Activity overview
Latest activity by Chris.Allen
Profiling stopped
Its not uncommon for profiling to exacerbate issues like this. There's not a lot that can be done except to try a different mode of profiling or to fully debug the whole process.
Time has shown that errors likw this appear to be an issue caused by a bug in the .NET profiling interface. Some of the int32 arrays being allocated are not being reported by .NET as objects to the profiler, although the objects that reference them are (and are apparently still valid as objects and not candidates for garbage collection).
What we have found that works is disabling the server garbage collector as described here: <http://support.microsoft.com/kb/911716>
Remember- "server GC" actually means multi-core GC.
If that doesn't work, mail us at support@red-gate.com for an update which works better in some cases like these. / comments
Time has shown that errors likw this appear to be an issue caused by a bug in the .NET profiling interface. Some of the int32 arrays being allocated are not being reported by .NET as objects to the...
We've acknowledged this as a UI bug (PP-895) and will hopefully enhance the presentation of such data in the next version. / comments
We've acknowledged this as a UI bug (PP-895) and will hopefully enhance the presentation of such data in the next version.
Yes- I see what you mean, when represented in the methods view grod, its only claiming 3.6 milliseconds but in the source code view its 3.6 seconds. This is rather odd, is it easily reproducible? / comments
Yes- I see what you mean, when represented in the methods view grod, its only claiming 3.6 milliseconds but in the source code view its 3.6 seconds. This is rather odd, is it easily reproducible?
Can you post a screenshot of what the results look like in the source code view please (wth line-level timings). / comments
Can you post a screenshot of what the results look like in the source code view please (wth line-level timings).
Just a very quick check- are you using "line level" mode of profiling (most detail)?
(I think the problem is actually deeper than this so I'll check with the developer but I just thought I'd ask this one :-) / comments
Just a very quick check- are you using "line level" mode of profiling (most detail)?
(I think the problem is actually deeper than this so I'll check with the developer but I just thought I'd ask th...
No. Sorry. But teh application has been designed throughout to be as efficient and simple to use as possible so the way we would recommend you profile is by using "CPU Time", not "Wall clock" time- then you shouldn't get any high-latency method shsw up significantly in your results. I hope this helps- let me know if I misunderstood... / comments
No. Sorry. But teh application has been designed throughout to be as efficient and simple to use as possible so the way we would recommend you profile is by using "CPU Time", not "Wall clock" time-...
Thanks for finding the solution. I will suggest this enhancement to our guys. / comments
Thanks for finding the solution. I will suggest this enhancement to our guys.
There's no indication of a problem and every reason to think the profiler is reliable. Counters are only updated by the CLR on a garbage collect. A GC happens on every snapshot. / comments
There's no indication of a problem and every reason to think the profiler is reliable. Counters are only updated by the CLR on a garbage collect. A GC happens on every snapshot.
Hmm, Catastrophic failure from the add-in.
Can you try replacing the existing addin dll with this please: ftp://support.red-gate.com/Patches/ANTS ... .AddIn.dll / comments
Hmm, Catastrophic failure from the add-in.
Can you try replacing the existing addin dll with this please:ftp://support.red-gate.com/Patches/ANTS ... .AddIn.dll