index
:
debian-clblast
debian/sid
upstream/latest
Debian package for CLBlast.
gspr@nonempty.org
summary
refs
log
tree
commit
diff
log msg
author
committer
range
path:
root
/
include
/
internal
Age
Commit message (
Collapse
)
Author
2016-01-30
Added first auto-generated database headers from the Python database; only ↵
Cedric Nugteren
K40 and Iris supported now
2015-10-23
Added alpha and beta to tuner meta-data
CNugteren
2015-10-12
Routine names are now all default arguments defined in the header
CNugteren
2015-09-26
Added TRMV/TBMV/TPMV routines
CNugteren
2015-09-26
Made buffer copying a const-method for the source
CNugteren
2015-09-19
Added SBMV and SPMV routines
CNugteren
2015-09-19
Added the HPMV routine
CNugteren
2015-09-19
Added infrastructure for packed matrices
CNugteren
2015-09-19
Added the HBMV routine
CNugteren
2015-09-18
Improved the organization and performance of level 2 routines
CNugteren
2015-09-18
Added first version of banded matrix-vector multiplication
CNugteren
2015-09-14
Added xDOT/xDOTU/xDOTC dot-product routines
CNugteren
2015-09-14
Added extra temporary buffer to tuners in preparation of Xdot routines
CNugteren
2015-09-14
Added support for the dot buffer and offset argument
CNugteren
2015-08-22
Added the XSWAP, XSCAL and XCOPY level-1 routines
CNugteren
2015-08-19
Add check for supported precision to the tuners
CNugteren
2015-08-19
Moved precision tester to utilities
CNugteren
2015-08-19
Added precision to the JSON output
CNugteren
2015-08-13
Added argument m,n,k metadata to JSON files
CNugteren
2015-08-09
Refactored the tuners, added JSON output
CNugteren
2015-08-04
Added distinguished names for GEMV inherited HEMV/SYMV
CNugteren
2015-07-31
Added HEMV routine
CNugteren
2015-07-31
Added SYMV routine
CNugteren
2015-07-27
Now using the new Claduc C++11 OpenCL header
CNugteren
2015-07-22
Set the correct name for AMD OpenCL devices
CNugteren
2015-07-22
Updated GEMM tuning results for Tahiti
CNugteren
2015-07-22
Added workgroup shuffle option to transpose kernel for AMD GPUs
CNugteren
2015-07-19
Kernel caching is now based on a routine's name
CNugteren
2015-07-19
The kernel source string is now a routine's member variable
CNugteren
2015-07-19
Fixed complex performance on Intel Iris
CNugteren
2015-07-13
Updated interface of the PadCopyTransposeMatrix method
CNugteren
2015-07-12
Added subfolders for the level1/2/3 routines
CNugteren
2015-07-12
Added the HEMM routine, tester, and client
CNugteren
2015-07-10
Added the HER2K routine, tester, and client
CNugteren
2015-07-10
Added the HERK routine, tester, and client
CNugteren
2015-07-08
Added option to set the imaginary part of the diagonal to zero
CNugteren
2015-07-02
Added the TRMM routine, tester, and client
CNugteren
2015-07-01
Added the unit/non-unit diagonal enum
CNugteren
2015-06-28
Added buffer structure and sizes to arguments
CNugteren
2015-06-26
Added the SYR2K routine, tester, and client
CNugteren
2015-06-24
Added the SYRK routine, tester, and client
CNugteren
2015-06-23
Added a condition to update only lower/upper triangular parts in the un-pad ↵
CNugteren
kernels
2015-06-20
Automatically skips tests with unsupported precision
CNugteren
2015-06-20
Distinguish between a short smoke test and a full test
CNugteren
2015-06-20
Added additional absolute error checking when testing
CNugteren
2015-06-19
Added const-ref accessors to all CL++11 classes
CNugteren
2015-06-18
Now returns program from database by reference
CNugteren
2015-06-16
Added support for complex conjugate transpose
CNugteren
2015-06-15
Added tuning for DGEMV on Iris and SGEMV on K40m
CNugteren
2015-06-14
Split the three variations of the GEMV kernel for maximal tuning freedom
CNugteren
[next]