Closed
Bug 1055825
Opened 10 years ago
Closed 10 years ago
Regression in Private Bytes/RSS after Bug 1026561 and Bug 990322
Categories
(Firefox :: New Tab Page, defect)
Tracking
()
RESOLVED
WONTFIX
People
(Reporter: khuey, Assigned: Mardak)
References
Details
(Keywords: memory-footprint, regression, Whiteboard: [MemShrink:P2])
Attachments
(1 file)
3.21 MB,
application/x-bzip2
|
Details |
Unsurprisingly Bug 1026561 made us use more memory. We probably don't care, but filing anyways.
Comment 1•10 years ago
|
||
This needs investigation because the increase seems larger than a few more images would account for.
Whiteboard: [MemShrink] → [MemShrink:P2]
Reporter | ||
Comment 3•10 years ago
|
||
[Tracking Requested - why for this release]: regression - nom for tracking
status-firefox33:
--- → unaffected
status-firefox34:
--- → affected
tracking-firefox34:
--- → ?
Keywords: regression
Assignee | ||
Comment 4•10 years ago
|
||
I noted in bug 990322 comment 24 (which is the next push after bug 1026561) that when it originally landed by itself (push 3f3f9b729ee7) then backed out, it caused a tp5o regression. http://graphs.mozilla.org/graph.html#tests=[[261,132,25],[261,132,33],[261,132,35],[261,132,37],[257,132,33],[257,132,35],[257,132,37],[257,132,25]]&sel=1407538811795.6667,1407543114017.889,140994743.69493282,633178961.0544211&displayrange=30&datatype=running bug 990322 linux64 rss: 7.9% To be clear, bug 1026561 (push 05f052f7c894) landed and regressed, and the followup push with bug 990322 (push de9d70a14a16) did not regress as much as it did solo (graphs above). http://graphs.mozilla.org/graph.html#tests=[[261,132,25],[261,132,33],[261,132,35],[261,132,37],[257,132,33],[257,132,35],[257,132,37],[257,132,25]]&sel=1408248504368.4443,1408265793257.3333,99233772.57553017,644124361.0607649&displayrange=30&datatype=running bug 1026561 linux64 rss: 9.2% bug 990322 linux64 rss: 2.2% We landed bug 1055261 (push dd0e4667d948) reducing it from 24 tiles (8 columns) to 15 tiles (5 columns), so still higher than the original 9 tiles (3 columns), but that seems to have had negligible effect: http://graphs.mozilla.org/graph.html#tests=[[261,132,33],[257,132,35],[257,132,33],[261,132,35],[261,132,37],[261,132,25],[257,132,37],[257,132,25]]&sel=1408418989784.3333,1408423182006.5557,170305932.5938471,586749527.2401186&displayrange=7&datatype=running bug 1055261 linux64 rss: -0.1%
Comment 5•10 years ago
|
||
Memory and DMD reports after loading about:newtab from before the regression started, the bug that triggered the primary regression, and with the latest m-c (where the regression has gone away). There's a bit of noise from about:memory and DMD that should be ignored. baseline: bug_1055825/dmd-pre.txt.gz bug_1055825/memory-report-pre.json.gz regression: bug_1055825/dmd-regression.txt.gz bug_1055825/memory-report-regression.json.gz m-c: bug_1055825/memory-report-latest.json.gz bug_1055825/dmd-latest.txt.gz
Comment 6•10 years ago
|
||
Diff of baseline vs regression, all that changed b/w builds were the pref changes. Of note if you ignore about:memory: - window-objects is +0.99 MB - heap-unclassified is +7.81 MB - js-non-window is +5.66 MB - workers/workers(chrome) is +2.13 MB - images is +0.91 MB From the DMD report for the regressed build the top offenders appear to all be CSS/Image/URI related. Diff: 30.91 MB (100.0%) -- explicit ├──11.57 MB (37.45%) -- window-objects │ ├──10.58 MB (34.23%) ++ top(about:memory, id=7) │ ├───1.88 MB (06.10%) ++ top(about:newtab, id=16) │ ├──-1.47 MB (-4.75%) ++ top(about:newtab, id=19) │ ├───1.45 MB (04.68%) ++ top(about:newtab, id=21) │ ├──-1.20 MB (-3.88%) ++ top(about:newtab, id=25) │ ├──-0.97 MB (-3.13%) ++ top(about:memory, id=13) │ ├───0.76 MB (02.46%) ++ top(chrome://browser/content/browser.xul, id=3) │ └───0.54 MB (01.74%) ++ top(chrome://browser/content/hiddenWindow.xul, id=1) ├───7.81 MB (25.26%) ── heap-unclassified ├───5.66 MB (18.33%) -- js-non-window │ ├──3.12 MB (10.11%) -- runtime │ │ ├──1.31 MB (04.25%) ++ code │ │ ├──0.70 MB (02.28%) ── uncompressed-source-cache │ │ ├──0.52 MB (01.69%) ++ script-sources │ │ ├──0.43 MB (01.40%) ── script-data │ │ └──0.15 MB (00.49%) ++ (3 tiny) │ ├──2.45 MB (07.91%) ++ zones/zone(0xNNN) │ └──0.09 MB (00.30%) ── gc-heap/chunk-admin ├───2.13 MB (06.90%) -- workers/workers(chrome) │ ├──2.13 MB (06.90%) ++ worker(resource://gre/modules/PageThumbsWorker.js, 0xNNN) │ └──0.00 MB (00.01%) ++ (2 tiny) ├───1.69 MB (05.48%) ++ dmd ├───1.06 MB (03.44%) ++ storage ├───0.91 MB (02.95%) ++ images ├──-0.25 MB (-0.81%) ++ heap-overhead ├──-0.42 MB (-1.36%) ++ network ├───0.43 MB (01.38%) ++ xpconnect └───0.30 MB (00.98%) ++ (12 tiny) DMD top hits: Unreported { 3 blocks in heap block record 1 of 2,318 1,875,968 bytes (1,869,037 requested / 6,931 slop) 2.44% of the heap (2.44% cumulative) 6.46% of unreported (6.46% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc moz_malloc mozilla::gfx::AlignedArray<unsigned char, 16>::Realloc(unsigned long) mozilla::gfx::DrawTargetCG::Init(mozilla::gfx::BackendType, unsigned char*, mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits> const&, int, mozilla::gfx::SurfaceFormat) mozilla::gfx::DrawTargetCG::Init(mozilla::gfx::BackendType, mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits> const&, mozilla::gfx::SurfaceFormat&) mozilla::gfx::Factory::CreateDrawTarget(mozilla::gfx::BackendType, mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits> const&, mozilla::gfx::SurfaceFormat) gfxPlatform::CreateDrawTargetForBackend(mozilla::gfx::BackendType, mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits> const&, mozilla::gfx::SurfaceFormat) gfxPlatform::CreateOffscreenContentDrawTarget(mozilla::gfx::IntSizeTyped<mozilla::gfx::UnknownUnits> const&, mozilla::gfx::SurfaceFormat) mozilla::layers::ContentClientBasic::CreateBuffer(gfxContentType, nsIntRect const&, unsigned int, mozilla::RefPtr<mozilla::gfx::DrawTarget>*, mozilla::RefPtr<mozilla::gfx::DrawTarget>*) non-virtual thunk to mozilla::layers::ContentClientBasic::CreateBuffer(gfxContentType, nsIntRect const&, unsigned int, mozilla::RefPtr<mozilla::gfx::DrawTarget>*, mozilla::RefPtr<mozilla::gfx::DrawTarget>*) mozilla::layers::RotatedContentBuffer::BeginPaint(mozilla::layers::ThebesLayer*, unsigned int) mozilla::layers::ContentClientBasic::BeginPaintBuffer(mozilla::layers::ThebesLayer*, unsigned int) mozilla::layers::BasicThebesLayer::Validate(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::ReadbackProcessor*) non-virtual thunk to mozilla::layers::BasicThebesLayer::Validate(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::ReadbackProcessor*) mozilla::layers::BasicContainerLayer::Validate(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::ReadbackProcessor*) non-virtual thunk to mozilla::layers::BasicContainerLayer::Validate(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::ReadbackProcessor*) mozilla::layers::BasicLayerManager::EndTransactionInternal(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::LayerManager::EndTransactionFlags) mozilla::layers::BasicLayerManager::EndTransaction(void (*)(mozilla::layers::ThebesLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::LayerManager::EndTransactionFlags) nsDisplayList::PaintForFrame(nsDisplayListBuilder*, nsRenderingContext*, nsIFrame*, unsigned