Activity overview
Latest activity by dannytuppeny
Brian Donahue wrote:
I don't believe that this is the major reason for the code actually running more slowly
I've no problem with the code actually runnign slowler, that's expected and not a problem. The problem is that if the profiler doesn't accurately remove its overhead the results are quite skewed. The results make it look like a part of our application takes over 40 seconds (not good in a web page request!) when in reality this wasn't the case.
It's not a big deal, and we changed some things around that seemed to solve it. I've also been playing with the new version and I'll keep an eye on how things go with that one [image] / comments
Brian Donahue wrote:
I don't believe that this is the major reason for the code actually running more slowly
I've no problem with the code actually runnign slowler, that's expected and not a pr...
I don't understand the difference between "Wall Clock" and "CPU", but both seem to show the same numbers. / comments
I don't understand the difference between "Wall Clock" and "CPU", but both seem to show the same numbers.
Inaccurate timings
I posted a thread in the v3 forum about innacurate timings we had in our application. I stripped it down to a small sample posted here:http://www.red-gate.com/MessageBoard/vi ... php?t=7117The code...
I've just read a few posts about tight loops where the overhead calculation massively slows thigns down and inflates reported answers. Maybe that's what we're seeing? / comments
I've just read a few posts about tight loops where the overhead calculation massively slows thigns down and inflates reported answers. Maybe that's what we're seeing?
BUG: Profiler not correctly removing overhead from timings
I think I've found a bug in the profiler to do with it not removing its own overhead properly. The code below might seem a little strange, but it was the best example I could come up with. I think ...
Strange hit counts on if statements
I'm not sure if this is deliberate, but the hit counts on my if statements seem to be the sum of the number of times the condition was evaluated plus the number of times the condition evaluated to ...
That solved the problem! With our changes the process took 30 mins and the profiler says it was 14 mins - sounds resonable, since the last time I noticed the profiler was roughly doubling the exceution time :-) / comments
That solved the problem! With our changes the process took 30 mins and the profiler says it was 14 mins - sounds resonable, since the last time I noticed the profiler was roughly doubling the exceu...
Thanks. I'll install that update and see if it makes a difference [image] / comments
Thanks. I'll install that update and see if it makes a difference
Just a thought - it took 2.5 hours total - with the profiler attached. Since the profiler should be removing its own overhead, I assume the figures should be even lower, and it's out by even more [image] / comments
Just a thought - it took 2.5 hours total - with the profiler attached. Since the profiler should be removing its own overhead, I assume the figures should be even lower, and it's out by even more
Timings way off? 25 hours vs 2.5 hours
I've been profiling an application recently but the timings seem to be off by a factor of 10. The process takes 2.5 hours and the profiler reports 90,000 seconds (25 hours).
It's a multi-threaded a...