Hello,
Hello Mindagaus, All
My name is Mindaugas.
I have finished porting some parts of my appln's
report-generation of routines to take advantage of
MEMIO instead of disk files and have found the following.
One of reports creates a table, say, rpa80025.dbf and
associated index, ttt14071.z01. The numbers part is pulled
with hb_random to ensure that file name is unique in every
run, regardless of thread or process creating it.
The tables are created properly, and at the end, dbDrop()
reports true for both .dbf and .z01.
I run the same report various times and got these facts
RUN# TimeToFinish Mem Usage
(excluding ( Task
display ) Manager )
1 906
2 859
3 875
4 906
5 937
6 953
7 985
8 1016
9 1032
10 1062
11 1078
12 1109
13 1141
14 1156
15 1203
16 1203
17 1250 14548
18 1266 14568
19 1282 14568
20 1312 14576
21 1329 14588
22 1375 14604
23 1390 14608
24 1422 14612
Is it a normal behavior ?
Or should I put more rigorous checks
to measure the performance ?
Or something is still required in MemIO implementation ?
Or may be I am missing something completely ?
Can you try to make self contained sample by changing real report to
some simple DBAPPEND(), FIELDPUT()? I can not reproduce such bahaviour.
Regards,
Mindaugas
_______________________________________________
Harbour mailing list
Harbour@harbour-project.org
http://lists.harbour-project.org/mailman/listinfo/harbour
