 -0800
Subject: [PATCH] simplify
Created using spr 1.3.4
---
llvm/lib/Transforms/Vector
@@ -1986,10 +1986,12 @@ void
VPReverseVectorPointerRecipe::execute(VPTransformState &State) {
// LastLane = 1 - RunTimeVF
Value *LastLane = Builder.CreateSub(ConstantInt::get(IndexTy, 1), RunTimeVF);
Value *Ptr = State.get(getOperand(0), VPLane(0));
- Value *ResultPtr =
https://github.com/arsenm edited
https://github.com/llvm/llvm-project/pull/120598
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -146,6 +146,54 @@ define <1 x i64> @atomic_vec1_i64_align(ptr %x) nounwind {
ret <1 x i64> %ret
}
+define <2 x i8> @atomic_vec2_i8(ptr %x) {
+; CHECK3-LABEL: atomic_vec2_i8:
+; CHECK3: ## %bb.0:
+; CHECK3-NEXT:movzwl (%rdi), %eax
+; CHECK3-NEXT:movd %eax, %x
https://github.com/arsenm commented:
Description says it's invalid, but it is now valid
https://github.com/llvm/llvm-project/pull/120598
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/lis
llvmbot wrote:
@llvm/pr-subscribers-llvm-transforms
Author: Florian Mayer (fmayer)
Changes
---
Full diff: https://github.com/llvm/llvm-project/pull/120730.diff
2 Files Affected:
- (modified) llvm/lib/Transforms/Vectorize/VPlanRecipes.cpp (+6-4)
- (modified) llvm/test/Transforms/LoopV
https://github.com/fmayer created
https://github.com/llvm/llvm-project/pull/120730
None
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
@@ -145,3 +145,40 @@ define <1 x i64> @atomic_vec1_i64_align(ptr %x) nounwind {
%ret = load atomic <1 x i64>, ptr %x acquire, align 8
ret <1 x i64> %ret
}
+
+define <1 x half> @atomic_vec1_half(ptr %x) {
+; CHECK3-LABEL: atomic_vec1_half:
+; CHECK3: ## %bb.0:
+; CHECK
@@ -145,3 +145,40 @@ define <1 x i64> @atomic_vec1_i64_align(ptr %x) nounwind {
%ret = load atomic <1 x i64>, ptr %x acquire, align 8
ret <1 x i64> %ret
}
+
+define <1 x half> @atomic_vec1_half(ptr %x) {
arsenm wrote:
I'm not sure why bfloat is special suc
@@ -1146,6 +1146,9 @@ void DAGTypeLegalizer::SplitVectorResult(SDNode *N,
unsigned ResNo) {
SplitVecRes_STEP_VECTOR(N, Lo, Hi);
break;
case ISD::SIGN_EXTEND_INREG: SplitVecRes_InregOp(N, Lo, Hi); break;
+ case ISD::ATOMIC_LOAD:
+SplitVecRes_ATOMIC_LOAD(cast(N),
@@ -983,3 +983,60 @@ define void @store_atomic_f64__seq_cst(ptr %ptr, double
%val1) {
store atomic double %val1, ptr %ptr seq_cst, align 8
ret void
}
+
+define <1 x ptr> @atomic_vec1_ptr(ptr %x) #0 {
arsenm wrote:
Should get IR test in test/Transforms/Ato
@@ -1884,7 +1884,10 @@ bool AtomicExpandImpl::expandAtomicOpToLibcall(
IRBuilder<> Builder(I);
IRBuilder<> AllocaBuilder(&I->getFunction()->getEntryBlock().front());
- bool UseSizedLibcall = canUseSizedAtomicCall(Size, Alignment, DL);
+ const bool IsAtomic =
+ isa(I
arsenm wrote:
> Should we test any of the unaligned ones in this series? Certain aligned
> tests actually also lower to calls, like `%ret = load atomic <2 x
> {double,i64}>, ptr %x acquire, align 16` -> ` callq ___atomic_load_16`, yet
> these have always worked.
