Bugzilla – Full Text Bug Listing
| Summary: | Low performance while drawing nodes within the Parallel Runtime perspective | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Tools] PTP | Reporter: | Ivan Fernandez <rayban1984> | ||||
| Component: | RM.PBS | Assignee: | Roland Schulz <roland> | ||||
| Status: | CLOSED FIXED | QA Contact: | |||||
| Severity: | enhancement | ||||||
| Priority: | P3 | CC: | roland | ||||
| Version: | unspecified | ||||||
| Target Milestone: | --- | ||||||
| Hardware: | PC | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
Created attachment 173005 [details]
Patch AbstractRuntimeResourceManager
The problem is that for every single node addNodes(machine, newNodes);
is executed and it shouldn't. It is faster if it is executed only once
but with more nodes in newNodes.
Ivan, could you please add the following declaration to this bug report:
Also please add the following statement to bugzilla:
I, {name}, declare that:
1) This code contains no cryptography.
2) The code has been developed from scratch (i.e. without incorporating content
from elsewhere or relying on the intellectual property of others).
3) All code in the contribution is licensed under the EPL.
thanks
I, Iván Fernández Hernández, declare that: 1) This code contains no cryptography. 2) The code has been developed from scratch (i.e. without incorporating content from elsewhere or relying on the intellectual property of others). 3) All code in the contribution is licensed under the EPL. Applied to HEAD and 4.0 branch. Thanks! |
Build Identifier: I20100312-1448 Hi all, We have been doing some tests with the PBS Proxy (in our Juropa supercomputer) and the LML Proxy (in a fictional System like BlueGene/P), and we could experience a slow performance in the drawing of the node elements and job elements on the screen (a user filter for jobs was not used). The reason is that one we commented a month ago, public void handleEvent(IRuntimeNewNodeEvent e) {, within the AbstractRuntimeResourceManager class. The problem is that for every single node addNodes(machine, newNodes); is executed and it shouldn't. It is faster if it is executed only once but with more nodes in newNodes. Let's try in your code this modification. Kindly regards, Reproducible: Always