Smalltalk › Cincom › VisualWorks

[vwnc] [enh] VW7.6 Speeding up the low space process

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

22 messages Options

Paul Baumann

Re: [vwnc] [enh] VW7.6 Speeding up the low space process

That was understood. It was only a question of whether bindings got flushed from GC activity too because the overstated profiler measurements tended to fall on bindings. Bindings stood out most because they normally shouldn't cost much in loops. Bindings were not being flushed because they would have shown different in the profiler.

Paul Baumann

From: [hidden email] [mailto:[hidden email]] On Behalf Of Holger Guhl
Sent: Friday, January 09, 2009 7:13 AM
To: Valloud, Andres
Cc: VWNC
Subject: Re: [vwnc] [enh] VW7.6 Speeding up the low space process

Flushing bindings happens always after compiling/adding classes. See SystemUtils class>>linkSystem, sent by #modifySystem:. So if the process that forces the system into a low space situation is adding classes, it is pretty normal that all bindings are flushed.
Just an idea for a scenario that might impose these troubles...

Holger Guhl
-- 
Senior Consultant * Certified Scrum Master * [hidden email]
Tel: +49 231 9 75 99 21 * Fax: +49 231 9 75 99 20
Georg Heeg eK Dortmund
Handelsregister: Amtsgericht Dortmund  A 12812

This message may contain confidential information and is intended for specific recipients unless explicitly noted otherwise. If you have reason to believe you are not an intended recipient of this message, please delete it and notify the sender. This message may not represent the opinion of IntercontinentalExchange, Inc. (ICE), its subsidiaries or affiliates, and does not constitute a contract or guarantee. Unencrypted electronic mail is not secure and the recipient of this message is expected to provide safeguards from viruses and pursue alternate means of communication where privacy or a binding message is desired.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Eliot Miranda-2

Re: [vwnc] [enh] VW7.6 Speeding up the low space process

Paul,

the profiler can only attribute a sample at a suspension point. It is a high-priority process spinning waiting on a delay. When it proceeds from the delay it samples the pc of the process it is profiling. So it only sees the process where it suspends. There is no suspension point in the incremental GC primitive once the primitive has started. The next suspension point is the send of #value to the binding in the low-space loop. SO what you think is time going into binding is actually time going into the incremental GC primitive being misattributed to bindings because the profiler cannot interrupt teh process in the incremental gc primitive until it has returned from the primitive and sent a message to the binding.

On Fri, Jan 9, 2009 at 1:18 PM, Paul Baumann <[hidden email]> wrote:

That was understood. It was only a question of whether bindings got flushed from GC activity too because the overstated profiler measurements tended to fall on bindings. Bindings stood out most because they normally shouldn't cost much in loops. Bindings were not being flushed because they would have shown different in the profiler.

Paul Baumann

From: [hidden email] [mailto:[hidden email]] On Behalf Of Holger Guhl
Sent: Friday, January 09, 2009 7:13 AM
To: Valloud, Andres
Cc: VWNC

Subject: Re: [vwnc] [enh] VW7.6 Speeding up the low space process
Flushing bindings happens always after compiling/adding classes. See SystemUtils class>>linkSystem, sent by #modifySystem:. So if the process that forces the system into a low space situation is adding classes, it is pretty normal that all bindings are flushed.
Just an idea for a scenario that might impose these troubles...
Holger Guhl
-- 
Senior Consultant * Certified Scrum Master * [hidden email]
Tel: +49 231 9 75 99 21 * Fax: +49 231 9 75 99 20
Georg Heeg eK Dortmund
Handelsregister: Amtsgericht Dortmund  A 12812
This message may contain confidential information and is intended for specific recipients unless explicitly noted otherwise. If you have reason to believe you are not an intended recipient of this message, please delete it and notify the sender. This message may not represent the opinion of IntercontinentalExchange, Inc. (ICE), its subsidiaries or affiliates, and does not constitute a contract or guarantee. Unencrypted electronic mail is not secure and the recipient of this message is expected to provide safeguards from viruses and pursue alternate means of communication where privacy or a binding message is desired.
_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc