77458 - Cache is flaky, a lot of the time slower

Status: VERIFIED WONTFIX

(Core :: Networking: Cache, defect, P3)

Description

[Mike Pinkerton (not reading bugmail)]

Look at the attached page runs on jrgm's tests, the second through fifth 
runs are generally slower than the initial page load. This smells like the 
cache isn't really working all that well, at least on mac.

Comment 1

[Mike Pinkerton (not reading bugmail)]

Attachment: page load times from jrgm's tests

Comment 2

[gordon]

This may get better with the landing of the DISKCACHE2_BRANCH, but if not, I 
won't have time to investigate it more thoroughly until 0.9.2.

Comment 3

[Ronald van Kuijk]

Don't know if it's related but a shift-reload is even worse (see bug 83721).
It's 2 to 4 times slower than an initial load. The more images in a page the
worse it seems.

Comment 4

[gordon]

No, this is not related to shift-reload.  There seems to be a large degree of 
randomness in the test results that has not been explained, though the overall 
averages seem reliable.

Comment 5

[Gagan]

gordon sez he won't get to it before 0.9.3. off to 0.9.3

Comment 6

[gordon]

I'm continuing to investigate the performance results.

Comment 7

[gordon]

I want to take a look at this after fixing bug 81724, which should boost
performance a bit.

Comment 8

[gordon]

Setting dependency on bug 81724, which has been moved to mozilla1.0.

Comment 9

[scottputterman]

Moving Netscape owned 0.9.9 and 1.0 bugs that don't have an nsbeta1, nsbeta1+,
topembed, topembed+, Mozilla0.9.9+ or Mozilla1.0+ keyword.  Please send any
questions or feedback about this to [e-mail].  You can search for
"Moving bugs not scheduled for a project" to quickly delete this bugmail.

Comment 10

[S Woodside]

Is this still a problem?

Comment 11

[gordon]

We need to re-run the performance tests to determine if it's still a problem.  I
wouldn't be surprised, but it may involve underlying limitations of the file
system.  More testing and analysis is needed, so I'm leaving this bug open for now.

Comment 12

[Simon Paquet [:sipaq]]

This bug is targeted at a Mac classic platform/OS, which is no longer supported
by mozilla.org. Please re-target it to another platform/OS if this bug applies
there as well or resolve this bug.

I will resolve this bug as WONTFIX in four weeks if no action has been taken.
To filter this and similar messages out, please filter for "mac_cla_reorg".

Comment 13

[benc]

-> defaults.

Simon, do you have a test case for this problem?

Disk I/O is so different between the two flavors of Mac OS...

Comment 14

[Darin Fisher]

huh?  so, we have a bug about poor cache performance of the "phase 1" cache
rewrite on OS/9 (read CFM build) that is over 2 years old... and now, we just
transfer this bug over to OSX where we use a completely different i/o system? 
how does that make any sense?

WONTFIX because CFM build is dead.

Comment 15

[benc]

V.