Readthedocs documentation support (#71)

docs/run_doc.sh

@@ -0,0 +1,16 @@
+#!/bin/bash
+
+if [ -d docBin ]; then
+    rm -rf docBin
+fi
+
+sed -e 's/ROCFFT_EXPORT //g' ../library/include/rocfft.h > rocfft.h
+doxygen Doxyfile
+
+cd source
+make clean
+make html
+cd ..
+
+rm rocfft.h
+

docs/run_doxygen.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+
+if [ -d docBin ]; then
+    rm -rf docBin
+fi
+
+rm nccl.h
+
+sed -e 's/ROCFFT_EXPORT //g' ../src/nccl.h.in > nccl.h
+doxygen Doxyfile
+#rm nccl.h
+

docs/source/Makefile

@@ -0,0 +1,20 @@
+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line.
+SPHINXOPTS    =
+SPHINXBUILD   = sphinx-build
+SPHINXPROJ    = RCCL
+SOURCEDIR     = .
+BUILDDIR      = _build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

docs/source/allapi.rst

@@ -0,0 +1,11 @@
+.. toctree::
+   :maxdepth: 4 
+   :caption: Contents:
+
+=======
+All API
+=======
+
+.. doxygenindex::
+
+ 

docs/source/api.rst

@@ -0,0 +1,103 @@
+.. toctree::
+   :maxdepth: 4 
+   :caption: Contents:
+
+===
+API
+===
+
+This section provides details of the library API
+
+Communicator Functions
+----------------------
+
+.. doxygenfunction:: ncclGetUniqueId
+
+.. doxygenfunction:: ncclCommInitRank
+
+.. doxygenfunction:: ncclCommInitAll
+
+.. doxygenfunction:: ncclCommDestroy
+
+.. doxygenfunction:: ncclCommCount
+
+.. doxygenfunction:: ncclCommCuDevice
+
+.. doxygenfunction:: ncclCommUserRank
+
+Collection Communication Operations
+-----------------------------------
+
+Collective communication operations must be called separately for each communicator in a communicator clique.
+
+They return when operations have been enqueued on the hipstream.
+
+Since they may perform inter-CPU synchronization, each call has to be done from a different thread or process, or need to use Group Semantics (see below).
+
+.. doxygenfunction:: ncclReduce
+
+.. doxygenfunction:: ncclBcast
+
+.. doxygenfunction:: ncclBroadcast
+
+.. doxygenfunction:: ncclAllReduce
+
+.. doxygenfunction:: ncclReduceScatter
+
+.. doxygenfunction:: ncclAllGather
+
+
+Group Semantics
+---------------
+When managing multiple GPUs from a single thread, and since NCCL collective
+calls may perform inter-CPU synchronization, we need to "group" calls for
+different ranks/devices into a single call.
+
+Grouping NCCL calls as being part of the same collective operation is done
+using ncclGroupStart and ncclGroupEnd. ncclGroupStart will enqueue all
+collective calls until the ncclGroupEnd call, which will wait for all calls
+to be complete. Note that for collective communication, ncclGroupEnd only
+guarantees that the operations are enqueued on the streams, not that
+the operation is effectively done.
+
+Both collective communication and ncclCommInitRank can be used in conjunction
+of ncclGroupStart/ncclGroupEnd.
+
+.. doxygenfunction:: ncclGroupStart
+
+.. doxygenfunction:: ncclGroupEnd
+
+Library Functions
+-----------------
+
+.. doxygenfunction:: ncclGetVersion
+
+.. doxygenfunction:: ncclGetErrorString
+
+Types
+-----
+
+There are few data structures that are internal to the library. The pointer types to these
+structures are given below. The user would need to use these types to create handles and pass them
+between different library functions.
+
+.. doxygentypedef:: ncclComm_t
+
+.. doxygenstruct:: ncclUniqueId
+
+
+
+Enumerations
+------------
+
+This section provides all the enumerations used.
+
+.. doxygenenum:: ncclResult_t
+
+.. doxygenenum:: ncclRedOp_t
+
+.. doxygenenum:: ncclDataType_t
+
+
+
+