Yes, include both. But just th
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120386
>From d08f5c32417cf434e994d0132c5e113682b990e8 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:38:23 -0500
Subject: [PATCH] [X86] Manage atomic load of fp -> int promotion in DAG
When lowering atom
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120598
>From 23c9ff2277008ffeae536deb280affba256de4c3 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Thu, 19 Dec 2024 11:19:39 -0500
Subject: [PATCH] [SelectionDAG] Widen vector types for atomic load
`load atomic ` is not
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120598
>From 23c9ff2277008ffeae536deb280affba256de4c3 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Thu, 19 Dec 2024 11:19:39 -0500
Subject: [PATCH] [SelectionDAG] Widen vector types for atomic load
`load atomic ` is not
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120640
>From db674f8a2f88652bd8cd80fae1e7f9ee8d214f49 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Thu, 19 Dec 2024 16:25:55 -0500
Subject: [PATCH] [SelectionDAG] Split vector types for atomic load
`load atomic ` is not
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120387
>From 592fa22e90dd2fec06f23f95af29795204394c8b Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:40:32 -0500
Subject: [PATCH] [X86] Add atomic vector tests for unaligned >1 sizes.
Unaligned atomic ve
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120385
>From c63ec44f68659a9a8022f9d8125236b6869f1266 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:37:17 -0500
Subject: [PATCH] [SelectionDAG] Legalize <1 x T> vector types for atomic load
`load atomic
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120387
>From 592fa22e90dd2fec06f23f95af29795204394c8b Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:40:32 -0500
Subject: [PATCH] [X86] Add atomic vector tests for unaligned >1 sizes.
Unaligned atomic ve
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120385
>From c63ec44f68659a9a8022f9d8125236b6869f1266 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:37:17 -0500
Subject: [PATCH] [SelectionDAG] Legalize <1 x T> vector types for atomic load
`load atomic
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120716
>From 1a9eae9445c20f8f611964b14315c4d57a93b0cd Mon Sep 17 00:00:00 2001
From: jofrn
Date: Fri, 20 Dec 2024 06:14:28 -0500
Subject: [PATCH] [AtomicExpand] Avoid sized call when expanding load atomic
vector
Atomic
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120385
>From c63ec44f68659a9a8022f9d8125236b6869f1266 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:37:17 -0500
Subject: [PATCH] [SelectionDAG] Legalize <1 x T> vector types for atomic load
`load atomic
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120387
>From 592fa22e90dd2fec06f23f95af29795204394c8b Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:40:32 -0500
Subject: [PATCH] [X86] Add atomic vector tests for unaligned >1 sizes.
Unaligned atomic ve
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120386
>From d08f5c32417cf434e994d0132c5e113682b990e8 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:38:23 -0500
Subject: [PATCH] [X86] Manage atomic load of fp -> int promotion in DAG
When lowering atom
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120386
>From d08f5c32417cf434e994d0132c5e113682b990e8 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Wed, 18 Dec 2024 03:38:23 -0500
Subject: [PATCH] [X86] Manage atomic load of fp -> int promotion in DAG
When lowering atom
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120640
>From db674f8a2f88652bd8cd80fae1e7f9ee8d214f49 Mon Sep 17 00:00:00 2001
From: jofrn
Date: Thu, 19 Dec 2024 16:25:55 -0500
Subject: [PATCH] [SelectionDAG] Split vector types for atomic load
`load atomic ` is not
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120716
>From 1a9eae9445c20f8f611964b14315c4d57a93b0cd Mon Sep 17 00:00:00 2001
From: jofrn
Date: Fri, 20 Dec 2024 06:14:28 -0500
Subject: [PATCH] [AtomicExpand] Avoid sized call when expanding load atomic
vector
Atomic
https://github.com/jofrn updated
https://github.com/llvm/llvm-project/pull/120716
>From 1a9eae9445c20f8f611964b14315c4d57a93b0cd Mon Sep 17 00:00:00 2001
From: jofrn
Date: Fri, 20 Dec 2024 06:14:28 -0500
Subject: [PATCH] [AtomicExpand] Avoid sized call when expanding load atomic
vector
Atomic
@@ -145,3 +145,40 @@ define <1 x i64> @atomic_vec1_i64_align(ptr %x) nounwind {
%ret = load atomic <1 x i64>, ptr %x acquire, align 8
ret <1 x i64> %ret
}
+
+define <1 x half> @atomic_vec1_half(ptr %x) {
jofrn wrote:
`1 x bfloat` is added here:
https://gi
jofrn wrote:
> Title change doesn't make sense. It's less useful to test the underaligned
> cases. The baseline tests must include properly aligned atomics. Those are
> the case that we are interested in surviving until the DAG. The unaligned
> case are a separate point of interest as a supple
https://github.com/arsenm edited
https://github.com/llvm/llvm-project/pull/120199
___
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
64 matches
Mail list logo
