Squeak-dev February 1999

squeak-dev@lists.squeakfoundation.org

148 participants
356 discussions

total newbie to Squeak
by Benjamin Shults 26 Feb '99

26 Feb '99

I can write methods in Squeak to extend the behavior of existing objects. Does the Squeak GUI provide a similar way to write new classes from scratch (inheriting from an existing class, of course?) Benji -- Benjamin Shults http://www2.kenyon.edu/People/shultsb/ Department of Mathematics mailto:shultsb@kenyon.edu Kenyon College Voice: 740 427 5817 Gambier, OH 43022 USA Office: Ascension 22

3 2

Heckbert's algorithm
by Andres Valloud 26 Feb '99

26 Feb '99

Hi. I forgot to mention, Color's #red, #green and #blue take more than 36% of the execution time!!! Here, reimplemented with #bitShift: and no method indirection, each of those take between 6 and 8 microseconds. Andres.

1 0

Color buried treasure
by Andres Valloud 26 Feb '99

26 Feb '99

Hi. I've found a #>> send in Color in a private method... it's about 4 times slower than #bitShift:, so I first thought about replacing #>> with #bitShift:, and then I just thought about using //... so... Taking advantage of so much VM talk, would it be troublesome to support fast smallinteger shifting, like making #bitShift: hardwired as $+ and so? It's 1 or 2 cpu cycles versus 39 (intel 32 bit division and multiplication take at least that, division is slower, perhaps with fast out determination for multiplication). And here's the fileout for the changed method in Color. I noticed they were slow checking them while trying to speed up Heckbert's quantization. Hehe, it works! Right now it quantizates a 512x384x32 bit image with 630 colors down to 256 colors including building the representatives in 2.469s (200mmx). I still have to build the renderer to change all colors in the source to those quantizated by Heckbert's algorithm. Andres. Content-Type: application/x-unknown-content-type-st_auto_file; name="Color-privateGreen.st" Content-Disposition: inline; filename="Color-privateGreen.st" Attachment converted: Anon:Color-privateGreen.st (????/----) (000072BF)

1 0

My SGI build is probably good enough.
by Bill Cattey 26 Feb '99

26 Feb '99

I decided to run some real benchmarks and find out whether or not I should be worried about the issue of performance with the SGI VM I built. Here are the summaries: VM Make CPU OS Slop Smop 2.3 Sun Ultra-10/300 Solaris 5.6 1.834 1.177 2.3 Sun Ultra-10/300 Solaris 2.5.1 1.649 1.185 2.3 SGI O2/200 IRIX 6.3 0.847 0.679 2.2 SGI O2/200 IRIX 5.3 0.790 0.622 2.2 SGI O2/200 IRIX 6.3 0.751 0.572 Comparison: Sun by Ian vs wdc: wdc Slop +11% Smop -1% SGI by Ian vs wdc: wdc Slop -5% Smop -9% Bottom line: When I build Squeak 2.2 the same way I built 2.3, I got a little less performance, but less than 10%, so I'm not going to worry about it. The VM I built using the egcs compiler is a little better at simple things, but a teeny bit worse at complex things. But there again, perhaps not enough to worry about. -wdc

1 0

Testing for modifiers in Morphic?
by Bert Freudenberg 26 Feb '99

26 Feb '99

I'd like to have a left-shift-hack in a morph's step method. But it doesn't work, at least not like expected: When shift is pressed, nothing happens (that is, Sensor shiftPressed answers false). If a key or mouse button is pressed while holding shift, it flips to true and remains like this even if shift is released. It flips back to false only when the next "unshifted" event is received. (For testing, I inserted Transcript show: Sensor shiftPressed printString; cr. into the morphs step method. Asking the hand for its last event is equivalent to this method.) What's going on? /bert -- Bert Freudenberg Department of Simulation and mailto:bert@isg.cs.uni-magdeburg.de Computer Graphics http://isgwww.cs.uni-magdeburg.de/isg/bert.html Univ. of Magdeburg

