Never mind. My fault. I was using -malloc_debug with the inode LU 
factorization

On Jun 7, 2011, at 10:16 AM, Barry Smith wrote:

> 
>    The flop rate for ILU numeric for AIJ matrices seems to have dropped 
> through the floor in petsc-dev. Any thoughts on what was changed there 
> recently?
> 
> 
>    Barry
> 


Reply via email to