Re: Performance hit with --enable-atomic-refcount

[Michael Goffioul]

On Mon, Jun 9, 2014 at 11:38 AM, Daniel J Sebald <address@hidden> wrote:

On 06/06/2014 12:40 PM, Rik wrote:

6/6/14

All,

>From the previous benchmarking, there was a doubling in execution time from

3.8.1 to the gui-release branch for a nested for loop.  It turns out that
this doubling is 100% correlated with using the --enable-atomic-refcount
option to configure.  I built the gui-release branch with this feature
disabled and the results are then equivalent to the 3.8.1 release.  It
seems like we need to explore a different solution than this configure
option so that we can both use Qt graphics and have reasonable performance.

The attomic reference count is, from what I understand, needed for the QtHandles code, which implements script callbacks (I say "script callbacks" to differentiate from source code callbacks).  Atomic reference isn't needed for Qt graphics without that.

No, atomic reference is a way to implement seamless concurrent access to octave_value objects (and other ref-counted objects), between the UI event loop and the octave thread.

 

However, I believe the atomic reference count can be eliminated with a well-organized bi-directional interface between UI/worker threads rather than using source-code callbacks.  This could make a good topic for OctConf '12.

It's a bit late for OctConf'12... ;)

Michael.