Is the way the create the large table crucial/relevant for the test?
currently is uses a MAL loop that performs
10000000 int to dbl conversions
10000000 mmath.sin() calls
10000000 bat.insert()
would a simple, say,
microbenchmark.uniform(0:oid,10000000,10000000)
to replace the entire MAL loop
also be OK for the purpose of this test?
Stefan
On Sat, Aug 27, 2011 at 12:02:26PM +0200, Martin Kersten wrote:
> In that case, trim down the size of the table being produced.
>
> The only reason for this test is to be able to see if parallelism
> processing works. This, however, can only be observed by either
> looking at top (-H) or using the stethoscope trace.
> For, the individual threads should have enough work to compete.
>
> As such, I would rather propose to remove the test from the batch
> and mark it for manual execution.
> (similar remarks hold for the datacell scenarios)
>
> A hardwired scheme to detect the level of parallelism could
> be considered for the future.
>
> On 8/27/11 11:40 AM, Stefan Manegold wrote:
> >Changeset: f4d001392870 for MonetDB
> >URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=f4d001392870
> >Modified Files:
> > monetdb5/optimizer/Tests/dataflow.timeout
> >Branch: Aug2011
> >Log Message:
> >
> >test runs for more than 3 minutes on our Sun T1000 machine
> >
> >
> >diffs (6 lines):
> >
> >diff --git a/monetdb5/optimizer/Tests/dataflow.timeout
> >b/monetdb5/optimizer/Tests/dataflow.timeout
> >--- a/monetdb5/optimizer/Tests/dataflow.timeout
> >+++ b/monetdb5/optimizer/Tests/dataflow.timeout
> >@@ -1,1 +1,1 @@
> >-2
> >+4
> >_______________________________________________
> >Checkin-list mailing list
> >[email protected]
> >http://mail.monetdb.org/mailman/listinfo/checkin-list
>
> _______________________________________________
> Checkin-list mailing list
> [email protected]
> http://mail.monetdb.org/mailman/listinfo/checkin-list
>
--
| Stefan.Manegold @ CWI.nl | DB Architectures (INS1) |
| http://CWI.nl/~manegold/ | Science Park 123 (L321) |
| Tel.: +31 (0)20 592-4212 | 1098 XG Amsterdam (NL) |
_______________________________________________
Checkin-list mailing list
[email protected]
http://mail.monetdb.org/mailman/listinfo/checkin-list
