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

Restore differential drawing to DirectWrite renderer #1069

New Issue
Closed
opened 2026-01-30 22:15:42 +00:00 by claunia · 10 comments
claunia commented 2026-01-30 22:15:42 +00:00
Owner
Copy Link

Originally created by @miniksa on GitHub (May 14, 2019).

Originally assigned to: @miniksa on GitHub.

For some reason, I turned this off in the haste to get things "working" well enough.

But it's definitely a performance hog. We should get differential drawing back inside the DirectWrite renderer.

Most of the provisions are already there with the invalid rectangle, we're just not using them.

I'm not sure if I'll be able to do cell-by-cell differential drawing like the GDI renderer because the more advanced glyphs/runs we support now (like ligatures) might necessitate that we instead just mark "invalid lines" and redraw the whole line. That would be mildly better than the whole screen.

This also would re-enable the use of the Present1 call to the DX infrastructure which is more efficient at scrolling and dirty regions.

Originally created by @miniksa on GitHub (May 14, 2019). Originally assigned to: @miniksa on GitHub. For some reason, I turned this off in the haste to get things "working" well enough. But it's definitely a performance hog. We should get differential drawing back inside the DirectWrite renderer. Most of the provisions are already there with the invalid rectangle, we're just not using them. I'm not sure if I'll be able to do cell-by-cell differential drawing like the GDI renderer because the more advanced glyphs/runs we support now (like ligatures) might necessitate that we instead just mark "invalid lines" and redraw the whole line. That would be mildly better than the whole screen. This also would re-enable the use of the `Present1` call to the DX infrastructure which is more efficient at scrolling and dirty regions.
claunia commented 2026-01-30 22:15:44 +00:00
Author
Owner
Copy Link

@oising commented on GitHub (May 14, 2019):

I think line invalidation is an honorable first pass and would be welcome.

@oising commented on GitHub (May 14, 2019): I think line invalidation is an honorable first pass and would be welcome.
claunia commented 2026-01-30 22:15:44 +00:00
Author
Owner
Copy Link

@skyline75489 commented on GitHub (Nov 11, 2019):

I've tried enable differential drawing on db79758092 . To my surprise, it does not help much of the performance as I originally anticipated.

First with #2959 and co the raw rendering performance is faster than before. In a way it is "fast enough" to deliquate the effect of differential drawing. Second, AFAIK most terminal operations requires entire screen redrawing, at which differential drawing makes little difference.

Update: I'm blinded by the PC I've be using myself. People with 4K display or higher would still need this.

@skyline75489 commented on GitHub (Nov 11, 2019): I've tried enable differential drawing on https://github.com/microsoft/terminal/commit/db797580923472c08613f054364097be06d73229 . To my surprise, it does not help much of the performance as I originally anticipated. First with #2959 and co the raw rendering performance is faster than before. In a way it is "fast enough" to deliquate the effect of differential drawing. Second, AFAIK most terminal operations requires entire screen redrawing, at which differential drawing makes little difference. **Update**: I'm blinded by the PC I've be using myself. People with 4K display or higher would still need this.
claunia commented 2026-01-30 22:15:45 +00:00
Author
Owner
Copy Link

@miniksa commented on GitHub (Mar 30, 2020):

A comparison:

Running each of these commands for ~15 seconds under WPR on the same dev box.
Cleaned up as many bg processes as possible, graphs scoped to just the WindowsTerminalDev package.

Master

  • PTY - individual cells
  • Term - full screen blit

DX cells

  • PTY - individual cells
  • Term - individual cells

note: Ligatures and multi-cell glyphs don't work well with this.

DX lines

  • PTY - individual cells
  • Term - individual complete rows

DX lines fixed

  • PTY - individual cells
  • Term - individual complete rows with patched til::bitmap algorithm to used optimized set method on dynamic_bitset<>

cmatrix

master

image

dx cells

image

dx lines

image

dx lines fixed

image

cmatrix -u0

master

image

dx cells

image

dx lines

image

dx lines fixed

image

cacafire

master

image

dx cells

image

dx lines

image

dx lines fixed

image

And for flavor, this is what cacafire graphs look like. It's kinda fun. You can see it ramping up in the amount of work required to draw the growing fire.
image

@miniksa commented on GitHub (Mar 30, 2020): A comparison: Running each of these commands for ~15 seconds under WPR on the same dev box. Cleaned up as many bg processes as possible, graphs scoped to just the `WindowsTerminalDev` package. **Master** - PTY - individual cells - Term - full screen blit **DX cells** - PTY - individual cells - Term - individual cells *note:* Ligatures and multi-cell glyphs don't work well with this. **DX lines** - PTY - individual cells - Term - individual complete rows **DX lines fixed** - PTY - individual cells - Term - individual complete rows with patched `til::bitmap` algorithm to used optimized set method on `dynamic_bitset<>` ## cmatrix ### master ![image](https://user-images.githubusercontent.com/18221333/77964754-790f6600-7294-11ea-9c1d-2e6969179c3c.png) ### dx cells ![image](https://user-images.githubusercontent.com/18221333/77964629-3a79ab80-7294-11ea-8639-4ced716157a7.png) ### dx lines ![image](https://user-images.githubusercontent.com/18221333/77964984-f509ae00-7294-11ea-87c5-0a456e575453.png) ### dx lines fixed ![image](https://user-images.githubusercontent.com/18221333/77967139-5a5f9e00-7299-11ea-9483-b0a46a1e9211.png) ## cmatrix -u0 ### master ![image](https://user-images.githubusercontent.com/18221333/77964838-a65c1400-7294-11ea-80fd-407cc24e9add.png) ### dx cells ![image](https://user-images.githubusercontent.com/18221333/77964723-64cb6900-7294-11ea-8273-c6dcd83711b6.png) ### dx lines ![image](https://user-images.githubusercontent.com/18221333/77964962-ec18dc80-7294-11ea-9c1a-d25aae94ccd5.png) ### dx lines fixed ![image](https://user-images.githubusercontent.com/18221333/77967424-ed003d00-7299-11ea-8b1f-35b31356ab49.png) ## cacafire ### master ![image](https://user-images.githubusercontent.com/18221333/77964858-b247d600-7294-11ea-9968-06fd5bd96e21.png) ### dx cells ![image](https://user-images.githubusercontent.com/18221333/77964710-5d0bc480-7294-11ea-8eb7-ed7c188915ef.png) ### dx lines ![image](https://user-images.githubusercontent.com/18221333/77964938-de635700-7294-11ea-8b1c-6cd792e6c0c2.png) ### dx lines fixed ![image](https://user-images.githubusercontent.com/18221333/77966723-6d25a300-7298-11ea-955a-71a19d1a9389.png) --------------- And for flavor, this is what `cacafire` graphs look like. It's kinda fun. You can see it ramping up in the amount of work required to draw the growing fire. ![image](https://user-images.githubusercontent.com/18221333/77965052-166a9a00-7295-11ea-8dcf-f83af8c13e24.png)
claunia commented 2026-01-30 22:15:45 +00:00
Author
Owner
Copy Link

