hi Stefan

with 3.8 6273

I get | allClasses allClassesSet block |
	allClasses := Smalltalk allClasses.
	block _ [allClassesSet _ allClasses asSet.
			[allClasses do: [:class | allClassesSet remove: class]] timeToRun].
	{block value. block value. block value}
  #(51 48 53)

and nearly the same in 6208

Stef

could you check because something certainly changed.

On 1 oct. 04, at 04:21, Stephan Rudlof wrote:

> Hello Stef,
>
> nice to see this reviewed!
>
> For me there is a dramatic improvement with the cs, copied from a 
> workspace:
>
> Test:
> doIt
> 	| allClasses allClassesSet block |
> 	allClasses := Smalltalk allClasses.
> 	block _ [allClassesSet _ allClasses asSet.
> 			[allClasses do: [:class | allClassesSet remove: class]] timeToRun].
> 	{block value. block value. block value}
>  #(7127 7135 7120) #(7102 7161 7140) #(7341 7353 7300) "before fileIn"	
>  #(251 250 255) #(257 255 250) "after fileIn"
> 	
> before and after filing in this cs.
> To see the problem again just doIt
> 	| allClasses allClassesSet block |
> 	allClasses := Smalltalk allClasses.
> 	block _ [allClassesSet _ allClasses asIdentitySet.
> 			[allClasses do: [:class | allClassesSet remove: class]] timeToRun].
> 	{block value. block value. block value}.
>  #(3677 3681 3664) #(3650 3688 3654) #(3656 3695 3648) "before"
>  #(3662 3686 3663) #(3645 3710 3647) "after"
>
>
> I'm using Squeak3.7#5988.
>
> ducasse at iam.unibe.ch wrote:
>> can somebody else have a look at this enh.
>>
>> I run the "tests"
>>
>> | allClasses allClassesSet block |
>> 	allClasses := Smalltalk allClasses.
>> 	block _ [allClassesSet _ allClasses asSet.
>> 			[allClasses do: [:class | allClassesSet remove: class]] timeToRun].
>> 	{block value. block value. block value}
>> 	
>>
>> before and after filing in this cs.
>> To see the problem again just doIt
>> 	| allClasses allClassesSet block |
>> 	allClasses := Smalltalk allClasses.
>> 	block _ [allClassesSet _ allClasses asIdentitySet.
>> 			[allClasses do: [:class | allClassesSet remove: class]] timeToRun].
>> 	{block value. block value. block value}.
>> 	
>> 	
>> after loading I got
>> 	-> #(64 57 63)
>>
>> To see the problem again just doIt	
>> 	-> #(34 33 39)
>> 	
>>
>
>> So I'm confused because I thought this would be the opposite so I may 
>> be
>> too tired.
>
> For me it is the opposite: e.g. 251 to 3662!
>
> 1. Observations:
>   - My numbers for the first doIt after fileIn (trying to show the
> improvement) seem to correspond to your numbers of *both* doIts in
> magnitude, if you have a GHz machine (I have a slow 366MHz Pentium).
>   - The proportion of your numbers *after* fileIn is similar to the
> proportion of mine *before*, but then mine are *much* slower.
>
> 2. Since you have been tired: the doIts differ, and the cs comment of
> v1.1 says *not* to fileOut the cs to fileIn it later again.
>
> 3. Questions:
>   - What have you gotten before fileIn for both doIts?
>   - What is the previous version stamp of String>>hash in your test
> image? For me it is SqR 8/13/2002 10:52.
>
> 4. Thoughts:
>   - Has there been a change of SystemDictionary>>allClasses? For me it
> returns the classes and not their names (symbols).
>   - Not very probable: Do you have a better identityHash?
>
> My VM:
> squeak -version
> 3.7b-5 #1 Mon Jun 14 18:05:35 CEST 2004 gcc 3.3.3
> Squeak3.7beta of '1 April 2004' [latest update: #5954]
> Linux karl 2.4.26sr #1 Sun May 23 19:28:24 CEST 2004 i686 GNU/Linux
> default plugin location: /usr/local/lib/squeak/3.7b-5/*.so
>
>
> Hope that helps,
> Stephan
>
>>
>> Stef
>>
>
> -- 
> Stephan Rudlof (sr at evolgo.de)
>    "Genius doesn't work on an assembly line basis.
>     You can't simply say, 'Today I will be brilliant.'"
>     -- Kirk, "The Ultimate Computer", stardate 4731.3
>