Subclassing Engines (was Re: Moore's law and why persistence
may not be necessary. (fwd))
Bijan Parsia
bparsia at email.unc.edu
Wed Jan 23 20:25:37 UTC 2002
On Wed, 23 Jan 2002, Bijan Parsia wrote:
[snip]
> That seems high, after niling all those things and gcing:
>
> memory 32,311,120 bytes
> old 28,114,280 bytes (87.0%)
> young 124,272 bytes (0.4%)
> used 28,238,552 bytes (87.4%)
> free 4,072,568 bytes (12.6%)
[snip]
After indexing 93.7 meg messages & gc.
memory 58,577,984 bytes
old 51,063,332 bytes (87.2%)
young 39,320 bytes (0.1%)
used 51,102,652 bytes (87.2%)
free 7,475,332 bytes (12.8%)
self anyOf: #(bijan scott mark) finds 3641 messages in 131 milliseconds
(woohoo!)
self anyOf: #(from) finds 19868 messages (all of them, I think) in 13359
milliseconds (copying the set?? why copy the set? ah, something with the
unioning; yes)
I don't really want to run a full particpantHas: search :)
I'm trying to write out the index to disk using ReferenceStream along the
following lines:
|rr|
rr _ ReferenceStream fileNamed: 'EMAIL.fulltext'.
rr nextPut: self.
rr close.
At a 13 meg index this is going *painfully* slowly. I'm up to 143030
written out, though.
I suspect I should have just put message ids in the document
index. IndexFile already compactly serializes the index, and I *really*
only need the ids, I imagine. Might speed up the set copying/merging
too. Advice is welcome!
Cheers,
Bijan Parsia.
More information about the Squeak-dev
mailing list
|