int) nsDisplayList::PaintRoot(nsDisplayListBuilder*, nsRenderingContext*, unsigned int) } } Unreported { ~55 blocks in heap block record 2 of 2,318 ~1,748,971 bytes (~1,642,395 requested / ~106,576 slop) 2.27% of the heap (4.71% cumulative) 6.02% of unreported (12.48% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc nsStringBuffer::Alloc(unsigned long) nsAString_internal::MutatePrep(unsigned int, char16_t**, unsigned int*) nsAString_internal::ReplacePrepInternal(unsigned int, unsigned int, unsigned int, unsigned int) nsAString_internal::ReplacePrep(unsigned int, unsigned int, unsigned int) nsAString_internal::Replace(unsigned int, unsigned int, char16_t const*, unsigned int, mozilla::fallible_t const&) nsAString_internal::Replace(unsigned int, unsigned int, char16_t const*, unsigned int) nsAString_internal::Append(char16_t const*, unsigned int) nsCSSScanner::GatherText(unsigned char, nsString&) nsCSSScanner::ScanString(nsCSSToken&) nsCSSScanner::NextURL(nsCSSToken&) nsCSSScanner::ScanIdent(nsCSSToken&) nsCSSScanner::Next(nsCSSToken&, bool) (anonymous namespace)::CSSParserImpl::GetToken(bool) (anonymous namespace)::CSSParserImpl::ParseVariant(nsCSSValue&, int, short const*) (anonymous namespace)::CSSParserImpl::ParseValueList(nsCSSProperty) (anonymous namespace)::CSSParserImpl::ParseProperty(nsCSSProperty) (anonymous namespace)::CSSParserImpl::ParseProperty(nsCSSProperty, nsAString_internal const&, nsIURI*, nsIURI*, nsIPrincipal*, mozilla::css::Declaration*, bool*, bool, bool) nsCSSParser::ParseProperty(nsCSSProperty, nsAString_internal const&, nsIURI*, nsIURI*, nsIPrincipal*, mozilla::css::Declaration*, bool*, bool, bool) nsDOMCSSDeclaration::ParsePropertyValue(nsCSSProperty, nsAString_internal const&, bool) } } Unreported { ~25 blocks in heap block record 3 of 2,318 ~1,347,581 bytes (~1,303,376 requested / ~44,205 slop) 1.75% of the heap (6.47% cumulative) 4.64% of unreported (17.12% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc moz_malloc nsTArrayFallibleAllocator::Malloc(unsigned long) nsTArray_base<nsTArrayFallibleAllocator, nsTArray_CopyWithMemutils>::EnsureCapacity(unsigned long, unsigned long) nsTArray_Impl<char, nsTArrayFallibleAllocator>::SetCapacity(unsigned long) mozilla::image::RasterImage::SetSourceSizeHint(unsigned int) mozilla::image::ImageFactory::CreateRasterImage(nsIRequest*, imgStatusTracker*, nsCString const&, mozilla::image::ImageURL*, unsigned int, unsigned int) mozilla::image::ImageFactory::CreateImage(nsIRequest*, imgStatusTracker*, nsCString const&, mozilla::image::ImageURL*, bool, unsigned int) imgRequest::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) ProxyListener::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) nsBaseChannel::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) non-virtual thunk to nsBaseChannel::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) nsInputStreamPump::OnStateTransfer() nsInputStreamPump::OnInputStreamReady(nsIAsyncInputStream*) non-virtual thunk to nsInputStreamPump::OnInputStreamReady(nsIAsyncInputStream*) nsInputStreamReadyEvent::Run() nsThread::ProcessNextEvent(bool, bool*) NS_ProcessPendingEvents(nsIThread*, unsigned int) nsBaseAppShell::NativeEventCallback() nsAppShell::ProcessGeckoEvents(void*) } } Unreported { ~44 blocks in heap block record 4 of 2,318 ~901,096 bytes (~817,808 requested / ~83,288 slop) 1.17% of the heap (7.64% cumulative) 3.10% of unreported (20.22% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc nsStringBuffer::Alloc(unsigned long) nsACString_internal::MutatePrep(unsigned int, char**, unsigned int*) nsACString_internal::ReplacePrepInternal(unsigned