Issue metadata
Sign in to add a comment
|
43.9 KiB regression in resource_sizes (MonochromePublic.apk) at 551586:551586 |
||||||||||||||||||||
Issue descriptionCaused by “[builtins] Re-enable embedded builtins” Chromium Commit: abc123abc123abc123abc123abc123abc123abcd Review: https://chromium-review.googlesource.com/#/c/1013570/ Link to size graph: https://chromeperf.appspot.com/report?sid=a097e74b1aa288511afb4cb616efe0f95ba4d347ad61d5e835072f23450938ba&num_points=10&rev=480214 Debugging size regressions is documented at: https://chromium.googlesource.com/chromium/src/+/master/docs/speed/apk_size_regressions.md#Debugging-Apk-Size-Increase Based on the graph: 43.9 KiB native code. It looks like this increase was probably unexpected or might be avoidable. Please have a look and either: Close as “Won't Fix” with a short justification, or Land a revert / fix-up.
,
Apr 18 2018
Assigning to v8-autoroll@chromium.org because this is the only CL in range: Update V8 to version 6.8.34. Summary of changes available at: https://chromium.googlesource.com/v8/v8/+log/059b1287..bde5be65 Please follow these instructions for assigning/CC'ing issues: https://github.com/v8/v8/wiki/Triaging%20issues Please close rolling in case of a roll revert: https://v8-roll.appspot.com/ This only works with a Google account. CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel;luci.chromium.try:android_optional_gpu_tests_rel TBR=hablich@chromium.org,machenbach@chromium.org,kozyatinskiy@chromium.org,sergiyb@chromium.org Change-Id: Iae7c5b0f4356f1f808e55c7ea16f912ca0ce88e1 Reviewed-on: https://chromium-review.googlesource.com/1016044 Reviewed-by: v8 autoroll <v8-autoroll@chromium.org> Commit-Queue: v8 autoroll <v8-autoroll@chromium.org> Cr-Commit-Position: refs/heads/master@{#551586}
,
Apr 18 2018
Justification for picking the CL: When I run: tools/binary_size/diagnose_bloat.py --cloud bf455d0506522c0fc24b9aa5b4b267433fd0d114 I get: ******* Common Metadata: apk_file_name=apks/MonochromePublic.apk elf_arch=arm elf_file_name=lib.unstripped/libmonochrome.so gn_args=ffmpeg_branding="Chrome" goma_dir="/b/build/slave/cache/goma_client" is_chrome_branded=true is_debug=false ... Section Sizes (Total=44.0kb (45063 bytes)): .bss: 0 bytes (0 bytes) (not included in totals) .data: 0 bytes (0 bytes) (0.0%) .data.rel.ro: 0 bytes (0 bytes) (0.0%) .dex: 0 bytes (0 bytes) (0.0%) .other: -263kb (-269377 bytes) (-597.8%) .pak.nontranslated: 0 bytes (0 bytes) (0.0%) .pak.translations: 0 bytes (0 bytes) (0.0%) .rel.dyn: 4 bytes (4 bytes) (0.0%) .rodata: 256 bytes (256 bytes) (0.6%) .text: 306kb (314176 bytes) (697.2%) 76 symbols added (+), 230 changed (~), 12 removed (-), 956604 unchanged (not shown) Of changed symbols, 169 grew, 149 shrank Number of unique symbols 603939 -> 603989 (+50) 2 paths added, 0 removed, 41 changed Showing 318 symbols (231 -> 291 unique) with total pss: 26278 bytes Histogram of symbols based on PSS: (-524288,-262144]: 1 (-128,-64]: 6 (-4,-2]: 3 [2,4): 29 [64,128): 15 [2048,4096): 1 (-32768,-16384]: 1 (-64,-32]: 7 (-2,-1]: 2 [4,8): 26 [128,256): 3 [262144,524288): 1 (-1024,-512]: 3 (-32,-16]: 80 (-1,0): 4 [8,16): 27 [256,512): 3 (-512,-256]: 6 (-16,-8]: 19 (0,1): 2 [16,32): 26 [512,1024): 2 (-256,-128]: 2 (-8,-4]: 15 [1,2): 4 [32,64): 28 [1024,2048): 2 .text=306kb .rodata=-18.1kb .data.rel.ro=0 bytes .data=0 bytes .bss=0 bytes .dex=0 bytes .dex.method=0 bytes .pak.translations=0 bytes .pak.nontranslated=0 bytes .other=-263kb total=25.7kb Number of unique paths: 48 Section Legend: t=.text, r=.rodata, R=.data.rel.ro, d=.data, b=.bss, x=.dex, m=.dex.method, p=.pak.translations, P=.pak.nontranslated, o=.other Index | Running Total | Section@Address | Δ PSS (Δ size_without_padding) | Path ------------------------------------------------------------ + 0) 311316 (1177.7%) t@0x169cb00 311316 (0->311296) $root_gen_dir/v8/embedded.cc ~ 1) 39428 (149.1%) o@0x0 -271888 (1436976->1165088) v8/snapshot_blob_32.bin + 2) 20634 (78.1%) r@0x0 -18794 (0->0) {no path} ** aggregate padding of diff'ed symbols ~ 3) 23418 (88.6%) t@0x1865c24 2784 (27216->30000) {no path} ** lld merge strings + 4) 24900 (94.2%) t@0x134db00 1482 (0->1480) v8/src/builtins/builtins.cc v8::internal::Builtins::IsIsolateIndependent ... ~ 27) 25485 (96.4%) t@0x135caa8 116 (916->1032) v8/src/compiler.cc ******************************Resource Sizes Diff****************************** MonochromePublic.apk_Breakdown (+45,063 bytes) +1,410 bytes Zip Overhead -271,888 bytes V8 Snapshots size +142 bytes unwind_cfi (dev and canary only) size +315,396 bytes Native code size +3 bytes Package metadata size MonochromePublic.apk_Specifics +44,921 bytes normalized apk size +315,396 bytes main lib size I 78732 See detailed diff results here: out/binary-size-results/diffs/f3150125ff552dc17f7706a76fc39e5c90c27347..bf455d0506522c0fc24b9aa5b4b267433fd0d114/diff_results.txt Diff Summary +44921 bytes MonochromePublic.apk_Specifics normalized apk size for range: f3150125ff552dc17f7706a76fc39e5c90c27347..bf455d0506522c0fc24b9aa5b4b267433fd0d114 I 78733 Enter supersize console via: tools/binary_size/supersize console out/binary-size-results/f3150125ff552dc17f7706a76fc39e5c90c27347/MonochromePublic.apk.size out/binary-size-results/bf455d0506522c0fc24b9aa5b4b267433fd0d114/MonochromePublic.apk.size ******* So items (1) and (2) are big additions followed by big reduction. (1) is due to embedded.cc, which matches description of https://chromium-review.googlesource.com/c/v8/v8/+/1013570 . A more thorough scan can be obtained from tools/binary_size/diagnose_bloat.py bde5be65 --reference-rev c7e6cf7^ --subrepo v8 --all -v but I cannot build chrome_public_pak for some reason, and this would be time consuming.
,
Apr 19 2018
I'm confused what items you mean by '(1)' and '(2)'. The CL is expected to significantly increase binary size (due to embedded.cc) and significantly decrease the snapshot size. There's no 1-to-1 correlation since the serialized format of the snapshot differs from the raw binary. Closing as WontFix, but please feel free to reopen in case I missed something that seems unexpected. |
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by 42576172...@developer.gserviceaccount.com
, Apr 18 2018