All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=920337

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=833c0e68dae40984bd6be4b3c6c94a5fcda56d5d0df4aab13d38f47168e6aed1


Bot(s) for this bug's original alert(s):

mac-10_13_laptop_high_end-perf

system_health.memory_desktop - Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/135ab251940000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/135ab251940000

Use Skia's decal mode for non-repeating image patterns by fmalita@chromium.org
https://chromium.googlesource.com/chromium/src/+/89d63e3a995cfca2be4419655d1767e72faf5d53
memory:chrome:all_processes:reported_by_chrome:cc:effective_size: 4.869e+07 → 4.992e+07 (+1.235e+06)

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

Benchmark documentation link:
  https://bit.ly/system-health-benchmarks

This is actually an overall progression:

  * https://chromium-review.googlesource.com/c/chromium/src/+/1376252 streamlined the handling of image patterns repeating in only one dimension - previously we were wrapping the image as a SkPictureImageGenerator-based shader to pad edges, now we're just building a shader from the original image

  * CC's ImageDecodeController knows how to handle decoding/caching for image shaders upfront -> thus there is additional RAM accounted against CC

  * previously, CC would let the picture image generator pass to Skia, which would handle the caching (multiplied by potential unique ID churn) -> I would expect to find a corresponding drop in RAM accounted against Skia

  * lo and behold, memory:chrome:all_processes:reported_by_chrome:skia:effective_size shows a corresponding drop for the same revision: https://chromeperf.appspot.com/report?sid=7983b5937294ba083c7786a195e7ea567d8ecee18e07005d7dbea5962d55c448&start_rev=615470&end_rev=619025


memory:chrome:all_processes:reported_by_chrome:cc:effective_size:   + 1491587 bytes
memory:chrome:all_processes:reported_by_chrome:skia:effective_size: -13359357 bytes

bottom line:                                                        -11867770 bytes

khushalsagar@, does this sound plausible?