docs/source/conf.py

@@ -0,0 +1,185 @@
+# -*- coding: utf-8 -*-
+#
+# RCCL documentation build configuration file, created by
+# sphinx-quickstart on Mon Jan  8 16:34:42 2018.
+#
+# This file is execfile()d with the current directory set to its
+# containing dir.
+#
+# Note that not all possible configuration values are present in this
+# autogenerated file.
+#
+# All configuration values have a default; values that are commented out
+# serve to show the default.
+
+# If extensions (or modules to document with autodoc) are in another directory,
+# add these directories to sys.path here. If the directory is relative to the
+# documentation root, use os.path.abspath to make it absolute, like shown here.
+#
+# import os
+# import sys
+# sys.path.insert(0, os.path.abspath('.'))
+
+import os
+import sys
+import subprocess
+
+read_the_docs_build = os.environ.get('READTHEDOCS', None) == 'True'
+
+if read_the_docs_build:
+    subprocess.call('cd ..; ./run_doxygen.sh; cd source', shell=True)
+
+# -- General configuration ------------------------------------------------
+
+# If your documentation needs a minimal Sphinx version, state it here.
+#
+# needs_sphinx = '1.0'
+
+# Add any Sphinx extension module names here, as strings. They can be
+# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom
+# ones.
+extensions = ['sphinx.ext.mathjax', 'breathe']
+breathe_projects = { "RCCL": "../docBin/xml" }
+breathe_default_project = "RCCL"
+
+# Add any paths that contain templates here, relative to this directory.
+templates_path = ['_templates']
+
+# The suffix(es) of source filenames.
+# You can specify multiple suffix as a list of string:
+#
+# source_suffix = ['.rst', '.md']
+source_suffix = '.rst'
+
+# The master toctree document.
+master_doc = 'index'
+
+# General information about the project.
+project = u'RCCL'
+copyright = u'2019, Advanced Mirco Devices'
+author = u'Advanced Mirco Devices'
+
+# The version info for the project you're documenting, acts as replacement for
+# |version| and |release|, also used in various other places throughout the
+# built documents.
+#
+# The short X.Y version.
+version = u'0.8'
+# The full version, including alpha/beta/rc tags.
+release = u'0.8'
+
+# The language for content autogenerated by Sphinx. Refer to documentation
+# for a list of supported languages.
+#
+# This is also used if you do content translation via gettext catalogs.
+# Usually you set "language" from the command line for these cases.
+language = None
+
+# List of patterns, relative to source directory, that match files and
+# directories to ignore when looking for source files.
+# This patterns also effect to html_static_path and html_extra_path
+exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store']
+
+# The name of the Pygments (syntax highlighting) style to use.
+pygments_style = 'sphinx'
+
+# If true, `todo` and `todoList` produce output, else they produce nothing.
+todo_include_todos = False
+
+
+# -- Options for HTML output ----------------------------------------------
+
+# The theme to use for HTML and HTML Help pages.  See the documentation for
+# a list of builtin themes.
+#
+# html_theme = 'alabaster'
+
+if read_the_docs_build:
+    html_theme = 'default'
+else:
+    import sphinx_rtd_theme
+    html_theme = "sphinx_rtd_theme"
+    html_theme_path = [sphinx_rtd_theme.get_html_theme_path()]
+
+# Theme options are theme-specific and customize the look and feel of a theme
+# further.  For a list of options available for each theme, see the
+# documentation.
+#
+# html_theme_options = {}
+
+# Add any paths that contain custom static files (such as style sheets) here,
+# relative to this directory. They are copied after the builtin static files,
+# so a file named "default.css" will overwrite the builtin "default.css".
+# html_static_path = ['_static']
+
+# Custom sidebar templates, must be a dictionary that maps document names
+# to template names.
+#
+# This is required for the alabaster theme
+# refs: http://alabaster.readthedocs.io/en/latest/installation.html#sidebars
+# html_sidebars = {
+#     '**': [
+#         'relations.html',  # needs 'show_related': True theme option to display
+#         'searchbox.html',
+#     ]
+# }
+
+
+# -- Options for HTMLHelp output ------------------------------------------
+
+# Output file base name for HTML help builder.
+htmlhelp_basename = 'RCCLdoc'
+
+
+# -- Options for LaTeX output ---------------------------------------------
+
+latex_elements = {
+    # The paper size ('letterpaper' or 'a4paper').
+    #
+    # 'papersize': 'letterpaper',
+
+    # The font size ('10pt', '11pt' or '12pt').
+    #
+    # 'pointsize': '10pt',
+
+    # Additional stuff for the LaTeX preamble.
+    #
+    # 'preamble': '',
+
+    # Latex figure (float) alignment
+    #
+    # 'figure_align': 'htbp',
+}
+
+# Grouping the document tree into LaTeX files. List of tuples
+# (source start file, target name, title,
+#  author, documentclass [howto, manual, or own class]).
+latex_documents = [
+    (master_doc, 'RCCL.tex', u'RCCL Documentation',
+     u'Advanced Mirco Devices', 'manual'),
+]
+
+
+# -- Options for manual page output ---------------------------------------
+
+# One entry per manual page. List of tuples
+# (source start file, name, description, authors, manual section).
+man_pages = [
+    (master_doc, 'RCCL', u'RCCL Documentation',
+     [author], 1)
+]
+
+
+# -- Options for Texinfo output -------------------------------------------
+
+# Grouping the document tree into Texinfo files. List of tuples
+# (source start file, target name, title, author,
+#  dir menu entry, description, category)
+texinfo_documents = [
+    (master_doc, 'RCCL', u'RCCL Documentation',
+     author, 'RCCL', 'One line description of project.',
+     'Miscellaneous'),
+]
+
+
+

docs/source/index.rst

@@ -0,0 +1,21 @@
+.. rocFFT documentation master file, created by
+   sphinx-quickstart on Mon Jan  8 09:51:41 2018.
+   You can adapt this file completely to your liking, but it should at least
+   contain the root `toctree` directive.
+
+Welcome to RCCL's documentation!
+==================================
+
+.. toctree::
+   :maxdepth: 4 
+   :caption: Contents:
+
+   library 
+   api
+   allapi
+
+Indices and tables
+==================
+
+* :ref:`genindex`
+* :ref:`search`

docs/source/library.rst

@@ -0,0 +1,13 @@
+
+.. toctree::
+   :maxdepth: 4 
+   :caption: Contents:
+
+======
+RCCL
+======
+
+Introduction
+------------
+
+The RCCL is an AMD port of NCCL. 

docs/source/requirements.txt

@@ -0,0 +1,3 @@
+
+breathe
+

src/nccl.h.in

@@ -23,14 +23,16 @@
 extern "C" {
 #endif
 
-/* Opaque handle to communicator */
+/*! @brief Opaque handle to communicator 
+*/
 typedef struct ncclComm* ncclComm_t;
 
 #define NCCL_UNIQUE_ID_BYTES 128
+/*! @brief struct to store ncclUniqueId */
 typedef struct { char internal[NCCL_UNIQUE_ID_BYTES]; } ncclUniqueId;
 
-/* Error type */
-typedef enum { ncclSuccess                 =  0,
+/*! @brief Error type */
+  typedef enum { ncclSuccess                 =  0, /**< Successfuly ran */
                ncclUnhandledCudaError      =  1,
                ncclSystemError             =  2,
                ncclInternalError           =  3,
@@ -38,65 +40,86 @@ typedef enum { ncclSuccess                 =  0,
                ncclInvalidUsage            =  5,
                ncclNumResults              =  6 } ncclResult_t;
 
-/* Return the NCCL_VERSION_CODE of the NCCL library in the supplied integer.
- * This integer is coded with the MAJOR, MINOR and PATCH level of the
+
+  
+/*! @brief Return the NCCL_VERSION_CODE of the NCCL library in the supplied integer.
+ * 
+ * @details This integer is coded with the MAJOR, MINOR and PATCH level of the
  * NCCL library
  */
 ncclResult_t ncclGetVersion(int *version);
 ncclResult_t pncclGetVersion(int *version);
 
-/* Generates an Id to be used in ncclCommInitRank. ncclGetUniqueId should be
- * called once and the Id should be distributed to all ranks in the
- * communicator before calling ncclCommInitRank. */
+/*! @brief Generates an ID for ncclCommInitRank
+
+    @details
+    Generates an ID to be used in ncclCommInitRank. ncclGetUniqueId should be
+    called once and the Id should be distributed to all ranks in the
+    communicator before calling ncclCommInitRank. 
+    
+    @param[in]
+    uniqueId     ncclUniqueId*
+                 pointer to uniqueId
+
+*/
 ncclResult_t  ncclGetUniqueId(ncclUniqueId* uniqueId);
 ncclResult_t pncclGetUniqueId(ncclUniqueId* uniqueId);
 
