Age | Commit message (Collapse) | Author |
|
|
|
imag is much larger than the other
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checks in the error checking to make the tests pass
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Refactor cache subsystem
|
|
|
|
This does not change much, but will become useful in next commits when
plugin support is introduced.
|
|
|
|
Implement a generalized Cache<K, V>. Two variants are provided: the
first one is based on std::map, using C++14-specific transparent
std::less<> and generalized std::map::find() to allow searching by tuple
of references. The second one is based on std::vector and O(n) lookup,
but remains C++11-compliant.
|
|
Assorted minor fixes
|
|
|
|
|
|
These errors are self-evident and enough to fail the test even if there is
no clBLAS reference to compare error codes with.
|
|
Thus do not prevent filling cache for float if the device does not support
e. g. double.
|
|
There was a missing return statement in the semi-warm path that made
CLBlast to continue to cold path after a cache hit.
|
|
This is to avoid spurious "induced" errors on destruction, if construction
failed for some reason.
|
|
The stored source/binary does not seem to serve any purpose, yet its
presence makes Program a heavy (not pure refcounted) object, which is
undesired esp. because it is copied from the cache in the hot path.
|
|
|
|
Otherwise, cl.h gets included through clblast.h before clpp11.hpp.
|
|
|
|
|
|
|
|
|
|
|
|
TRSM
|
|
|
|
|
|
OpenCL 1.1 or lower
|
|
|
|
|
|
|
|
|
|
|
|
executable and without re-running CMake
|
|
|
|
all devices
|
|
Netlib CBLAS API for CLBlast
|
|
|