• Open Menu Close Menu
  • Apple
  • Shopping Bag
  • Apple
  • Mac
  • iPad
  • iPhone
  • Watch
  • TV
  • Music
  • Support
  • Search apple.com
  • Shopping Bag

Lists

Open Menu Close Menu
  • Terms and Conditions
  • Lists hosted on this site
  • Email the Postmaster
  • Tips for posting to public mailing lists
Re: ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)


  • Subject: Re: ANN: Rosetta Booster Beta 1.0.1a (reference [NSTask] -launch return)
  • From: Greg Guerin <email@hidden>
  • Date: Mon, 15 Jun 2009 10:51:49 -0700

M Pulis wrote:

On a 4GB machine:

Boot time increased 2.533 seconds or +6.535% for a 39 second boot.

But did it solve the problem?

That is, did "priming the pump" with the execution of an exit(0) ppc process reliably decrease the launch-time of another ppc process, the one actually desired?

I ask this because you wrote:

No. Rosetta is a translator. Please read Tannenbaum, Apple, et al.

Being a translator, it must perform some amount of initial translation of every ppc executable before the translated code can run. What part of 2.533 secs is loading, and what is translation?


It seems clear that some of the initial 2.533 secs at boot time is due to Rosetta's libraries and data simply loading from disk. I think we can safely say that this work won't be repeated for a 2nd ppc process. But an unknown fraction of the 2.533 secs is spent translating the actual exit(0) code, and presumably, some of that will have to be repeated for each ppc executable.

Mitigating this is the observation that the ppc code run before main () is usually a common code sequence, and if Rosetta has half the wits of the Scarecrow (please read L. Frank Baum) it would already have that sequence cached. It is also possible that various ppc libraries would be similarly cached, and thus not translated again. So we are left with one question answered (the cost of pump priming) and one unanswered (the cost unique to each ppc process).

I invite you to apply for further research grants to answer this question: How much does a 2nd ppc executable, being the one whose execution was originally desired, benefit from the primed pump?

A 3rd question is whether one can automate the creation of a suitable "primer" executable, similar enough in its use of libraries to an arbitrary original executable, that the maximum benefit of priming the pump can be obtained. A marketing consultant suggested the name Optimize Primer for this, apparently a tie-in to a popular product related to inductors or transformers.

  -- GG

_______________________________________________

Cocoa-dev mailing list (email@hidden)

Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com

Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden


  • Prev by Date: Re: [iPhone] (newbie) navigation controller is null, combining navigation controller and tab bar controller
  • Next by Date: Re: Core Data, to-many relationships, and object graph consistency
  • Previous by thread: spell checking
  • Next by thread: Font Panel and storing font user default
  • Index(es):
    • Date
    • Thread