2 1

Another Smalltalk VM architecture
by Allen Wirfs-Brock 26 Feb '99

26 Feb '99

The recent interesting discussion of performance and vm architectures motivated to get permission to publish the specification of one of the architectures I developed over the years. It can be found at: http://www.smalltalksystems.com/publications/avmarch.pdf The following is from the document's introduction: This document describes a bytecode architecture for an experimental Smalltalk virtual machine. The architecture was originally developed by the authors at Digitalk Inc. in 1993-1994. This document is published with the permission of ObjectShare Inc, the successor company to Digitalk. This architecture is intended to be used for the representation of compiled method in the context of dynamic translating ("JIT") virtual machine. As such, its main purpose is to provide information that can be easily and efficiently processed by such a translator. An important design goal was the minimization of the space needed to represent compiled methods. Because it was design to be the input to a translator, minimization of decoding time was not a consideration of the design. It was not intended for direct interpretation and certain features, such as the "Label" construct would be inefficient if used by an interpreter. Allen_Wirfs-Brock(a)Instantiations.com

2 1

Re: total newbie to Squeak
by Thomas Fahey 25 Feb '99

25 Feb '99

Benjamin Shults wrote: > > I can write methods in Squeak to extend the behavior of > existing objects. Does the Squeak GUI provide a similar > way to write new classes from scratch (inheriting from > an existing class, of course?) > > Benji Just take the sample class "template" that you see in the SystemBrowser when you select a class category (this one happened to be the category named 'Network-Protocols'): Object subclass: #NameOfClass instanceVariableNames: 'instVarName1 instVarName2' classVariableNames: 'ClassVarName1 ClassVarName2' poolDictionaries: '' category: 'Network-Protocols' Change 'NameOfClass' to some new class name that doesn't exist in the system, change the instVarName1... and ClassVarName1... names to whatever you want (or delete them). Then choose 'accept' from the popup menu and, voila', you have a new class. BTW, you can also change 'Object' to whatever superclass you want to use, as long as it already exists. Enjoy! == Tom Fahey First Union Corporation Charlotte, NC GO SMALLTALK! _________________________________________________________ DO YOU YAHOO!? Get your free @yahoo.com address at http://mail.yahoo.com

1 0

Converting VisualWorks based code to use Squeak
by Thomas A Peterson 25 Feb '99

25 Feb '99

Does anyone have any experience with converting a VisualWorks based application to be Squeak based? I am interested in finding out what would be the major hurdles to performing such an activity. Thanks, Tom

2 1

For the curious...
by Ian Piumarta 25 Feb '99

25 Feb '99

....the following is temporarily available on alix: ftp://alix.inria.fr/pub/squeak/unix/Jitter-2.990225.2055 It's configured for Linux (only), and should not yet be considered as suitable for general consumption. VM hackers might be interested though. The README contains build instructions. Enjoy! Ian ------------------------------- projet SOR ------------------------------- Ian Piumarta, INRIA Rocquencourt, Internet: Ian.Piumarta(a)inria.fr BP105, 78153 Le Chesnay Cedex, FRANCE Voice: +33 1 39 63 52 87 ----------------------- Systemes a Objets Repartis -----------------------

1 0

Re: Optimizing Squeak
by Ian Piumarta 25 Feb '99

25 Feb '99

> Fundamentally, in threaded code, the execution tokens are the addresses > of the interpretive routines [...] > So the core of the interpreter dispatch loop becomes trivial -- and > very fast. The crux of threaded code is that the interpreter loop goes away entirely. Each "opcode" ends by dispatching to the next opcode, so there is no return to a central loop -- e.g: #define DISPATCH() { \ register void *nextOp= (void *)fetchWord(); \ goto *nextOp; \ } pushReceiver: internalPush(receiver); DISPATCH(); pushTrue: internalPush(trueObj); DISPATCH(); pushFalse: internalPush(falseObj); DISPATCH(); and so on. (The addresses of the labels are the "opcodes" used to represent a program.) The word "threaded" comes from an analogy with sewing: execution flows through the opcode implementations, like a needle pulling thread. There is a truly excellent short article that I highly recommend to the archaeologists amongst us: James R.\@ Bell, Threaded Code, Communications of the ACM, 16(6):370--372, 1973. Ian

6 5

Jump to page: