summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2017-10-20Added tuning parameters for GeForce GTX 580, GeForce GTX 1080Ti, and Core ↵Cedric Nugteren
i5-4570
2017-10-18Moved CUmodule code from Kernel to Program class to not require ↵Cedric Nugteren
re-compilation every time
2017-10-17Fix an incompatibility with CUDA's FP16 definitionCedric Nugteren
2017-10-17Made buffers of batched routines read/write (was: read-only)Cedric Nugteren
2017-10-17CUDA kernel compilation fixesCedric Nugteren
2017-10-16Made all CUDA kernel launches synchronous; removed exception raisingCedric Nugteren
2017-10-15Added a missing OpenCL-to-CUDA function translationCedric Nugteren
2017-10-15Fixes for the CUDA API: first tests pass and the client runsCedric Nugteren
2017-10-15Added the SM-compute-arch version to the nv compile optionsCedric Nugteren
2017-10-15Various fixes to make the first CUDA examples workCedric Nugteren
2017-10-14Fixed a kernel/attribute order bug in the direct GEMM kernelsCedric Nugteren
2017-10-14Make local memory pointers a define in OpenCL; some fixes to the recently ↵Cedric Nugteren
changed transpose kernel code
2017-10-14Made transpose kernel struct init proper according to the C standardCedric Nugteren
2017-10-14Fixed several (not all) CUDA kernel compilation issuesCedric Nugteren
2017-10-14Various fixes to make the host code and sample compile with the CUDA APICedric Nugteren
2017-10-14Added OpenCL to CUDA translation header for the kernelsCedric Nugteren
2017-10-12CUDA API now takes context and device in instead of streamCedric Nugteren
2017-10-11Added first (untested) version of a CUDA APICedric Nugteren
2017-10-09Removed include of clpp11.hpp in places other than utilities.hppCedric Nugteren
2017-10-08Moved non-routine-specific API functions and includes to separate filesCedric Nugteren
2017-10-08Moved the remaining OpenCL specific host code to the clpp11.h header where ↵Cedric Nugteren
it belongs
2017-10-07Synchronizes clpp11.h with CLCudaAPI 9.0Cedric Nugteren
2017-10-07Fixed a small typoCedric Nugteren
2017-10-03Gemm in-direct implementation now uses only 1 larger instead of max 3 ↵Cedric Nugteren
optional temporary buffers
2017-10-01Allow OverrideParameters function to work before a kernel was first usedCedric Nugteren
2017-09-30Merge branch 'additional_tuners'Cedric Nugteren
2017-09-30Refactored the tuning architecture: less duplicate now; more defaultsCedric Nugteren
2017-09-24Updated database override function to work with the new database storage formatCedric Nugteren
2017-09-23Made program and binary databases dependent on the routine parameters on top ↵Cedric Nugteren
of the name
2017-09-23Made database-caching no longer dependent on device name but on ↵Cedric Nugteren
device/platform IDs
2017-09-19Fixed type conversion warnings under MSVC 2013Cedric Nugteren
2017-09-19Fixed compilation issues of the database for MSVC 2013Cedric Nugteren
2017-09-16Updated README with proper AMD device names; fixed device look-up for names ↵Cedric Nugteren
of length 50+
2017-09-16Added tuning results for Intel Core i7 6770HQCedric Nugteren
2017-09-16Fixed a compilation error and warning under MacOSCedric Nugteren
2017-09-16Fixed an issue with the NVIDIA compute capability not being retrieved properlyCedric Nugteren
2017-09-16Improved compilation time of the tuner databaseCedric Nugteren
2017-09-14Added a guard against missing AMD and NVIDIA extensionsCedric Nugteren
2017-09-14Added architecture layer in the tuning database for better performance on ↵Cedric Nugteren
unseen devices
2017-09-10Added the new vendor-architecture-name hierarchy to the tuners as wellCedric Nugteren
2017-09-08Introduced the notion of a device-architecture for the database and added ↵Cedric Nugteren
device and architecture name mappings
2017-09-06Split the database files over multiple directories and files; first step ↵Cedric Nugteren
towards separate compilation
2017-09-05Fixed a modulo and division issue manifesting on Apple OpenCL for im2colCedric Nugteren
2017-09-04Removed an assumption that the 'default' tuning parameters have to be stored ↵Cedric Nugteren
last; this is no longer needed
2017-08-31Fixed a bug in im2col: process only valid channel IDsCedric Nugteren
2017-08-31Fixed a bug in im2col confusing first and second workgroup size; made im2col ↵Cedric Nugteren
kernel 2d instead of 3d
2017-08-31Fixed some things in the tuner: bugs, style, and defaults to random searchCedric Nugteren
2017-08-24Merge branch 'master' into im_to_colCedric Nugteren
2017-08-24Completed im2col implementationCedric Nugteren
2017-08-23Made the im2col client properly handle the argumentsCedric Nugteren
generated by cgit v1.2.3 (git 2.39.1) at 2024-08-03 08:50:40 +0000