int, unsigned int, unsigned int, unsigned int) nsACString_internal::ReplacePrep(unsigned int, unsigned int, unsigned int) nsACString_internal::Assign(char const*, unsigned int, mozilla::fallible_t const&) nsACString_internal::Assign(nsACString_internal const&, mozilla::fallible_t const&) nsACString_internal::Assign(nsACString_internal const&) nsCString::operator=(nsACString_internal const&) nsSimpleURI::SetPath(nsACString_internal const&) nsSimpleURI::SetSpec(nsACString_internal const&) nsDataHandler::NewURI(nsACString_internal const&, char const*, nsIURI*, nsIURI**) nsIOService::NewURI(nsACString_internal const&, char const*, nsIURI*, nsIURI**) NS_NewURI(nsIURI**, nsACString_internal const&, char const*, nsIURI*, nsIIOService*) mozilla::css::URLValue::GetURI() const nsCSSValue::StartImageLoad(nsIDocument*) const TryToStartImageLoadOnValue(nsCSSValue const&, nsIDocument*, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) MapSinglePropertyInto(nsCSSProperty, nsCSSValue const*, nsCSSValue*, nsRuleData*) } } Unreported { ~11 blocks in heap block record 5 of 2,318 ~647,165 bytes (~627,252 requested / ~19,913 slop) 0.84% of the heap (8.48% cumulative) 2.23% of unreported (22.45% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc moz_malloc nsTArrayFallibleAllocator::Malloc(unsigned long) nsTArray_base<nsTArrayFallibleAllocator, nsTArray_CopyWithMemutils>::EnsureCapacity(unsigned long, unsigned long) nsTArray_Impl<char, nsTArrayFallibleAllocator>::SetCapacity(unsigned long) mozilla::image::RasterImage::SetSourceSizeHint(unsigned int) mozilla::image::ImageFactory::CreateRasterImage(nsIRequest*, imgStatusTracker*, nsCString const&, mozilla::image::ImageURL*, unsigned int, unsigned int) mozilla::image::ImageFactory::CreateImage(nsIRequest*, imgStatusTracker*, nsCString const&, mozilla::image::ImageURL*, bool, unsigned int) imgRequest::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) ProxyListener::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) imgCacheValidator::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) nsBaseChannel::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) non-virtual thunk to nsBaseChannel::OnDataAvailable(nsIRequest*, nsISupports*, nsIInputStream*, unsigned long long, unsigned int) nsInputStreamPump::OnStateTransfer() nsInputStreamPump::OnInputStreamReady(nsIAsyncInputStream*) non-virtual thunk to nsInputStreamPump::OnInputStreamReady(nsIAsyncInputStream*) nsInputStreamReadyEvent::Run() nsThread::ProcessNextEvent(bool, bool*) NS_ProcessPendingEvents(nsIThread*, unsigned int) nsBaseAppShell::NativeEventCallback() } } Unreported { ~13 blocks in heap block record 6 of 2,318 ~331,767 bytes (~326,865 requested / ~4,902 slop) 0.43% of the heap (8.91% cumulative) 1.14% of unreported (23.59% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc nsStringBuffer::Alloc(unsigned long) nsACString_internal::MutatePrep(unsigned int, char**, unsigned int*) nsACString_internal::Assign(nsCSubstringTuple const&, mozilla::fallible_t const&) nsACString_internal::Assign(nsCSubstringTuple const&) nsACString_internal::operator=(nsCSubstringTuple const&) nsSimpleURI::GetSpec(nsACString_internal&) mozilla::image::ImageURL::ImageURL(nsIURI*) mozilla::image::ImageURL::ImageURL(nsIURI*) imgRequest::Init(nsIURI*, nsIURI*, nsIRequest*, nsIChannel*, imgCacheEntry*, void*, nsIPrincipal*, int) imgLoader::LoadImage(nsIURI*, nsIURI*, nsIURI*, nsIPrincipal*, nsILoadGroup*, imgINotificationObserver*, nsISupports*, unsigned int, nsISupports*, nsIChannelPolicy*, nsAString_internal const&, imgRequestProxy**) nsContentUtils::LoadImage(nsIURI*, nsIDocument*, nsIPrincipal*, nsIURI*, imgINotificationObserver*, int, nsAString_internal