@miniksa commented on GitHub (Mar 30, 2020):

Figuring out why dx lines cacafire is so much slower than the others
image
The act of setting all the bits in the rows appears to be taking a long time.

Updated trace changing void til::bitmap::set(const til::rectangle) to use dynamic_bitset<>::set(pos, len) to set the whole row at once:
image

Back to almost normal.

@miniksa commented on GitHub (Mar 30, 2020): Figuring out why `dx lines` `cacafire` is so much slower than the others ![image](https://user-images.githubusercontent.com/18221333/77965806-ab21c780-7296-11ea-9a69-76d659640845.png) The act of setting all the bits in the rows appears to be taking a long time. Updated trace changing `void til::bitmap::set(const til::rectangle)` to use `dynamic_bitset<>::set(pos, len)` to set the whole row at once: ![image](https://user-images.githubusercontent.com/18221333/77966723-6d25a300-7298-11ea-955a-71a19d1a9389.png) Back to almost normal.
claunia commented 2026-01-30 22:15:45 +00:00
Author
Owner
Copy Link

@miniksa commented on GitHub (Mar 30, 2020):

dxlines cmatrix: (with fix to setting rectangles)
image

@miniksa commented on GitHub (Mar 30, 2020): `dxlines` `cmatrix`: (with fix to setting rectangles) ![image](https://user-images.githubusercontent.com/18221333/77967139-5a5f9e00-7299-11ea-9483-b0a46a1e9211.png)
claunia commented 2026-01-30 22:15:46 +00:00
Author
Owner
Copy Link

@miniksa commented on GitHub (Mar 30, 2020):

dxlines cmatrix -u0: (with fix to setting rectangles)
image

I think that's justification for using the row/line invalidator for DX. It's better than painting the whole screen and works for ligatures. Not as fast as individual cells, but scoping it down further can be a later problem.

@miniksa commented on GitHub (Mar 30, 2020): `dxlines` `cmatrix -u0`: (with fix to setting rectangles) ![image](https://user-images.githubusercontent.com/18221333/77967424-ed003d00-7299-11ea-8b1f-35b31356ab49.png) I think that's justification for using the row/line invalidator for DX. It's better than painting the whole screen and works for ligatures. Not as fast as individual cells, but scoping it down further can be a later problem.
claunia commented 2026-01-30 22:15:46 +00:00
Author
Owner
Copy Link

@miniksa commented on GitHub (Apr 15, 2020):

I have to back this out for 1.0. I can't land #5345. There are too many loose ends.

@miniksa commented on GitHub (Apr 15, 2020): I have to back this out for 1.0. I can't land #5345. There are too many loose ends.
claunia commented 2026-01-30 22:15:46 +00:00
Author
Owner
Copy Link

@DHowett-MSFT commented on GitHub (Apr 16, 2020):

Closing this one out again pending our decision on differential drawing. We think we can stick the landing for v1.

@DHowett-MSFT commented on GitHub (Apr 16, 2020): Closing this one out again pending our decision on differential drawing. We think we can stick the landing for v1.
claunia commented 2026-01-30 22:15:47 +00:00
Author
Owner
Copy Link

@ghost commented on GitHub (May 5, 2020):

:tada:This issue was addressed in #5185, which has now been successfully released as Windows Terminal Release Candidate v0.11.1251.0 (1.0rc1).🎉

Handy links:

@ghost commented on GitHub (May 5, 2020): :tada:This issue was addressed in #5185, which has now been successfully released as `Windows Terminal Release Candidate v0.11.1251.0 (1.0rc1)`.:tada: Handy links: * [Release Notes](https://github.com/microsoft/terminal/releases/tag/v0.11.1251.0 (1.0rc1)) * [Store Download](https://www.microsoft.com/store/apps/9n0dx20hk701?cid=storebadge&ocid=badge)
claunia commented 2026-01-30 22:15:47 +00:00
Author
Owner
Copy Link

@BarrettStephen commented on GitHub (May 6, 2020):

using the terminal is so much faster now. its just crazy fast. a git pull used to scroll so slowly and now its just boom done

@BarrettStephen commented on GitHub (May 6, 2020): using the terminal is *so much faster* now. its just crazy fast. a git pull used to scroll so slowly and now its just boom done
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-Performance Area-Rendering Issue-Task Priority-0 Product-Terminal Resolution-Fix-Committed
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#1069
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?