starred/terminal
1
0
Fork 0
mirror of https://github.com/microsoft/terminal.git synced 2026-07-08 18:16:28 +00:00
Code Issues 19.0k Packages Projects Releases 20 Wiki Activity

Use virtual memory for the console buffer #14609

New Issue
Closed
opened 2026-01-31 04:14:49 +00:00 by claunia · 3 comments
claunia commented 2026-01-31 04:14:49 +00:00
Owner
Copy Link

Originally created by @lhecker on GitHub (Jul 20, 2021).

The console buffer is currently implemented as something like this (simplified, source):

struct CharRow {
    small_vector<wchar_t, 120> chars;
};

struct TextBuffer {
    std::vector<CharRow> rows;
};

void TextBuffer::TextBuffer(int width = 120, int height = 9001) {
    rows.resize(height);

    for (const auto& row : rows) {
        row.chars.resize(width);
    }
}

This works reasonably well if your terminal is 120 characters wide, as the small_vector will prevent any heap allocations up to that size. But if your terminal is any larger it will allocate every row separately on the heap.
Additionally the use of C++ container classes force any memory to be initialized eagerly, which pessimisifies memory usage for large allocations, as all affected virtual memory is eagerly forced to be paged into physical memory.
Finally the large structure size of ROW causes instance constructions and copies, as well as iterations through the rows vector to be very costly, even if just data is read that's unrelated to the characters in the buffer (-25% performance impact).

Proposed technical implementation details

Allocate a large backing buffer for the entire TextBuffer using VirtualAlloc. Each CharRow will then only hold a begin/end pointer into the large buffer. This reduces the structure size and improves buffer performance by +20%.
Most importantly the memory used by a 160x48 conhost process is reduced from 12MB down to just 3MB (-75%) after startup.

Originally created by @lhecker on GitHub (Jul 20, 2021). The console buffer is currently implemented as something like this (simplified, [source](https://github.com/microsoft/terminal/blob/9c1331ab2ec517e832cf8fd4e0c88275103acc9f/src/buffer/out/textBuffer.cpp#L30-L52)): ```cpp struct CharRow { small_vector<wchar_t, 120> chars; }; struct TextBuffer { std::vector<CharRow> rows; }; void TextBuffer::TextBuffer(int width = 120, int height = 9001) { rows.resize(height); for (const auto& row : rows) { row.chars.resize(width); } } ``` This works reasonably well if your terminal is 120 characters wide, as the `small_vector` will prevent any heap allocations up to that size. But if your terminal is any larger it will allocate every row separately on the heap. Additionally the use of C++ container classes force any memory to be initialized eagerly, which pessimisifies memory usage for large allocations, as all affected virtual memory is eagerly forced to be paged into physical memory. Finally the large structure size of `ROW` causes instance constructions and copies, as well as iterations through the `rows` vector to be very costly, even if just data is read that's unrelated to the characters in the buffer (-25% performance impact). # Proposed technical implementation details Allocate a large backing buffer for the entire `TextBuffer` using `VirtualAlloc`. Each `CharRow` will then only hold a begin/end pointer into the large buffer. This reduces the structure size and improves buffer performance by +20%. Most importantly the memory used by a 160x48 conhost process is reduced from 12MB down to just 3MB (-75%) after startup.
claunia added the Product-ConhostArea-OutputIssue-TaskNeeds-Tag-Fix labels 2026-01-31 04:14:49 +00:00
claunia commented 2026-01-31 04:14:50 +00:00
Author
Owner
Copy Link

@zadjii-msft commented on GitHub (Jul 20, 2021):

Allocate a large backing buffer for the entire TextBuffer using VirtualAlloc. Each CharRow will then only hold a begin/end pointer into the large buffer. This reduces the structure size and improves buffer performance by +20%.
Most importantly the memory used by a 160x48 conhost process is reduced from 12MB down to just 3MB (-75%) after startup.

we've really come full circle, haven't we? IIRC, that's pretty similar to how the conhost buffer was designed at the start.

@zadjii-msft commented on GitHub (Jul 20, 2021): > Allocate a large backing buffer for the entire TextBuffer using VirtualAlloc. Each CharRow will then only hold a begin/end pointer into the large buffer. This reduces the structure size and improves buffer performance by +20%. > Most importantly the memory used by a 160x48 conhost process is reduced from 12MB down to just 3MB (-75%) after startup. we've really come full circle, haven't we? IIRC, that's pretty similar to how the conhost buffer was designed at the start.
claunia commented 2026-01-31 04:14:51 +00:00
Author
Owner
Copy Link

@DHowett commented on GitHub (Jul 20, 2021):

We will need to design this carefully to ensure that it works with the new column counting design in #8000, as well as can fault rows out to larger storage if they need to violate the implied "one wchar_t per column" invariant.

I'd rather use something a little less efficient but a little more "hands-off" like pmr and memory resources instead of having to come up with a defragmenting allocator to reign over a single block of memory 😉

@DHowett commented on GitHub (Jul 20, 2021): We will need to design this carefully to ensure that it works with the new column counting design in #8000, as well as can fault rows out to larger storage if they need to violate the implied "one `wchar_t` per column" invariant. I'd rather use something a little less efficient but a little more "hands-off" like pmr and memory resources instead of having to come up with a defragmenting allocator to reign over a single block of memory :wink:
claunia commented 2026-01-31 04:14:51 +00:00
Author
Owner
Copy Link

@lhecker commented on GitHub (Apr 29, 2023):

This was fixed in #13626

@lhecker commented on GitHub (Apr 29, 2023): This was fixed in #13626
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-Output Issue-Task Needs-Tag-Fix Product-Conhost
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#14609
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?