const&, imgRequestProxy**) mozilla::css::ImageLoader::LoadImage(nsIURI*, nsIPrincipal*, nsIURI*, mozilla::css::ImageValue*) mozilla::css::ImageValue::ImageValue(nsIURI*, nsStringBuffer*, nsIURI*, nsIPrincipal*, nsIDocument*) mozilla::css::ImageValue::ImageValue(nsIURI*, nsStringBuffer*, nsIURI*, nsIPrincipal*, nsIDocument*) nsCSSValue::StartImageLoad(nsIDocument*) const TryToStartImageLoadOnValue(nsCSSValue const&, nsIDocument*, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) MapSinglePropertyInto(nsCSSProperty, nsCSSValue const*, nsCSSValue*, nsRuleData*) } } Unreported { ~12 blocks in heap block record 7 of 2,318 ~327,674 bytes (~322,772 requested / ~4,902 slop) 0.43% of the heap (9.34% cumulative) 1.13% of unreported (24.72% cumulative) Allocated at { replace_malloc malloc_impl zone_malloc malloc_zone_malloc malloc nsStringBuffer::Alloc(unsigned long) nsACString_internal::MutatePrep(unsigned int, char**, unsigned int*) nsACString_internal::Assign(nsCSubstringTuple const&, mozilla::fallible_t const&) nsACString_internal::Assign(nsCSubstringTuple const&) nsACString_internal::operator=(nsCSubstringTuple const&) nsSimpleURI::GetSpec(nsACString_internal&) imgLoader::PutIntoCache(nsIURI*, imgCacheEntry*) imgLoader::LoadImage(nsIURI*, nsIURI*, nsIURI*, nsIPrincipal*, nsILoadGroup*, imgINotificationObserver*, nsISupports*, unsigned int, nsISupports*, nsIChannelPolicy*, nsAString_internal const&, imgRequestProxy**) nsContentUtils::LoadImage(nsIURI*, nsIDocument*, nsIPrincipal*, nsIURI*, imgINotificationObserver*, int, nsAString_internal const&, imgRequestProxy**) mozilla::css::ImageLoader::LoadImage(nsIURI*, nsIPrincipal*, nsIURI*, mozilla::css::ImageValue*) mozilla::css::ImageValue::ImageValue(nsIURI*, nsStringBuffer*, nsIURI*, nsIPrincipal*, nsIDocument*) mozilla::css::ImageValue::ImageValue(nsIURI*, nsStringBuffer*, nsIURI*, nsIPrincipal*, nsIDocument*) nsCSSValue::StartImageLoad(nsIDocument*) const TryToStartImageLoadOnValue(nsCSSValue const&, nsIDocument*, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) TryToStartImageLoad(nsCSSValue const&, nsIDocument*, nsCSSProperty, nsCSSValueTokenStream*) MapSinglePropertyInto(nsCSSProperty, nsCSSValue const*, nsCSSValue*, nsRuleData*) nsCSSCompressedDataBlock::MapRuleInfoInto(nsRuleData*) const mozilla::css::Declaration::MapNormalRuleInfoInto(nsRuleData*) const } }
Flags: needinfo?(erahm)
Comment 7•10 years ago
|
||
Diff of regressed with m-c: Again we'll want to ignore about:memory and DMD, of note: - window-objects is -1.71 MB - heap-unclassified is -6.08 MB - js-non-window is +2.17 MB - workers/workers(chrome) is -0.05 MB - images is +0.66 MB So the main difference here is heap-unclassified is way better and window-objects is better as well, but js-non-window went up, workers is still up, and images increased further. -21.29 MB (100.0%) -- explicit ├──-10.56 MB (49.59%) -- window-objects │ ├───-8.85 MB (41.57%) ++ top(about:memory, id=7) │ ├───-0.68 MB (03.20%) ++ top(chrome://browser/content/browser.xul, id=3) │ ├───-0.56 MB (02.65%) ++ top(chrome://browser/content/hiddenWindow.xul, id=1) │ ├───-0.28 MB (01.32%) ++ top(about:newtab, id=16) │ └───-0.18 MB (00.85%) -- (2 tiny) │ ├──-0.18 MB (00.84%) ++ top(about:newtab, id=21) │ └──-0.00 MB (00.00%) ++ top(about:blank, id=6)/active/window(about:blank)/dom ├───-6.08 MB (28.57%) ── heap-unclassified ├────2.88 MB (-13.54%) ++ startup-cache ├───-2.78 MB (13.06%) ++ heap-overhead ├───-2.17 MB (10.20%) -- js-non-window │ ├──-2.29 MB (10.74%) -- runtime │ │ ├──-0.70 MB (03.31%) ── uncompressed-source-cache │ │ ├──-0.69 