[llvm] LangRef: Claim llvm.{vp.,}vector.reduce.f{min,max} as deprecated (PR #118730)

[YunQiang Su via llvm-commits]

https://github.com/wzssyqa updated https://github.com/llvm/llvm-project/pull/118730

>From 6c9cda84636eeaaf4a519b9e7cfebadf08aaf026 Mon Sep 17 00:00:00 2001
From: YunQiang Su <yunqiang at isrc.iscas.ac.cn>
Date: Thu, 5 Dec 2024 10:00:15 +0800
Subject: [PATCH 1/2] LangRef: Claim llvm.{vp.,}vector.reduce.f{min,max} as
 deprecated

The semantics about sNaN of these instructions are not clear enough.
If there is a sNaN in the vector, the result may differs due to the
position of the sNaN value.

Since then, it is difficult to optimize with Vector or SIMD extensions.
---
 llvm/docs/LangRef.rst | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst
index 655ec4389456c..082009e7eb9da 100644
--- a/llvm/docs/LangRef.rst
+++ b/llvm/docs/LangRef.rst
@@ -19870,6 +19870,10 @@ This instruction has the same comparison semantics as the '``llvm.maxnum.*``'
 intrinsic.  If the intrinsic call has the ``nnan`` fast-math flag, then the
 operation can assume that NaNs are not present in the input vector.
 
+It is deprecated, since the different order of inputs may produce different
+outputs, and it is hard to optimize with Vector or SIMD extensions.
+Use '``llvm.vector.reduce.fmaximum``' or '``llvm.vector.reduce.fmaximumnum``' instead.
+
 Arguments:
 """"""""""
 The argument to this intrinsic must be a vector of floating-point values.
@@ -19899,6 +19903,10 @@ This instruction has the same comparison semantics as the '``llvm.minnum.*``'
 intrinsic. If the intrinsic call has the ``nnan`` fast-math flag, then the
 operation can assume that NaNs are not present in the input vector.
 
+It is deprecated, since the different order of inputs may produce different
+outputs, and it is hard to optimize with Vector or SIMD extensions.
+Use '``llvm.vector.reduce.fminimum``' or '``llvm.vector.reduce.fminimumnum``' instead.
+
 Arguments:
 """"""""""
 The argument to this intrinsic must be a vector of floating-point values.
@@ -23528,6 +23536,10 @@ This instruction has the same comparison semantics as the
 
 To ignore the start value, the neutral value can be used.
 
+It is deprecated, since the different order of inputs may produce different
+outputs, and it is hard to optimize with Vector or SIMD extensions.
+Use '``llvm.vp.vector.reduce.fmaximum``' or '``llvm.vp.vector.reduce.fmaximumnum``' instead.
+
 Examples:
 """""""""
 
@@ -23595,6 +23607,10 @@ This instruction has the same comparison semantics as the
 
 To ignore the start value, the neutral value can be used.
 
+It is deprecated, since the different order of inputs may produce different
+outputs, and it is hard to optimize with Vector or SIMD extensions.
+Use '``llvm.vp.vector.reduce.fminimum``' or '``llvm.vp.vector.reduce.fminimumnum``' instead.
+
 Examples:
 """""""""
 

>From 098c6737a7614091bfe0b9e0160f9794828dd4f5 Mon Sep 17 00:00:00 2001
From: YunQiang Su <yunqiang at isrc.iscas.ac.cn>
Date: Mon, 31 Mar 2025 09:47:42 +0800
Subject: [PATCH 2/2] Use wording 'don't use'

---
 llvm/docs/LangRef.rst | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/llvm/docs/LangRef.rst b/llvm/docs/LangRef.rst
index 082009e7eb9da..cb97cc4087090 100644
--- a/llvm/docs/LangRef.rst
+++ b/llvm/docs/LangRef.rst
@@ -19870,7 +19870,7 @@ This instruction has the same comparison semantics as the '``llvm.maxnum.*``'
 intrinsic.  If the intrinsic call has the ``nnan`` fast-math flag, then the
 operation can assume that NaNs are not present in the input vector.
 
-It is deprecated, since the different order of inputs may produce different
+Don't use it if possible, since the different order of inputs may produce different
 outputs, and it is hard to optimize with Vector or SIMD extensions.
 Use '``llvm.vector.reduce.fmaximum``' or '``llvm.vector.reduce.fmaximumnum``' instead.
 
@@ -19903,7 +19903,7 @@ This instruction has the same comparison semantics as the '``llvm.minnum.*``'
 intrinsic. If the intrinsic call has the ``nnan`` fast-math flag, then the
 operation can assume that NaNs are not present in the input vector.
 
-It is deprecated, since the different order of inputs may produce different
+Don't use it if possible, since the different order of inputs may produce different
 outputs, and it is hard to optimize with Vector or SIMD extensions.
 Use '``llvm.vector.reduce.fminimum``' or '``llvm.vector.reduce.fminimumnum``' instead.
 
@@ -23536,7 +23536,7 @@ This instruction has the same comparison semantics as the
 
 To ignore the start value, the neutral value can be used.
 
-It is deprecated, since the different order of inputs may produce different
+Don't use it if possible, since the different order of inputs may produce different
 outputs, and it is hard to optimize with Vector or SIMD extensions.
 Use '``llvm.vp.vector.reduce.fmaximum``' or '``llvm.vp.vector.reduce.fmaximumnum``' instead.
 
@@ -23607,7 +23607,7 @@ This instruction has the same comparison semantics as the
 
 To ignore the start value, the neutral value can be used.
 
-It is deprecated, since the different order of inputs may produce different
+Don't use it if possible, since the different order of inputs may produce different
 outputs, and it is hard to optimize with Vector or SIMD extensions.
 Use '``llvm.vp.vector.reduce.fminimum``' or '``llvm.vp.vector.reduce.fminimumnum``' instead.