James,

I have installed the latest beta v19 release. I have not encountered any issues with this latest release..but I am seeing a definite capture performance degradation when compared to earlier betas.

You mentioned that the v19 beta release had some debugging code removed...were there other changes as well? The performance I am seeing, at least thus far, seems to be about what I was seeing in the latest production release. When I started using the beta I observed huge capture performance gains...this appears to be reduced in V19.

Thanks,

GKG

Gary K. Griffey wrote:You mentioned that the v19 beta release had some debugging code removed...were there other changes as well?

Nothing significant, beyond the changes to how concurrent operations are handled. I would not expect that to significantly affect capture performance.

The performance I am seeing, at least thus far, seems to be about what I was seeing in the latest production release. When I started using the beta I observed huge capture performance gains...this appears to be reduced in V19.

That's unexpected. It could be QRecall or it could be OS X 10.6.5. One thing I've noticed with 10.6.5 is processes spawning an inordinate number of threads. Even after the "fix" in b18 to address this, processes still run with considerably more threads than they did in 10.6.4. Yet while I find this perplexing, I still wouldn't expect additional threads to seriously degrade capture performance.

I'll investigate this here. One thing you can do is gather some samples. You, and anyone else reading this thread, are welcome to send a diagnostic report after running b19 for a week or so. Say something like "b19 performance" in the comments. I can then write a program to extract the performance statistics from the logs for the past couple of months and see if there's a correlation between the upgrade and a drop in capture speeds. I'd be particularly interested in getting a report from anyone who's upgraded to QRecall 1.2.0b19 but is still running OS X 10.6.4.

James,

Thanks for the response...I like your plan...I will let v19 run for a week or so on my machines as it did before...and then go back and compare with previous beta timings..

Thanks,

GKG

James,

Just a quick update on this issue.

After comparing results of the current V19 beta vs. previous beta's..the performance appears to be roughly the same...i.e., much faster than the current 1.1.4 "gold" code....even on 10.6.5...

Sorry for raising a red flag on this...all seems to be working well in V19...

Thanks,

GKG