Hi all,
I have a question about the results of the profiler.
When I profile a simple program like:
DEFINE VARIABLE i AS INTEGER NO-UNDO.
DO WHILE i < 10000:
ASSIGN i = i + 1.
END.
on my windows client i get the following results:
1 08/20/2010 "test" 15:08:05 "VlasblomK"
.
1 "P:\Profiler\Config\p01396_Untitled7.ped" "" 63908
.
0 0 1 1
.
0 0 1 0.000000 0.027474
1 9 10001 0.005640 0.005640
1 11 10000 0.003336 0.003336
1 10 10000 0.005382 0.005382
.
.
.
.
But when I run the same program on the Linux server I get the following results:
1 08/23/2010 "test" 10:11:16 ""
.
1 "/tmp/p20028_Untitled1.ped" "" 43980
.
0 0 1 1
.
0 0 1 0.000000 140733193388032.026738
1 9 10001 0.006129 140733193388032.006129
1 11 10000 0.002513 140733193388032.002513
1 10 10000 0.005279 140733193388032.005279
.
.
As you can see the cumulative times are strange.
Can anyone explain why this is happening and does this number 140733193388032 depends on some config setting or something? Can I "fix" the results just by adjusting the results with this number?
regards
Kees Vlasblom
proc.mon file has also the same values sometimes.
I think it is oe bug when messured time is calculated as negative.
Ordinary I noticed this on heavy loaded systems.
I just ignore this statictics and look for other examples when the same procedure call occurs.
Maxim.
It sounds like you may be encountering a resolved issue. In 10.2A03 and 10.2B01, OpenEdge development resovled an issue on 64-bit Unix systems, in which the time value was displayed incorrectly in the Profiler.
Hi Evan,
Sounds like that's the issue, server is indeed on 64-bit and the client is 32-bit.
But upgrading just for this issue is not an option. I will try to adjust the numbers when importing the data in the DB.
Thanks
Kees