diff options
author | Koundinya Veluri <kouvel@users.noreply.github.com> | 2018-07-16 22:04:07 -0700 |
---|---|---|
committer | GitHub <noreply@github.com> | 2018-07-16 22:04:07 -0700 |
commit | 6b403ca4422f2bf3df9d25a32790cc4c0d4b6ee4 (patch) | |
tree | 74e1aecfcbd2054d410555217deae83b0d901a7e /src/vm/ceemain.cpp | |
parent | 2a4db3b265eccb9d56e3706d260237b3687da9ca (diff) | |
download | coreclr-6b403ca4422f2bf3df9d25a32790cc4c0d4b6ee4.tar.gz coreclr-6b403ca4422f2bf3df9d25a32790cc4c0d4b6ee4.tar.bz2 coreclr-6b403ca4422f2bf3df9d25a32790cc4c0d4b6ee4.zip |
Apply tiering's call counting delay more broadly (#18610)
Apply tiering's call counting delay more broadly
Issues
- When some time passes between process startup and first significant use of the app, startup perf with tiering can be slower because the call counting delay is no longer in effect
- This is especially true when the process is affinitized to one cpu
Fixes
- Initiate and prolong the call counting delay upon tier 0 activity (jitting or r2r code lookup for a new method)
- Stop call counting for a called method when the delay is in effect
- Stop (and don't start) tier 1 jitting when the delay is in effect
- After the delay resume call counting and tier 1 jitting
- If the process is affinitized to one cpu at process startup, multiply the delay by 10
No change in benchmarks.
Diffstat (limited to 'src/vm/ceemain.cpp')
-rw-r--r-- | src/vm/ceemain.cpp | 7 |
1 files changed, 0 insertions, 7 deletions
diff --git a/src/vm/ceemain.cpp b/src/vm/ceemain.cpp index fddc12768f..fb1832a83f 100644 --- a/src/vm/ceemain.cpp +++ b/src/vm/ceemain.cpp @@ -1095,13 +1095,6 @@ void EEStartupHelper(COINITIEE fFlags) #ifndef CROSSGEN_COMPILE -#ifdef FEATURE_TIERED_COMPILATION - if (g_pConfig->TieredCompilation()) - { - SystemDomain::System()->DefaultDomain()->GetTieredCompilationManager()->InitiateTier1CountingDelay(); - } -#endif - #ifdef _DEBUG //if g_fEEStarted was false when we loaded the System Module, we did not run ExpandAll on it. In |