Move general documentation into Sphinx (#2227)

Summary:
- Move general build, installation, and test documentation into Sphinx

Pull Request resolved: #2227

Reviewed By: spcyppt

Differential Revision: D52323411

Pulled By: q10

fbshipit-source-id: acf3f71af2241d1da7cd5092d1f3520afa14d367

Author: q10

.github/scripts/fbgemm_gpu_build.bash

@@ -77,6 +77,8 @@ __configure_fbgemm_gpu_build_rocm () {
       echo "[BUILD] Architectures list from rocminfo: ${arch_list}"
 
       if [ "$arch_list" == "" ]; then
+        # It is possible to build FBGEMM_GPU-ROCm on a machine without AMD
+        # cards, in which case the arch_list will be empty.
         echo "[BUILD] rocminfo did not return anything valid!"
 
         # By default, we build just for MI100 and MI250 to save time.  This list

.github/workflows/fbgemm_gpu_docs.yml

@@ -37,7 +37,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: [ "3.11" ]
+        python-version: [ "3.12" ]
 
     steps:
     - name: Setup Build Container

fbgemm_gpu/README.md

@@ -20,15 +20,15 @@ See our [Documentation](docs/README.md) for more information.
 
 The full installation instructions
 for the CUDA, ROCm, and CPU-only variants of FBGEMM_GPU can be found
-[here](docs/InstallationInstructions.md).  In addition, instructions for running
-example tests and benchmarks can be found [here](docs/TestInstructions.md).
+[here](docs/src/general/InstallationInstructions.rst).  In addition, instructions for running
+example tests and benchmarks can be found [here](docs/src/general/TestInstructions.rst).
 
 
 ## Build Instructions
 
 This section is intended for FBGEMM_GPU developers only.  The full build
 instructions for the CUDA, ROCm, and CPU-only variants of FBGEMM_GPU can be
-found [here](docs/BuildInstructions.md).
+found [here](docs/src/general/BuildInstructions.rst).
 
 
 ## Join the FBGEMM_GPU Community

fbgemm_gpu/codegen/embedding_bounds_check_host.cpp

@@ -16,7 +16,7 @@
 
 using Tensor = at::Tensor;
 
-///@defgroup embedding-cuda Embedding CUDA Operators
+/// @defgroup embedding-cuda Embedding CUDA Operators
 
 ///@ingroup embedding-cuda
 void bounds_check_indices_cuda(

fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp

@@ -17,7 +17,7 @@
 using Tensor = at::Tensor;
 using namespace fbgemm_gpu;
 
-///@defgroup embedding-cpu Embedding CPU Operators
+/// @defgroup embedding-cpu Embedding CPU Operators
 ///
 
 namespace {

fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp

@@ -184,7 +184,7 @@ void process_uvm_cache_stats(
 } // namespace
 #endif
 
-///@defgroup embedding-cuda Embedding CUDA Operators
+/// @defgroup embedding-cuda Embedding CUDA Operators
 ///
 
 Tensor int_nbit_split_embedding_codegen_forward_unweighted_cuda(

fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp

@@ -24,7 +24,7 @@
 using Tensor = at::Tensor;
 using namespace fbgemm_gpu;
 
-///@defgroup embedding-cpu Embedding CPU Operators
+/// @defgroup embedding-cpu Embedding CPU Operators
 ///
 
 Tensor int_nbit_split_embedding_codegen_forward_unweighted_cpu(