Reduce the HistoryDatabase's page cache from 6000 (24MB) to 1000 (4MB) pages.

Reduce the HistoryDatabase's page cache from 6000 (24MB) to 1000 (4MB) pages.

Shrinking the history databases page cache doesn't seem to impact performance
of the QueryHistory time metric, or the page cycler.

BUG=243769
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=208036

[rmcilroy, 2013-06-19 11:03:35]

I would like to propose reducing the size of the history database's cache from
24MB to 4MB.  I did some tests with a large history database (my own Chrome
profile, of which this "History" file is about 10MB on disk) and reducing the
cache size didn't seem to affect page cycler performance (even reducing to about
500K didn't make much of an impact).  The results are at:
https://docs.google.com/a/google.com/spreadsheet/ccc?key=0As0cjoYvJ58sdGtlVFB...

There also didn't seem to be much effect on the History.QueryHistory time
histogram when entering some test queries.

Please let me know if there are any other experiment you think would be useful
to validate that this change doesn't impact performance.

[Scott Hess - ex-Googler, 2013-06-19 17:32:14]

On 2013/06/19 11:03:35, Ross Mcilroy wrote:
> I would like to propose reducing the size of the history database's cache from
> 24MB to 4MB.  I did some tests with a large history database (my own Chrome
> profile, of which this "History" file is about 10MB on disk) and reducing the
> cache size didn't seem to affect page cycler performance (even reducing to
about
> 500K didn't make much of an impact).  The results are at:
>
https://docs.google.com/a/google.com/spreadsheet/ccc?key=0As0cjoYvJ58sdGtlVFB...
> 
> There also didn't seem to be much effect on the History.QueryHistory time
> histogram when entering some test queries.
> 
> Please let me know if there are any other experiment you think would be useful
> to validate that this change doesn't impact performance.

The only thing I think this would affect would be big index segment merges for
the fts table.  In the worst case it will manually swap to the journal file,
which would suck.  But it's effectively a merge sort, so it wouldn't be the
worst case worst case, and I'd be somewhat surprised if anyone ever hits the bit
ones anyhow, and since you're getting rid of that feature it wouldn't even come
up...

[Scott Hess - ex-Googler, 2013-06-19 17:32:26]

Oh, and LGTM.

[commit-bot: I haz the power, 2013-06-20 09:30:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[rmcilroy, 2013-06-20 09:35:48]

On 2013/06/20 09:30:55, I haz the power (commit-bot) wrote:
> CQ is trying da patch. Follow status at
> https://chromium-status.appspot.com/cq/rmcilroy%40chromium.org/17416002/1

Brett, could you add your LGTM to this please?  Needs lgtm from an owner of the
history service.

[commit-bot: I haz the power, 2013-06-20 10:48:02]

Retried try job too often on chromium_presubmit for step(s) presubmit
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=chromium_p...

[brettw, 2013-06-21 22:59:39]

lgtm

[commit-bot: I haz the power, 2013-06-22 01:46:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[commit-bot: I haz the power, 2013-06-22 02:42:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[commit-bot: I haz the power, 2013-06-22 03:06:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[commit-bot: I haz the power, 2013-06-22 03:34:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[commit-bot: I haz the power, 2013-06-22 03:42:37]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rmcilroy@chromium.org/17416002/1

[commit-bot: I haz the power, 2013-06-22 14:22:04]

Change committed as 208036