tree 8f15d4672342135f56aa20bd50e3f828ad16fe62
parent fdb42dafc0e2010d0b021cedd4fb1ca6f42939ab
author bjacob <benoitjacob@google.com> 1662046563 -0400
committer GitHub <noreply@github.com> 1662046563 +0000
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJjENFjCRBK7hj4Ov3rIwAAcFoIAFHE7opyY3I8Rpv/nMdhZexA
 dtV4ZPQFuoTqiyYnadsV7xZMgc4ThmKxN1pJobyi9DDM7dxo8i0ol/MegF6vBjRB
 RQJY5LmyWDtOJllbIoQeSseSs20et7b4IIjehJTAcpZVt2uumn7n1Zoszjhw/FkE
 qBwgR2zDGoj+965ejbJ5oigzENmQauzeit/5WbJq6gJozA9YFnv6tMzzpT6t5oSs
 b/HvwxZcgefRXzzyOpB5BIOdDtTtq6mYaVOPeUPkQ8P3o2jg4zaYvrKYOpj9QxDH
 5aU2xZKdVhrg7xNb59rvLMzWDCNeuJb9Ykdo18PLrxexZDgnBF1CAV91AEDL/Pg=
 =LBWp
 -----END PGP SIGNATURE-----
 

Fix mapping of buffers in mmt4d module functions. (#10269)

When mapping the 4D buffers using the 2D buffer-mapping macro, I needed to multiply the 3 inner dims sizes together to form the coalesced inner dim of the 2D shape.

Rationalizing this also allowed to write a better explanatory comment: not longer had to make excuses for an "abuse" of the 2D macro.