> >>
> >>
> >> Jerome Nilmeier, PhD
> >> Data Scientist and Engineer
> >> IBM Spark Technology Center
> >> http://www.spark.tc/
> >>
> >>
> >>
> >> ----- Original message -
> >> From: Matthias Boehm <mboe...@g
mail.com>
> To: dev@systemml.apache.org
> Cc: Qifan Pu <qifan...@gmail.com>, Jeremy Nilmeier <nilme...@us.ibm.com>
> Subject: Re: distributed cholesky on systemml
> Date: Sun, Apr 22, 2018 2:41 PM
>
> thanks for the context Jeremy - that helps. I also
com>
>> To: dev@systemml.apache.org
>> Cc: Qifan Pu <qifan...@gmail.com>, Jeremy Nilmeier <nilme...@us.ibm.com>
>> Subject: Re: distributed cholesky on systemml
>> Date: Sun, Apr 22, 2018 2:41 PM
>>
>> thanks for the context Jeremy - that helps. I al
with Sasha about this some time last year. Thanks
>> >>> for
>> >>> following up.
>> >>>
>> >>> Cheers, J
>> >>>
>> >>>
>> >>> Jerome Nilmeier, PhD
>> >>> Data Scientist and En
sure no problem - thanks again for catching this issue that was hidden
for a while.
Yes, the same depth-first characteristic applies to the Cholesky
function as well. In contrast to U_triangular_inv, however, there are
data dependencies between the blocks per level (at least in the
current
