You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This task is automatically imported from the old Task Issue Board and it was originally created by Michael Mauderer.
Original issue is here.
Why
As Enso user
**I want the backend engine to defer unnecessary processing of my and standard library sources **
**So that my new and/or existing Enso project is ready faster and the backend engine consumes less memory **
Acceptance Criteria
Scenario:Given empty default project
When it is opened and ready to use
Then the amount of used RSS memory shall be 10% lower than now.
Notes:
We are primarily interested in speeding up the time to process/open the project. The assumption of this issue is: if we avoid doing some work (by consuming less memory) we also speed something up. The correlation isn't 100%, but certainly by allocating less on startup we can't make the startup slower.
Comments:
Jaroslav Tulach reports a new STANDUP for yesterday (2022-04-28):
Progress: Performance investigations ~ 10% of memory saved by not allocating Truffle nodes. Can we use native-image? Dmitry's profile the language server work: #3389 - we need to eliminate the amount of dependencies - also adjust them in the upstream project as well; Review of Marcin's static drop-downs implementation. It should be finished by 2022-04-29.
Next Day: Performance investigation and improvements. (Enso Bot - Apr 29, 2022)
**Jaroslav Tulach** reports a new **🔴 DELAY** for yesterday (2022-04-29):
Possible solutions: A week deadline is certainly more appropriate than a day (Enso Bot - Apr 30, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-04-29):
Progress: ./run --no-backend PR: #3426; ess dependencies of sampler: 3fd211c; Requires another upstream update: apache/netbeans#4052 - shall be part of NetBeans 14 in a month or so It should be finished by 2022-05-06.
Next Day: More of performance investigation and improvements. (Enso Bot - Apr 30, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-02):
Progress: Investigating suggestions.db format. logging "actions" next to .npss file for easy analysis: d882dc8 Talking about statics. Talking about compiler passes. Talking about performance. It should be finished by 2022-05-06.
Next Day: More detailed performance measurements. Focus on compiler passes. (Enso Bot - May 3, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-03):
Progress: MFF UK lecture. Display content of of .npss file: https://docs.google.com/document/d/1vRcM12m3tS1pagnCMOypaWFT6Zra7P6SnX46qSo7dLU/edit + few ideas discussed with Michael for #3389; Analyzing 7s of startup sequence: 431 ms org.enso.syntax.text.Parser - caches are in use, but one(?) file still needs to be parsed; 500ms + 545 ms org.enso.compiler.codegen.IrToTruffle (especially ExpressionProcessor); 1000 ms + 513 ms + 312 ms in org.enso.compiler.context.SuggestionBuilder - composing & sending the message to the IDE
the rest is mostly spent in deserializing caches and resolving imports; re-measure with fix to delay creation of TruffleNodes: 1,5s faster It should be finished by 2022-05-06.
Next Day: Delaying TruffleNode creation results in few failing tests. Gonna analyze them now. (Enso Bot - May 4, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-04):
This task is automatically imported from the old Task Issue Board and it was originally created by Michael Mauderer.
Original issue is here.
Why
As Enso user
**I want the backend engine to defer unnecessary processing of my and standard library sources **
**So that my new and/or existing Enso project is ready faster and the backend engine consumes less memory **
Acceptance Criteria
Notes:
We are primarily interested in speeding up the time to process/open the project. The assumption of this issue is: if we avoid doing some work (by consuming less memory) we also speed something up. The correlation isn't 100%, but certainly by allocating less on startup we can't make the startup slower.
Comments:
Jaroslav Tulach reports a new STANDUP for yesterday (2022-04-28):
Progress: Performance investigations ~ 10% of memory saved by not allocating Truffle nodes. Can we use native-image? Dmitry's profile the language server work: #3389 - we need to eliminate the amount of dependencies - also adjust them in the upstream project as well; Review of Marcin's static drop-downs implementation. It should be finished by 2022-04-29.
Next Day: Performance investigation and improvements. (Enso Bot - Apr 29, 2022)
**Jaroslav Tulach** reports a new **🔴 DELAY** for yesterday (2022-04-29):
Summary: There is 7 days delay in implementation of the Delay creation of Truffle nodes for function bodies (#182024911) task.
It will cause 7 days delay for the delivery of this weekly plan.
Mistakenly entered wrong end date on Thursday
Delay Cause: Wrong initial end date
Possible solutions: A week deadline is certainly more appropriate than a day (Enso Bot - Apr 30, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-04-29):
Progress: ./run --no-backend PR: #3426; ess dependencies of sampler: 3fd211c; Requires another upstream update: apache/netbeans#4052 - shall be part of NetBeans 14 in a month or so It should be finished by 2022-05-06.
Next Day: More of performance investigation and improvements. (Enso Bot - Apr 30, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-02):
Progress: Investigating suggestions.db format. logging "actions" next to .npss file for easy analysis: d882dc8 Talking about statics. Talking about compiler passes. Talking about performance. It should be finished by 2022-05-06.
Next Day: More detailed performance measurements. Focus on compiler passes. (Enso Bot - May 3, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-03):
Progress: MFF UK lecture. Display content of of .npss file: https://docs.google.com/document/d/1vRcM12m3tS1pagnCMOypaWFT6Zra7P6SnX46qSo7dLU/edit + few ideas discussed with Michael for #3389; Analyzing 7s of startup sequence: 431 ms
org.enso.syntax.text.Parser
- caches are in use, but one(?) file still needs to be parsed; 500ms + 545 msorg.enso.compiler.codegen.IrToTruffle
(especiallyExpressionProcessor
); 1000 ms + 513 ms + 312 ms inorg.enso.compiler.context.SuggestionBuilder
- composing & sending the message to the IDENext Day: Delaying TruffleNode creation results in few failing tests. Gonna analyze them now. (Enso Bot - May 4, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-04):
Progress: Polishing TruffleNodes PR: #3429; reporting SuggestionBuilder&Rope user story: https://www.pivotaltracker.com/n/projects/2539304/stories/182072427; 1:1s; Demos It should be finished by 2022-05-06.
Next Day: Delaying TruffleNode creation. (Enso Bot - May 5, 2022)
**Jaroslav Tulach** reports a new **STANDUP** for yesterday (2022-05-05):
Progress: 1.5s out of 7s speedup: #3429; writing unit test in Java 18; Enso language debugger: #3432 It should be finished by 2022-05-06.
Next Day: Infrastructure to profile the engine language server: #3389 (Enso Bot - May 6, 2022)
Regressions seem to be fixed: https://github.com//pull/3441 - we can mark this accepted again, I believe. (jaroslavtulach - May 11, 2022)
The text was updated successfully, but these errors were encountered: