ReneEnjilian opened a new pull request, #2022:
URL: https://github.com/apache/systemds/pull/2022
I noticed that the fallback implementation for sparse-dense matrix multiply
produces wrong results or sometimes throws an index out of bounds exception.
This patch fixes this issue. Further, this part of the code is only reached
when LOW_LEVEL_OPTIMIZATION is set to false (by default it is true). Hence,
this is bug should generally not occur.
**Example:**
A %*% B = C , where A is in sparse format, B and C are dense.
A = {{1, 0, 2, 0},
{0, 1, 0, 0},
{0, 0, 3, 0},
{0, 5, 0, 0}}
B = {{1, 2, 3, 4},
{5, 6, 7, 8},
{9, 10, 11, 12},
{13, 14, 15, 16}}
For A utilizing CSR, it throws and out of bound exception and when using
MCSR it produced a wrong result. I will check if this or similar errors occur
in other matrix multiply operators (specifically those called when
LOW_LEVEL_OPTIMIZATION is true).
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: dev-unsubscr...@systemds.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
