diff options
| author |  Mark D Ryan <mark.d.ryan@intel.com> | 2019-01-11 14:02:09 +0100 |
|---|---|---|
| committer | Mark D Ryan <mark.d.ryan@intel.com> | 2019-01-11 14:02:09 +0100 |
| commit | 3c9add6598cc35e5317788627dfa81f517e89e07 (patch) | |
| tree | 71b030f5c999821a6d73c693cc3bcbd29c94a040 /lapack/slamch.f | |
| parent | 0522460a0d01d4253183349a49144b5ad8ba2f9f (diff) | |
Remove reinterpret_cast from AVX512 complex implementation
The reinterpret_casts used in ptranspose(PacketBlock<Packet8cf,4>&)
ptranspose(PacketBlock<Packet8cf,8>&) don't appear to be working
correctly. They're used to convert the kernel parameters to
PacketBlock<Packet8d,T>& so that the complex number versions of
ptranspose can be written using the existing double implementations.
Unfortunately, they don't seem to work and are responsible for 9 unit
test failures in the AVX512 build of tensorflow master. This commit
fixes the issue by manually initialising PacketBlock<Packet8d,T>
variables with the contents of the kernel parameter before calling
the double version of ptranspose, and then copying the resulting
values back into the kernel parameter before returning.
Diffstat (limited to 'lapack/slamch.f')
0 files changed, 0 insertions, 0 deletions