[llvm] AMDGPU/Docs: Fix target properties for gfx9-4-generic (PR #125593)

[Konstantin Zhuravlyov via llvm-commits]

https://github.com/kzhuravl updated https://github.com/llvm/llvm-project/pull/125593

>From e3fa0c489906c3c5b9790969561a5253e9cebc26 Mon Sep 17 00:00:00 2001
From: Konstantin Zhuravlyov <kzhuravl_dev at outlook.com>
Date: Mon, 3 Feb 2025 17:48:49 -0500
Subject: [PATCH 1/3] AMDGPU/Docs: Fix target properties for gfx9-4-generic

gfx9-4-generic has architected flat scratch, not absolute
---
 llvm/docs/AMDGPUUsage.rst | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/llvm/docs/AMDGPUUsage.rst b/llvm/docs/AMDGPUUsage.rst
index b646621d12eb0db..41c7aa5a1125446 100644
--- a/llvm/docs/AMDGPUUsage.rst
+++ b/llvm/docs/AMDGPUUsage.rst
@@ -583,8 +583,8 @@ Generic processor code objects are versioned. See :ref:`amdgpu-generic-processor
                                                                                                   - ``v_dot2_f32_f16``
 
 
-     ``gfx9-4-generic``   ``amdgcn``     - ``gfx940``      - xnack            - Absolute flat   FP8 and BF8 instructions,
-                                         - ``gfx941``      - sramecc            scratch         FP8 and BF8 conversion instructions,
+     ``gfx9-4-generic``   ``amdgcn``     - ``gfx940``      - xnack            - Architected     FP8 and BF8 instructions,
+                                         - ``gfx941``      - sramecc            flat scratch    FP8 and BF8 conversion instructions,
                                          - ``gfx942``                                           as well as instructions with XF32 format support
                                          - ``gfx950``                                           are not available.
 

>From a18fb710513b12569f8c274a55b2ce53311fec9b Mon Sep 17 00:00:00 2001
From: Konstantin Zhuravlyov <kzhuravl_dev at outlook.com>
Date: Tue, 4 Feb 2025 14:54:39 -0500
Subject: [PATCH 2/3] Address review feedback from Joe and Shilei

---
 llvm/docs/AMDGPUUsage.rst | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/llvm/docs/AMDGPUUsage.rst b/llvm/docs/AMDGPUUsage.rst
index 41c7aa5a1125446..dde6adf72483297 100644
--- a/llvm/docs/AMDGPUUsage.rst
+++ b/llvm/docs/AMDGPUUsage.rst
@@ -583,11 +583,11 @@ Generic processor code objects are versioned. See :ref:`amdgpu-generic-processor
                                                                                                   - ``v_dot2_f32_f16``
 
 
-     ``gfx9-4-generic``   ``amdgcn``     - ``gfx940``      - xnack            - Architected     FP8 and BF8 instructions,
-                                         - ``gfx941``      - sramecc            flat scratch    FP8 and BF8 conversion instructions,
-                                         - ``gfx942``                                           as well as instructions with XF32 format support
-                                         - ``gfx950``                                           are not available.
-
+     ``gfx9-4-generic``   ``amdgcn``     - ``gfx940``      - sramecc          - Architected     FP8 and BF8 instructions,
+                                         - ``gfx941``      - tgsplit            flat scratch    FP8 and BF8 conversion instructions,
+                                         - ``gfx942``      - xnack            - Packed          as well as instructions with XF32 format support
+                                         - ``gfx950``      - kernarg preload    work-item        are not available.
+                                                                                IDs
 
      ``gfx10-1-generic``  ``amdgcn``     - ``gfx1010``     - xnack            - Absolute flat   - The following instructions are
                                          - ``gfx1011``     - wavefrontsize64    scratch           not available on ``gfx1011``

>From 319c9bfadd652ec202b90f65fd17fa82859d1486 Mon Sep 17 00:00:00 2001
From: Konstantin Zhuravlyov <kzhuravl_dev at outlook.com>
Date: Tue, 4 Feb 2025 19:39:39 -0500
Subject: [PATCH 3/3] Fix formatting warnings

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

diff --git a/llvm/docs/AMDGPUUsage.rst b/llvm/docs/AMDGPUUsage.rst
index dde6adf72483297..84980d0c31d4f98 100644
--- a/llvm/docs/AMDGPUUsage.rst
+++ b/llvm/docs/AMDGPUUsage.rst
@@ -584,10 +584,10 @@ Generic processor code objects are versioned. See :ref:`amdgpu-generic-processor
 
 
      ``gfx9-4-generic``   ``amdgcn``     - ``gfx940``      - sramecc          - Architected     FP8 and BF8 instructions,
-                                         - ``gfx941``      - tgsplit            flat scratch    FP8 and BF8 conversion instructions,
-                                         - ``gfx942``      - xnack            - Packed          as well as instructions with XF32 format support
-                                         - ``gfx950``      - kernarg preload    work-item        are not available.
-                                                                                IDs
+                                         - ``gfx941``      - tgsplit            flat scratch    FP8 and BF8 conversion
+                                         - ``gfx942``      - xnack            - Packed          instructions, as well as
+                                         - ``gfx950``      - kernarg preload    work-item       instructions with XF32 format
+                                                                                IDs             support are not available.
 
      ``gfx10-1-generic``  ``amdgcn``     - ``gfx1010``     - xnack            - Absolute flat   - The following instructions are
                                          - ``gfx1011``     - wavefrontsize64    scratch           not available on ``gfx1011``