-/* Creates a new communicator (multi thread/process version).
- * rank must be between 0 and nranks-1 and unique within a communicator clique.
- * Each rank is associated to a CUDA device, which has to be set before calling
- * ncclCommInitRank.
- * ncclCommInitRank implicitly syncronizes with other ranks, so it must be
- * called by different threads/processes or use ncclGroupStart/ncclGroupEnd. */
+/*! @brief Creates a new communicator (multi thread/process version).
+    
+    @details
+    rank must be between 0 and nranks-1 and unique within a communicator clique.
+    Each rank is associated to a CUDA device, which has to be set before calling
+    ncclCommInitRank.
+    ncclCommInitRank implicitly syncronizes with other ranks, so it must be
+    called by different threads/processes or use ncclGroupStart/ncclGroupEnd. 
+    
+    @param[in]
+    comm        ncclComm_t*
+                communicator struct pointer
+    */
 ncclResult_t  ncclCommInitRank(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank);
 ncclResult_t pncclCommInitRank(ncclComm_t* comm, int nranks, ncclUniqueId commId, int rank);
 
-/* Creates a clique of communicators (single process version).
- * This is a convenience function to create a single-process communicator clique.
+/*! @brief Creates a clique of communicators (single process version).
+ *
+ * @details This is a convenience function to create a single-process communicator clique.
  * Returns an array of ndev newly initialized communicators in comm.
  * comm should be pre-allocated with size at least ndev*sizeof(ncclComm_t).
- * If devlist is NULL, the first ndev CUDA devices are used.
- * Order of devlist defines user-order of processors within the communicator. */
+ * If devlist is NULL, the first ndev HIP devices are used.
+ * Order of devlist defines user-order of processors within the communicator. 
+ * */
 ncclResult_t  ncclCommInitAll(ncclComm_t* comm, int ndev, const int* devlist);
 ncclResult_t pncclCommInitAll(ncclComm_t* comm, int ndev, const int* devlist);
 
-/* Frees resources associated with communicator object. */
+/*! @brief Frees resources associated with communicator object. */
 ncclResult_t  ncclCommDestroy(ncclComm_t comm);
 ncclResult_t pncclCommDestroy(ncclComm_t comm);
 
-/* Returns a human-readable error message. */
+/*! @brief Returns a human-readable error message. */
 const char*  ncclGetErrorString(ncclResult_t result);
 const char* pncclGetErrorString(ncclResult_t result);
 
-/* Gets the number of ranks in the communicator clique. */
+/*! @brief Gets the number of ranks in the communicator clique. */
 ncclResult_t  ncclCommCount(const ncclComm_t comm, int* count);
 ncclResult_t pncclCommCount(const ncclComm_t comm, int* count);
 
-/* Returns the cuda device number associated with the communicator. */
+/*! @brief Returns the rocm device number associated with the communicator. */
 ncclResult_t  ncclCommCuDevice(const ncclComm_t comm, int* device);
 ncclResult_t pncclCommCuDevice(const ncclComm_t comm, int* device);
 
-/* Returns the user-ordered "rank" associated with the communicator. */
+/*! @brief Returns the user-ordered "rank" associated with the communicator. */
 ncclResult_t  ncclCommUserRank(const ncclComm_t comm, int* rank);
 ncclResult_t pncclCommUserRank(const ncclComm_t comm, int* rank);
 
-/* Reduction operation selector */
+/*! @brief Reduction operation selector */
 typedef enum { ncclSum        = 0,
                ncclProd       = 1,
                ncclMax        = 2,
                ncclMin        = 3,
                ncclNumOps     = 4 } ncclRedOp_t;
 
-/* Data types */
+/*! @brief Data types */
 typedef enum { ncclInt8       = 0, ncclChar       = 0,
                ncclUint8      = 1,
                ncclInt32      = 2, ncclInt        = 2,
@@ -114,17 +137,17 @@ typedef enum { ncclInt8       = 0, ncclChar       = 0,
  * Collective communication operations must be called separately for each
  * communicator in a communicator clique.
  *
- * They return when operations have been enqueued on the CUDA stream.
+ * They return when operations have been enqueued on the hipstream.
  *
  * Since they may perform inter-CPU synchronization, each call has to be done
  * from a different thread or process, or need to use Group Semantics (see
  * below).
  */
 
-/*
- * Reduce
+/*!
+ * @brief Reduce collective communication
  *
- * Reduces data arrays of length count in sendbuff into recvbuff using op
+ * @details Reduces data arrays of length count in sendbuff into recvbuff using op
  * operation.
  * recvbuff may be NULL on all calls except for root device.
  * root is the rank (not the CUDA device) where data will reside after the
@@ -137,10 +160,9 @@ ncclResult_t  ncclReduce(const void* sendbuff, void* recvbuff, size_t count, ncc
 ncclResult_t pncclReduce(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype,
     ncclRedOp_t op, int root, ncclComm_t comm, hipStream_t stream);
 
-/*
- * (deprecated) Broadcast (in-place)
+/*! @brief (deprecated) Broadcast (in-place)
  *
- * Copies count values from root to all other devices.
+ * @details Copies count values from root to all other devices.
  * root is the rank (not the CUDA device) where data resides before the
  * operation is started.
  *
@@ -151,11 +173,10 @@ ncclResult_t  ncclBcast(void* buff, size_t count, ncclDataType_t datatype, int r
 ncclResult_t pncclBcast(void* buff, size_t count, ncclDataType_t datatype, int root,
     ncclComm_t comm, hipStream_t stream);
 
-/*
- * Broadcast
+/*! @brief Broadcast
  *
- * Copies count values from root to all other devices.
- * root is the rank (not the CUDA device) where data resides before the
+ * @details Copies count values from root to all other devices.
+ * root is the rank (not the HIP device) where data resides before the
  * operation is started.
  *
  * In-place operation will happen if sendbuff == recvbuff.
@@ -165,10 +186,9 @@ ncclResult_t  ncclBroadcast(const void* sendbuff, void* recvbuff, size_t count,
 ncclResult_t pncclBroadcast(const void* sendbuff, void* recvbuff, size_t count, ncclDataType_t datatype, int root,
     ncclComm_t comm, hipStream_t stream);
 
-/*
- * All-Reduce
+/*! @brief All-Reduce
  *
- * Reduces data arrays of length count in sendbuff using op operation, and
+ * @details Reduces data arrays of length count in sendbuff using op operation, and
  * leaves identical copies of result on each recvbuff.
  *
  * In-place operation will happen if sendbuff == recvbuff.
@@ -178,10 +198,10 @@ ncclResult_t  ncclAllReduce(const void* sendbuff, void* recvbuff, size_t count,
 ncclResult_t pncclAllReduce(const void* sendbuff, void* recvbuff, size_t count,
     ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm, hipStream_t stream);
 
-/*
- * Reduce-Scatter
+/*!
+ * @brief Reduce-Scatter
  *
- * Reduces data in sendbuff using op operation and leaves reduced result
+ * @details Reduces data in sendbuff using op operation and leaves reduced result
  * scattered over the devices so that recvbuff on rank i will contain the i-th
  * block of the result.
  * Assumes sendcount is equal to nranks*recvcount, which means that sendbuff
@@ -196,10 +216,9 @@ ncclResult_t pncclReduceScatter(const void* sendbuff, void* recvbuff,
     size_t recvcount, ncclDataType_t datatype, ncclRedOp_t op, ncclComm_t comm,
     hipStream_t stream);
 
-/*
- * All-Gather
+/*! @brief All-Gather
  *
- * Each device gathers sendcount values from other GPUs into recvbuff,
+ * @details Each device gathers sendcount values from other GPUs into recvbuff,
  * receiving data from rank i at offset i*sendcount.
  * Assumes recvcount is equal to nranks*sendcount, which means that recvbuff
  * should have a size of at least nranks*sendcount elements.
@@ -229,18 +248,16 @@ ncclResult_t pncclAllGather(const void* sendbuff, void* recvbuff, size_t sendcou
  * of ncclGroupStart/ncclGroupEnd.
  */
 
-/*
- * Group Start
+/*! @brief Group Start
  *
- * Start a group call. All subsequent calls to NCCL may not block due to
+ * @details Start a group call. All subsequent calls to NCCL may not block due to
  * inter-CPU synchronization.
  */
 ncclResult_t ncclGroupStart();
 
-/*
- * Group End
+/*! @brief Group End
  *
- * End a group call. Wait for all calls since ncclGroupStart to complete
+ * @details End a group call. Wait for all calls since ncclGroupStart to complete
  * before returning.
  */
 ncclResult_t ncclGroupEnd();