Skip to content

PERFORMANCE.md

Latest commit

 

History

History
88 lines (72 loc) · 3.19 KB

PERFORMANCE.md

File metadata and controls

88 lines (72 loc) · 3.19 KB

Performance

MiniTaskbar is designed to remain quiet while idle and to avoid desktop frame capture. This page documents the measurements collected for version 1.0.1.

Design Principles

  • A compositor-managed DWM thumbnail displays the existing desktop host. MiniTaskbar does not capture or repaint wallpaper frames.
  • Explorer-scoped WinEvent hooks wake centering immediately for relevant application-button changes.
  • Accessibility state-only notifications are ignored because they do not change button geometry.
  • Application-button and real foreground changes reassert taskbar transparency immediately, then once more after a short settling interval. Foreground-location changes do not perform that extra style work.
  • Animation advances at approximately 60 fps only while the application button group is moving.
  • Inexpensive fallback checks recover from Explorer changes without a busy loop.

Method

The reference measurements were collected on Windows 10 22H2 (19045) with a 1920x1080 desktop and a single resident MiniTaskbar process. The idle sample observed the installed process for 60 seconds after its taskbar state had settled.

The figures measure MiniTaskbar process behavior. They do not claim to measure complete electrical consumption or establish a universal result for every machine.

Results

IdleSampleSeconds      : 60
CpuPercentOfOneCore    : 0.417
WorkingSetMiB          : 25.18
PrivateMemoryMiB       : 23.69
Handles                : 219
Threads                : 5
BuiltExecutableBytes   : 24064

Visual And Functional Verification

Focused probes also exercised the user-visible paths:

TaskbarOverlayRect              : [0,1040,1920,1080]
WallpaperSource                 : WorkerW
CenteredApplicationGroupDeltaPx : 0
FullscreenWinKeyBackingShownMs  : <=47
FullscreenWinKeyWallpaperMs     : <=408
FullscreenReturnBackingHiddenMs : <=29
AppOpenAccentRecoveryMs         : <=40
FallbackAccentCadenceMs         : 5000
IconOpenFirstMoveDeltaPx        : 23
PurpleFallbackObserved          : false
UnderlyingWindowVisible         : false

The fullscreen probe used a true foreground fullscreen window, deliberately forced a purple taskbar state, pressed the Windows key, and returned to the fullscreen application. A separate probe forced the same purple state before opening an application and observed event-driven recovery before its first usable 40 ms sample. The icon probe added and removed an application button and observed the centered slide.

The exact executable published for v1.0.1 was also exercised against the original reported path with a real foreground Chrome fullscreen window:

InitialFullscreenBackingHidden  : true
WinKeyBackingShownMs            : 46
ForcedPurpleSample              : [255,0,255]
AppButtonAccentRecoveryMs       : 117
RecoveredWallpaperSample        : [101.1,56.3,56.9]
FullscreenReturnBackingHiddenMs : 66
FinalBackingVisible             : false
ResidentProcesses               : 1

Interpretation

Measurements depend on the machine, desktop layout, Explorer behavior, and wallpaper host. They provide a concrete regression baseline for this project, not a promise that every environment will report identical values.