[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: profiling, what takes so long?
From: |
Dave Hylands |
Subject: |
Re: profiling, what takes so long? |
Date: |
Thu, 2 Feb 2006 10:11:37 -0800 |
> Dynamic views (particularly if read/write) are the whole point of using
> Clearcase in the first place, they are what make it powerful. However in
> exchange for that power you have terrible performance. Each stat() requires
> a lookup, over a network link, on the CC server. Snapshot views are just
> like souped-up CVS.
Well, I use snapshot views for all of my day to day building, nightly
builds and whatnot. It's faster to populate a snapshot and build from
it than it is to build from a dynamic view.
I use the dynamic views for merging between streams, where I really
want the power of ClearCase. The real secret is that you can have both
a dynamic view and a snapshot view both using the same stream.
We've changed source code control systems several times, and still
occaisonally use a different one, so I'd much rather use GNU make,
which is agnostic to the source code control system being used (that's
my on-topic blurb).
--
Dave Hylands
Vancouver, BC, Canada
http://www.DaveHylands.com/
- profiling, what takes so long?, Torsten Mohr, 2006/02/01
- Re: profiling, what takes so long?, Brendan Heading, 2006/02/01
- Re: profiling, what takes so long?, Paul D. Smith, 2006/02/01
- Re: profiling, what takes so long?, Dan Kegel, 2006/02/01
- Re: profiling, what takes so long?, Johan Bezem, 2006/02/01
- Re: profiling, what takes so long?, Torsten Mohr, 2006/02/04