Just a couple of more tests...
Test file:
2.7MB IRC log, 11500 table rows
Netscape 4.6:
Computer: P-II 266, clocked to 291MHz, 96MB RAM, Solaris 7 x86, about 930kkeys RC5
Result: Reading and parsing took 330 seconds (5 min 30 sec)
Computer: P166, 46MB RAM, about 230kkeys/s RC5, billdoze 95
Result: Reading and parsing took 630 seconds (10 min 30 sec)
Internet exploiter 5:
Computer: P166, 48MB RAM, about 230kkeys/s RC5, billdoze 95
Result: Reading and parsing took 120 seconds (2 min)
Voyager 3p2:
Computer: Amiga 4000, 060-50 (604e 200), 146MB RAM, about 115kkeys/s RC5
Result: Reading and parsign took 50 seconds (just 50 sec)
AWeb-II 3.2 Demo:
Computer: Amiga 4000, 060-50 (604e 200), 146MB RAM, about 115kkeys/s RC5
Result: Reading and parsing took 330 seconds (5 min 30 sec)
Notes: The only of the tested browsers that allowed reading of
the so-far parsed file almost all the time during parsing.
The result got not so good because AWeb paused every now and then
to reformat the table when something didn't fit in a cell (total
time for reformatting was about 3 minutes)
Voyager required 6.5 times less time parsing the table than netscape.
The machine running netscape was about 8 times faster on the RC5 benchmark.
Calculating some kind of "speed index"
P-II 291 ns 4.6: 330s, 930kkeys = 306900
P 166 ns 4.6: 630s, 230kkeys = 144900
P 166 ie 5: 120s, 230kkeys = 27600
A4k060-50 v3p2: 50s, 115kkeys = 5750
A4k060-50 aw3.2: 330s, 115kkeys = 37950
Lower value = better
So... v3p2 is 53 times faster than ns 4.6 on Solaris (P-II 291), 25 times faster than
ns 4.6
on billdoze (P 166), about 5 times faster than ie 5 on billdoze (P 166), and about 6.5
times
faster than AWeb 3.2 parsing tables.
/PeO
____________________________________________________________
Voyager Mailing List - Info & Archive: http://www.vapor.com/
For Listserver Help: <[EMAIL PROTECTED]>, "HELP"
To Unsubscribe: <[EMAIL PROTECTED]>, "UNSUBSCRIBE"
