starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-02-04 05:35:20 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

Dedicated "paint" interface for VtRenderer #14438

New Issue
Open
opened 2026-01-31 04:10:23 +00:00 by claunia · 0 comments
claunia commented 2026-01-31 04:10:23 +00:00
Owner
Copy Link

Originally created by @skyline75489 on GitHub (Jul 9, 2021).

Description

As shown in #10563, ConPTY suffers from significant performance drop with too much scrolling request. A typical repro is:

time bash -c 'yes | head -n1000000'

This is a typical trace of OpenConsole.exe under the above workload:

image

Discussion

The main issue that caused this, I think, is that VtRenderer suffers from using the same interface as other graphic renderer. To break it down:

  1. Carriage return & linefeed can trigger graphic renderer to circling the buffer & force repaint. But for VtRenderer, it dosen't really need to do too much, other than print the newly added line. It doesn't really need to flush the buffer (EndPaint as shown above), because the actual text added to the buffer is only a single line (120 chars in a typical display).
  2. The _PaintBufferOutputHelper method insider renderer is obviously graphic-oriented. After all these years of development, I don't think it suits the need for VtRenderer anymore. An example is #10567, in which I try to reduce the cost of reassemble the line buffer. The reason why we need to reassemble the clusters, is because they are produced by splitting up the original buffer in _PaintBufferOutputHelper, which is heavily required by graphic renderers. If we have a dedicated interface, unnecessary operations like these can be eliminated.
  3. In VtEngine & XtermEngine, the actual number of methods that're both overridden and practically useful is rather small, comparing to the large amount of base methods in IRenderEngine. This indicates that the fundamental goal of VtEngine isn't really the same as the graphic renderers.

Proposed solution

I'm proposing a three-phase refactoring process for VtEngine to move away from IRenderEngine:

  1. Separate the methods in VtEngine that are heavily burdened by the current design, such as _PaintBufferOutputHelper. We add them as base methods to IRenderEngine. VtEngine is still a child class of IRenderEngine, but we add dedicated IO methods in it and test the performance gain.
  2. Separate the rest the methods in VtEngine same way as above.
  3. When we finished the first two steps, we should have a clear path towards adding a dedicated interface just for VtEngine, probably name it IVtEngine. Then we can create IVtEngine from IRenderEngine, and let IRenderEngine be what's left in it.

CC @DHowett @miniksa @zadjii-msft @lhecker @j4james for discussion.

(How can I just At everyone who's collaborator of this project. That would be helpful.)

Originally created by @skyline75489 on GitHub (Jul 9, 2021). ### Description As shown in #10563, ConPTY suffers from significant performance drop with too much scrolling request. A typical repro is: ```bash time bash -c 'yes | head -n1000000' ``` This is a typical trace of `OpenConsole.exe` under the above workload: ![image](https://user-images.githubusercontent.com/4710575/125062000-809cfc00-e0e0-11eb-9d0d-c9ec9e95884b.png) ### Discussion The main issue that caused this, I think, is that VtRenderer suffers from using the same interface as other graphic renderer. To break it down: 1. Carriage return & linefeed can trigger graphic renderer to circling the buffer & force repaint. But for VtRenderer, it dosen't really need to do too much, other than print the newly added line. It doesn't really need to flush the buffer (`EndPaint` as shown above), because the actual text added to the buffer is only a single line (120 chars in a typical display). 2. The `_PaintBufferOutputHelper` method insider `renderer` is obviously graphic-oriented. After all these years of development, I don't think it suits the need for VtRenderer anymore. An example is #10567, in which I try to reduce the cost of reassemble the line buffer. The reason why we need to reassemble the clusters, is because they are produced by splitting up the original buffer in `_PaintBufferOutputHelper`, which is heavily required by graphic renderers. If we have a dedicated interface, unnecessary operations like these can be eliminated. 3. In `VtEngine` & `XtermEngine`, the actual number of methods that're both overridden and practically useful is rather small, comparing to the large amount of base methods in `IRenderEngine`. This indicates that the fundamental goal of `VtEngine` isn't really the same as the graphic renderers. ### Proposed solution I'm proposing a three-phase refactoring process for VtEngine to move away from `IRenderEngine`: 1. Separate the methods in `VtEngine` that are heavily burdened by the current design, such as `_PaintBufferOutputHelper`. We add them as base methods to `IRenderEngine`. `VtEngine` is still a child class of `IRenderEngine`, but we add dedicated IO methods in it and test the performance gain. 1. Separate the rest the methods in `VtEngine` same way as above. 1. When we finished the first two steps, we should have a clear path towards adding a dedicated interface just for `VtEngine`, probably name it `IVtEngine`. Then we can create `IVtEngine` from `IRenderEngine`, and let `IRenderEngine` be what's left in it. CC @DHowett @miniksa @zadjii-msft @lhecker @j4james for discussion. (How can I just At everyone who's collaborator of this project. That would be helpful.)
claunia added the Issue-FeatureNeeds-TriageArea-VTProduct-Conpty labels 2026-01-31 04:10:24 +00:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
⛺ Reserved
A11yCO
A11yMAS
A11ySev1
A11ySev2
A11ySev3
A11yTTValidated
A11yUsable
A11yVoiceAccess
A11yWCAG
Area-Accessibility
Area-AtlasEngine
Area-AzureShell
Area-Build
Area-Build
Area-Chat
Area-CmdPal
Area-CodeHealth
Area-Commandline
Area-CookedRead
Area-DefApp
Area-Extensibility
Area-Fonts
Area-GroupPolicy
Area-i18n
Area-Input
Area-Interaction
Area-Interop
Area-Localization
Area-Output
Area-Performance
Area-Portable
Area-Quality
Area-Remoting
Area-Rendering
Area-Schema
Area-Server
Area-Settings
Area-SettingsUI
Area-ShellExtension
Area-ShellExtension
Area-ShellExtension
Area-Suggestions
Area-Suggestions
Area-TerminalConnection
Area-TerminalControl
Area-Theming
Area-UserInterface
Area-VT
Area-Windowing
Area-WPFControl
AutoMerge
Blocking-Ingestion
Culprit-Centennial
Culprit-WinUI
Disability-All
Disability-Blind
Disability-LowVision
Disability-Mobility
External-Blocked-WinUI3
Fixed
Gathering-Data
good first issue
HCL-E+D
HCL-WindowsTerminal
Help Wanted
Impact-Compatibility
Impact-Compliance
Impact-Correctness
Impact-Visual
In-PR
InclusionBacklog
InclusionBacklog-Windows TerminalWin32
InclusionCommitted-202206
Issue-Bug
Issue-Docs
Issue-Feature
Issue-Feature
Issue-Question
Issue-Samples
Issue-Scenario
Issue-Task
Needs-Attention
Needs-Author-Feedback
Needs-Bisect
Needs-Discussion
Needs-Repro
Needs-Tag-Fix
Needs-Tag-Fix
Needs-Triage
No-Recent-Activity
Priority-0
Priority-1
Priority-2
Priority-3
Product-Cmd.exe
Product-Colortool
Product-Colortool
Product-Colortool
Product-Conhost
Product-Conpty
Product-Meta
Product-Powershell
Product-Terminal
Product-WSL
pull-request
Mirrored from GitHub Pull Request
 Resolution-Answered
Resolution-By-Design
Resolution-Duplicate
Resolution-External
Resolution-Fix-Available
Resolution-Fix-Committed
Resolution-No-Repro
Resolution-Won't-Fix
Severity-Blocking
Severity-Crash
Severity-DataLoss
spam
this-will-be-a-breaking-change
Tracking-External
WindowsTerminal_Win32
Work-Item
zAskModeBug
zInbox-Bug
No Label Area-VT Issue-Feature Needs-Triage Product-Conpty
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
claunia
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: starred/terminal#14438
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?