added requirements and preparations for version 0.1.0

CHANGELOG.md

@@ -3,14 +3,24 @@ All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
 ## [Unreleased]
+
+
+
+
+## [0.1.0] - 2016-11-02
 ### Changed
 - verbose now also prints debug output when correctness check fails
+- restructured the utility functions into util and core
 - restructured the code to prepare for different strategies
 - shortened the output printed by the tune_kernel
 - allowing numpy integers for specifying problem size
 
 ### Added
+- a public roadmap
+- requirements.txt
+- example showing GPU code unit testing with the Kernel Tuner
 - support for passing a (list of) filenames instead of kernel string
+- runner that takes a random sample of 10 percent
 - support for OpenCL platform selection
 - support for using tuning parameter names in the problem size
 

doc/source/conf.py

@@ -57,9 +57,9 @@
 # built documents.
 #
 # The short X.Y version.
-version = u'0.0.1'
+version = u'0.1.0'
 # The full version, including alpha/beta/rc tags.
-release = u'0.0.1'
+release = u'0.1.0'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

requirements.txt

@@ -0,0 +1,6 @@
+mock>=2.0.0
+nose>=1.3.7
+numpy>=1.7.1
+pycuda>=2016.1.1
+pyopencl>=2015.2.4
+

roadmap.md

@@ -0,0 +1,38 @@
+# Roadmap for the Kernel Tuner
+
+This roadmap presents an overview of the features we are currently planning to
+implement. Please note that this is a living document that will evolve as
+priorities grow and shift.
+
+### version 0.2.0
+
+This is the list of features that we want to have implemented by the next version.
+
+ * Option to store tuning results in a file (e.g. json, csv, ... )
+ * Option to set a function that performs output verfication, instead of numpy.allclose()
+ * Option to change defaults for 'block_size_x', and so on
+ * Option to set a function that computes search space restriction, instead of a list of strings
+ * Option to set compiler name, when using C backend
+ * Option to set compiler options
+
+### version 1.0.0
+
+These functions are to be implemented by version 1.0.0, but may already be
+implemented in earlier versions.
+
+ * Tuning kernels in parallel on a single node
+ * Tuning kernels in parallel on a set of nodes in a GPU clusters
+ * Tuning kernels using machine learning or search strategies
+ * Store tuning results in a database and provide an API for analysis
+
+### Low priority
+
+These are the things that we would like to implement, but we currently have no
+demand for it. If you are interesting in any of these, let us know!
+
+ * Tuning compiler options in combination other parameters kernel
+ * Example that tunes a kernel using thread block re-indexing
+ * Example host code that runs a pipeline of kernels
+ * Example CUDA host code that uses runtime compilation
+
+

setup.py

@@ -1,17 +1,20 @@
 import os
 from setuptools import setup
 
+with open('requirements.txt') as f:
+    required = f.read().splitlines()
+
 def read(fname):
     return open(os.path.join(os.path.dirname(__file__), fname)).read()
 
 setup(
     name = "kernel_tuner",
-    version = "0.0.1",
+    version = "0.1.0",
     author = "Ben van Werkhoven",
     author_email = "b.vanwerkhoven@esciencecenter.nl",
-    description = ("A simple CUDA kernel tuner in Python"),
+    description = ("A simple CUDA/OpenCL kernel tuner in Python"),
     license = "Apache 2.0",
-    keywords = "auto-tuning gpu pycuda cuda pyopencl opencl",
+    keywords = "auto-tuning gpu computing pycuda cuda pyopencl opencl",
     url = "http://benvanwerkhoven.github.io/kernel_tuner/",
     packages=['kernel_tuner', 'kernel_tuner.runners'],
     long_description=read('README.md'),
@@ -30,5 +33,7 @@ def read(fname):
         'Topic :: System :: Distributed Computing',
         'Development Status :: 4 - Beta',
     ],
+    install_requires=required,
+
 )