MB (03.23%) ++ code │ │ ├──-0.47 MB (02.23%) -- script-sources │ │ │ ├──-0.48 MB (02.25%) ++ (14 tiny) │ │ │ ├───0.23 MB (-1.09%) ++ source(scripts=1564, <non-notable files>) │ │ │ └──-0.23 MB (01.07%) ++ source(scripts=1549, <non-notable files>) │ │ ├──-0.36 MB (01.67%) ── script-data │ │ └──-0.06 MB (00.30%) ++ (3 tiny) │ ├───0.19 MB (-0.90%) ++ zones/zone(0xNNN) │ └──-0.08 MB (00.37%) ── gc-heap/chunk-admin ├───-1.14 MB (05.35%) ++ dmd ├───-0.66 MB (03.08%) ++ storage ├───-0.66 MB (03.08%) ++ images ├────0.50 MB (-2.36%) ++ network ├───-0.33 MB (01.53%) ++ (14 tiny) │ ├──-0.05 MB (00.23%) -- workers/workers(chrome) └───-0.31 MB (01.44%) ++ xpconnect
Comment 8•10 years ago
|
||
And diff of baseline w/ m-c, this is a bit harder to interpret due the amount of commits in between, but we can still look at some of the same numbers: - window-objects is +1.02 MB - heap-unclassified is +1.73 MB - js-non-window is +3.49 MB - workers/workers(chrome) is +2.09 MB - images is +0.26 MB So we might argue that there's still a sizable regression. 9.62 MB (100.0%) -- explicit ├──2.88 MB (29.96%) ++ startup-cache ├──3.49 MB (36.30%) -- js-non-window │ ├──2.64 MB (27.42%) ++ zones/zone(0xNNN) │ ├──0.84 MB (08.72%) ++ runtime │ └──0.02 MB (00.16%) ── gc-heap/chunk-admin ├──-3.03 MB (-31.50%) ++ heap-overhead ├──2.09 MB (21.67%) -- workers/workers(chrome) │ ├──2.01 MB (20.94%) ++ worker(resource://gre/modules/PageThumbsWorker.js, 0xNNN) │ └──0.07 MB (00.73%) ++ (2 tiny) ├──1.02 MB (10.58%) -- window-objects │ ├──1.73 MB (17.99%) ++ top(about:memory, id=7) │ ├──1.60 MB (16.65%) ++ top(about:newtab, id=16) │ ├──-1.47 MB (-15.27%) ++ top(about:newtab, id=19) │ ├──1.27 MB (13.19%) ++ top(about:newtab, id=21) │ ├──-1.20 MB (-12.45%) ++ top(about:newtab, id=25) │ ├──-0.97 MB (-10.06%) ++ top(about:memory, id=13) │ ├──0.08 MB (00.82%) ++ top(chrome://browser/content/browser.xul, id=3) │ └──-0.03 MB (-0.29%) ++ (2 tiny) ├──1.73 MB (17.94%) ── heap-unclassified ├──0.56 MB (05.78%) ++ dmd/stack-traces ├──0.41 MB (04.25%) ++ storage ├──0.26 MB (02.68%) ++ images ├──0.12 MB (01.26%) ++ xpconnect └──0.10 MB (01.08%) ++ (14 tiny)
Comment 10•10 years ago
|
||
this set of regressions showed up on Aurora last week via the talos alerts.
Assignee | ||
Comment 11•10 years ago
|
||
Nod. Those bugs were uplifted to 33 for bug 1057602.
Blocks: 1057602
Comment 13•10 years ago
|
||
we still have linux32/64 regressions in private bytes on tp5o on aurora (v.34) as a result of the uplift.
Comment 14•10 years ago
|
||
Ed - I have set you as the owner as this seems to be your bug. Given that we still have a regression (at least we did a month ago), what is the next step here as 34 uplifts to Aurora on Monday?
Assignee: nobody → edilee
Assignee | ||
Comment 15•10 years ago
|
||
Per comment 6 showing the memory changes of the specific regression, they seem to be directly tied to showing more tiles (which involves loading more images, creating more tiles with more html/css/uris). One potential idea that I'll also note in bug 1059558, which has active work in improving performance of the newtab page, is that I believe our test machines run with high resolution and are able to show all tiles, but for many users, fewer than 3 rows and fewer than 5 columns will be visible -- we could possibly lazily load tiles that aren't immediately visible.
Status: NEW → RESOLVED
Closed: 10 years ago
Flags: needinfo?(edilee)
Resolution: --- → WONTFIX
Comment 16•10 years ago
|
||
Dropping tracking based on comment 15. I will note that bug 1059558 is already tracked.
You need to log in
before you can comment on or make changes to this bug.
Description
•