Subject: Craig Chambers: Re: a query concerning dispatch tree statistics
From: Craig Chambers (chambers@cs.washington.edu)
Date: Mon Jan 24 2000 - 08:17:51 PST
------- Forwarded Message
To: mnaik@hss.hns.com
Subject: Re: a query concerning dispatch tree statistics
In-reply-to: Your message of Tue, 18 Jan 2000 15:57:56 +0530.
<6525686A.00399393.00@sandesh.hss.hns.com>
Date: Mon, 24 Jan 2000 16:17:05 GMT
From: Craig Chambers <chambers@klamath.cs.washington.edu>
I can only observe that the [] tests were rarely used, and this didn't
affect the average dispatch time nor the total space cost. I looked
over the raw data files to confirm this. It's possible that there's a
problem in my data collection in the face of [] tests, but I can't see
it now.
- -- Craig Chambers
On Tue, 18 Jan 2000 15:57:56 +0530, mnaik@hss.hns.com wrote:
>
>
> Hello
>
> This is with reference to 'Table 3: Dispatch Tree Statistics' in the
> paper Efficient Multiple and Predicate Dispatching.
>
> The values of 'size' and 'avg. dispatch time' are *exactly* the same
> for dispatching using (=, <, []) and using (=, <). Could you please
> explain that?
>
> Thanks.
>
> -- Mayur
>
>
------- End of Forwarded Message
This archive was generated by hypermail 2b25 : Tue Oct 03 2000 - 15:21:21 PDT