Necko/MobileCache: Difference between revisions

Necko/MobileCache (view source)

Revision as of 11:11, 2 September 2011

437 bytes removed , 2 September 2011

→‎TODO list

BjarneG

97

edits

@@ Line 25: / Line 25: @@
 Right now this TODO list is mostly focused on getting more info about what's going on, since it'll be hard to make progress until we know why we're slow.
-* ''Bjarne/Michal'': write some xpcshell scripts to get performance data:
+* ''Bjarne/Michal'': continue work on scripts to get performance data (see  [[current state of microbenchmarks]])
-** Figure out a way to turn disk caching on/off for xpcshell tests (I believe an env variable already exists: $NECKO_DEV_ENABLE_DISK_CACHE).  If need be, revisit [https://bugzilla.mozilla.org/show_bug.cgi?id==584283 bug 584283]
+** Figure out a way to turn disk caching on/off for xpcshell tests (I believe an env variable already exists: $NECKO_DEV_ENABLE_DISK_CACHE).  If need be, revisit {{bug|584283}}.
-** Write microbenchmark that measures average time to load channel if it gets a cache miss (run with disk cache on and off)
+** If cache miss takes longer on mobile when disk cache on, presumably it's from the disk IO we do to check for entry, which delays network request.  Measure to see how much that's true.
-*** If cache miss takes longer on mobile when disk cache on, presumably it's from the disk IO we do to check for entry, which delays network request.  Measure to see how much that's true.
-** Write test to measure how long channel takes if cache hit (versus immediate network load when cache off).
-*** write test so that we measure both small/large channel payloads, since performance is likely to vary (mobile disks may have similar latency to network, but presumably higher bandwidth?).
 * ''mobile-team'':  run the above tests and give us the perf numbers for 3G and wifi latencies/bandwidth.