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

Significant SPEC CPU2017 perf regression due to ConPTY window visibility #19126

New Issue
Open
opened 2026-01-31 06:34:32 +00:00 by claunia · 0 comments
claunia commented 2026-01-31 06:34:32 +00:00
Owner
Copy Link

Originally created by @lhecker on GitHub (Dec 23, 2022).

Originally assigned to: @DHowett, @lhecker, @zadjii-msft on GitHub.

Steps to reproduce

As per an internal "Spec benchmark QoS observations on Surface and Samsung system with latest OS" discussion:

Here are steps to run the SPEC CPU2017 ST workload:

  1. Open CMD window [not as administrator] and CD into the directory "SPEC.CPU2017_Windows_IC2022.1". You should see the install.bat file in the current working directory [with DIR command].
  2. On the CMD prompt, type install.bat. Confirm the directory question and allow the script to finish. You should see 'success'.
  3. Run the script "run-1copy-1subtest-ic2021.2-win-core-avx2-smt-on-20210430.bat" for a quick 5 minute run. This script will run only test #531 "deepsjeng" for 1 copy and 1 iteration.
  4. When it finishes, in the result directory you should see newly created files such as "CPU2017.001.intrate.refrate.pdf" with the geomean in the top right of the first page highlighted in gray. Look for "SPECrate2017_int_base ".
  5. In the broken state, this score will be about 3 (+/- 0.5) and in the working state it should be about 7 (+/- 0.5)

Additionally,

  1. The above run is all via cmd line automation
  2. The cmd line that runs this is always in focus
  3. Not run as admin
  4. Nothing else is blocking the window. Spec doesn't have any UI associated

Expected Behavior

No response

Actual Behavior

The benchmark process gets assigned a "utility" QoS, due to it seemingly being in the background from the perspective of Windows' "Background Activity Manager" (BAM):

The difference I observed between working and broken set of traces is that WindowsTerminal is being used to launch the benchmark in broken case and cmd.exe is being used in working case. BAM understands when either one of those are in focus.

Benchmarks process is running without a window and BAM and uses parent/child relationship heuristics to determine appropriate PPM policy. In working case benchmark process is descendant of in-focus cmd.exe and that means it inherits its in-focus state.

In broken case though WindowsTerminal is not an ancestor of benchmark process. Benchmark has somewhat long ancestor chain which goes something like explorer.exe->cmd.exe->specinvoke.exe->cmd.exe->specinvoke.exe->benchmark. WindowsTerminal is created by DcomLaunch and it doesn’t launch any other processes.

The cause for this bug (and a potential solution!) is explained in detail in these two comments:

  1. https://github.com/microsoft/terminal/issues/11823#issuecomment-979185489
  2. https://github.com/microsoft/terminal/issues/11823#issuecomment-982193256
Originally created by @lhecker on GitHub (Dec 23, 2022). Originally assigned to: @DHowett, @lhecker, @zadjii-msft on GitHub. ### Steps to reproduce As per an internal "Spec benchmark QoS observations on Surface and Samsung system with latest OS" discussion: > Here are steps to run the SPEC CPU2017 ST workload: > > 1. Open CMD window [not as administrator] and CD into the directory "SPEC.CPU2017_Windows_IC2022.1". You should see the install.bat file in the current working directory [with DIR command]. > 2. On the CMD prompt, type install.bat. Confirm the directory question and allow the script to finish. You should see 'success'. > 3. Run the script "run-1copy-1subtest-ic2021.2-win-core-avx2-smt-on-20210430.bat" for a quick 5 minute run. This script will run only test #531 "deepsjeng" for 1 copy and 1 iteration. > 4. When it finishes, in the result directory you should see newly created files such as "CPU2017.001.intrate.refrate.pdf" with the geomean in the top right of the first page highlighted in gray. Look for "SPECrate2017_int_base ". > 5. In the broken state, this score will be about 3 (+/- 0.5) and in the working state it should be about 7 (+/- 0.5) > > Additionally, > 1. The above run is all via cmd line automation > 2. The cmd line that runs this is always in focus > 3. Not run as admin > 4. Nothing else is blocking the window. Spec doesn't have any UI associated ### Expected Behavior _No response_ ### Actual Behavior The benchmark process gets assigned a "utility" QoS, due to it seemingly being in the background from the perspective of Windows' "Background Activity Manager" (BAM): > The difference I observed between working and broken set of traces is that WindowsTerminal is being used to launch the benchmark in broken case and cmd.exe is being used in working case. BAM understands when either one of those are in focus. > > Benchmarks process is running without a window and BAM and uses parent/child relationship heuristics to determine appropriate PPM policy. In working case benchmark process is descendant of in-focus cmd.exe and that means it inherits its in-focus state. > > In broken case though WindowsTerminal is not an ancestor of benchmark process. Benchmark has somewhat long ancestor chain which goes something like explorer.exe->cmd.exe->specinvoke.exe->cmd.exe->specinvoke.exe->benchmark. WindowsTerminal is created by DcomLaunch and it doesn’t launch any other processes. The cause for this bug (and a potential solution!) is explained in detail in these two comments: 1. https://github.com/microsoft/terminal/issues/11823#issuecomment-979185489 2. https://github.com/microsoft/terminal/issues/11823#issuecomment-982193256
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-DefApp Area-Quality Issue-Bug Needs-Tag-Fix Priority-1 Product-Conpty Product-Terminal
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#19126
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?