From ee8cb4c8d89bce73f97236e807975dc69d77a600 Mon Sep 17 00:00:00 2001
From: ChrisRackauckas <accounts@chrisrackauckas.com>
Date: Sun, 10 Aug 2025 12:31:56 -0400
Subject: [PATCH 1/2] Add AMDGPU methods to documentation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

- Added AMDGPUOffloadLUFactorization and AMDGPUOffloadQRFactorization to solver docs
- Added example usage in GPU tutorial section
- Documented requirement for AMDGPU.jl package

🤖 Generated with Claude Code

Co-Authored-By: Claude <noreply@anthropic.com>
---
 docs/src/solvers/solvers.md | 13 +++++++++++++
 docs/src/tutorials/gpu.md   | 12 +++++++++++-
 2 files changed, 24 insertions(+), 1 deletion(-)

diff --git a/docs/src/solvers/solvers.md b/docs/src/solvers/solvers.md
index 7bcb11a56..286a5ead2 100644
--- a/docs/src/solvers/solvers.md
+++ b/docs/src/solvers/solvers.md
@@ -238,6 +238,19 @@ The following are non-standard GPU factorization routines.
 CudaOffloadFactorization
 ```
 
+### AMDGPU.jl
+
+The following are GPU factorization routines for AMD GPUs using the ROCm stack.
+
+!!! note
+    
+    Using these solvers requires adding the package AMDGPU.jl, i.e. `using AMDGPU`
+
+```@docs
+AMDGPUOffloadLUFactorization
+AMDGPUOffloadQRFactorization
+```
+
 ### CUSOLVERRF.jl
 
 !!! note
diff --git a/docs/src/tutorials/gpu.md b/docs/src/tutorials/gpu.md
index ee737668c..2bc5fc573 100644
--- a/docs/src/tutorials/gpu.md
+++ b/docs/src/tutorials/gpu.md
@@ -39,11 +39,21 @@ sol.u
 This computation can be moved to the GPU by the following:
 
 ```julia
-using CUDA # Add the GPU library
+using CUDA # Add the GPU library for NVIDIA GPUs
 sol = LS.solve(prob, LS.CudaOffloadFactorization())
 sol.u
 ```
 
+For AMD GPUs, you can use the AMDGPU.jl package:
+
+```julia
+using AMDGPU # Add the GPU library for AMD GPUs
+sol = LS.solve(prob, LS.AMDGPUOffloadLUFactorization())  # LU factorization
+# or
+sol = LS.solve(prob, LS.AMDGPUOffloadQRFactorization())  # QR factorization
+sol.u
+```
+
 ## GPUArray Interface
 
 For more manual control over the factorization setup, you can use the

From 4d78f35362377796e61639a6bb2779e33936aa3a Mon Sep 17 00:00:00 2001
From: Christopher Rackauckas <accounts@chrisrackauckas.com>
Date: Sun, 10 Aug 2025 12:38:10 -0400
Subject: [PATCH 2/2] Update docs/src/tutorials/gpu.md

---
 docs/src/tutorials/gpu.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/src/tutorials/gpu.md b/docs/src/tutorials/gpu.md
index 2bc5fc573..f348c647a 100644
--- a/docs/src/tutorials/gpu.md
+++ b/docs/src/tutorials/gpu.md
@@ -40,7 +40,7 @@ This computation can be moved to the GPU by the following:
 
 ```julia
 using CUDA # Add the GPU library for NVIDIA GPUs
-sol = LS.solve(prob, LS.CudaOffloadFactorization())
+sol = LS.solve(prob, LS.CudaOffloadLUFactorization())
 sol.u
 ```