core: workaround old gcc vec_mul{e,o} (Issue #15506)
authorPaul E. Murphy <pmur@users.noreply.github.com>
Thu, 12 Sep 2019 14:45:56 +0000 (09:45 -0500)
committerPaul E. Murphy <pmur@users.noreply.github.com>
Thu, 12 Sep 2019 14:54:02 +0000 (09:54 -0500)
commitb465c826963f9ea01d466f0281b568196e6ef933
treea4cbb4702d1b17dd5974cf3f92b454370ba9f5c0
parent8609198b050d568cdd19b4ed8c7dd80439791900
core: workaround old gcc vec_mul{e,o} (Issue #15506)

ISA 2.07 (aka POWER8) effectively extended the expanding multiply
operation to word types. The altivec intrinsics prior to gcc 8 did
not get the update.

Workaround this deficiency similar to other fixes.

This was exposed by commit 33fb253a66275abaa5060ef318c9a5cc87c5fd6e
which leverages the int -> dword expanding multiply.

This fixes Issue #15506
modules/core/include/opencv2/core/vsx_utils.hpp