Squeak-dev April 1998

squeak-dev@lists.squeakfoundation.org

107 participants
279 discussions

Re: Re[2]: Squeak and SMP
by Ken Dickey 25 Apr '98

25 Apr '98

>At 1:54 PM -0700 4/24/98, Ken Dickey wrote: > >>GC can be done at first by stopping all (non-gc) threads. This works to >>get things up. > >Then you violate real time that the original poster was after. Agreed. I was thinking of staged ways to get the above up and running incrementally. I guess when it comes from implementing stuff I just do better doing it in stages and having a working system at each stage. (Hey, it works for me. ;^) >>Then you can go to a concurrent collector (too big to >>discuss here). > >Do you have a line on a real time concurrent GC that is really >real time? (Some algorithms have been claimed to be real time >which aren't.) I would suggest looking at Springer Lecture Notes 637 and 986 (International Workshop In Memory Management '92 and '95) as well as Paul Wilson's survey. There are a bunch of papers with various assumption sets. I agree that many so called real-time collectors aren't. The implementation of Baker's Treadmill by Wade Hennesy used in Kaleida's ScriptX multi-media runtime is real-time. It really depends on what trade-offs you want to make. It really is too detailed a discussion for a quick answer. >> Also, if you have a 2-word compare&swap, you can do 'lockless' >> concurrent queueing (check a newer OS textbook for algorithms). >This seems neat. This is kind of like a "linear object" providing >unique (solitary) access? > >Do you have a good OS textbook you recommend, Ken? I originally learned about 'lockless queueing' reading about the Synthesis kernel, which has some really cool ideas, including specialization by runtime code generation. Henry Massalin "Synthesis: An Efficient Implementation of Fundamental Operating System Services", Columbia U Ph.D. Thesis, 1992. I have not bought any OS books lately, but note that compare&swap techniques are covered in Gray & Reuter's "Transaction Processing, Concepts and Techniques" (Morgan Kaufmann 1993) which you probably have gathering dust on your bookshelf, Patrick. 8^) Cheers, -Ken

1 0

Re: bit shifting
by Mike Klein 25 Apr '98

25 Apr '98

I think that sending highBit to a negative number should be an error, (or return an Infinite value) if we are going to expose 2's complement for bit operations. -- Mike Klein

1 0

Re: Re[2]: Squeak and SMP
by Ken Dickey 25 Apr '98

25 Apr '98

>Gemstone Java uses native threads to implement the JVM. Most of the >details are proprietary, unfortunately. I am not sure where the >boundaries are. Well, it is easy to bring up a browser and look in the "Kernel-Processes" area to see the things that have to be mapped. Basically the callouts become specialized primops. The hard part is verifying the threading assumptions about exclusion. Typically, there is a 2-step using an exclusion global flag which is accessed atomically (compare&swap or on older systems test&set). If the flag is unset, just do whatever, else you must call the underlying OS monitor/other exclusion method. The basic idea is that the flag is cheap and you don't have many collisions (e.g. on a uniprocessor) so you only make the expensive call where required. You can start without this. but it can make a big difference in performance. Also, if you have a 2-word compare&swap, you can do 'lockless' concurrent queueing (check a newer OS textbook for algorithms). GC can be done at first by stopping all (non-gc) threads. This works to get things up. Then you can go to a concurrent collector (too big to discuss here). Cheers, -Ken

2 1

Re: File times & 1.31 (win32)
by Andreas Raab 24 Apr '98

24 Apr '98

Steve, > When I open a FileList I noticed that the times reported for each file do > not match the times as displayed by the OS(Win95). I find that the time > that is reported in the FileList is out by exactly 6 hours. [...] > I find it interesting that the amount of the discrepancy exactly matches > our 6 hour time zone correction from GMT. Does/should the 1.31 Win32 > implementation take into account time zone differences? It didn't - I've just fixed it. Stay tuned for Squeak 2.0 - the Windows VM will have a number of fixes and improvements at this time. Andreas -- Linear algebra is your friend - Trigonometry is your enemy. +===== Andreas Raab ============= (raab(a)isg.cs.uni-magdeburg.de) =====+ I Department of Simulation and Graphics Phone: +49 391 671 8065 I I University of Magdeburg, Germany Fax: +49 391 671 1164 I +=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+

1 0

Re: Squeak and SMP
by James McCartney 24 Apr '98

24 Apr '98

At 10:02 AM -0700 4/22/98, Jim Ravan wrote: >One of my interests in Squeak is to implement SMP support. The eventual >target application is realtime generation of music. Can someone point a >newbie at existing research in this area? The target hardware would >probably be a two-processor Intel P2 box. The hard part will be SMP AND real time GC. I know of no good SMP and real time garbage collection algorithm. The only one that I've seen is in a paper by Herlihy and Moss, but would be very inefficient since it requires making a copy of an object every time it is modified. Thus would only seem to be useful for functional languages. --- james mccartney james(a)clyde.as.utexas.edu james(a)lcsaudio.com If you have a PowerMac check out SuperCollider, a real time synth program: http://www.audiosynth.com

3 7

Re: Squeak and SMP
by Dwight Hughes 24 Apr '98

24 Apr '98

[warning: the following may be entirely blue sky , but it is very *nice* blue sky <g>] "Concurrent Compacting Garbage Collection of a Persistent Heap" by James O'Toole, Scott Nettles, and David Gifford. December 1993. and "Concurrent Replicating Garbage Collection." by James O'Toole and Scott Nettles. April 1993. http://www.psrg.lcs.mit.edu/publications.html present a possible approach (real time GC is covered by other papers at this site). I very much want to try to convert Squeak over to the replicating GC approach presented here (real life has proven extremely time consuming lately however, sigh). Replication GC should be capable of at least soft real time constraints since it has no (incremental) time dependence on object size or number of objects and its incremental time slices can be tuned as needed. -- Dwight -----Original Message----- From: James McCartney <james(a)clyde.as.utexas.edu> To: squeak(a)cs.uiuc.edu <squeak(a)cs.uiuc.edu> Date: Friday, April 24, 1998 9:50 PM Subject: Re: Re[2]: Squeak and SMP [--snip--] > >Do you have a line on a real time concurrent GC that is really >real time? (Some algorithms have been claimed to be real time >which aren't.) > > > --- james mccartney james(a)clyde.as.utexas.edu james(a)lcsaudio.com

1 0

Re: bit shifting
by Christopher Oliver 24 Apr '98

24 Apr '98

Martin McClure <martin(a)hand2mouse.com> wrote: > I can think of historical and performance reasons for allowing bit > manipulations on integers, but I'd be interested to hear any *other* > reasons why people find this a good idea. Read any good cryptology papers lately? -- Christopher Oliver Traverse Internet Systems Coordinator 223 Grandview Pkwy, Suite 108 oliver(a)traverse.net Traverse City, Michigan, 49684 let magic f = fun x -> x and more_magic n f = fun x -> f ((n f) x);;

1 0

Re: QuickTime 3 support in Squeak
by Dan Ingalls 24 Apr '98

24 Apr '98

Philippe de Rochambeau <pr1(a)club-internet.fr> wrote... >I think that QuickTime 3 support in Squeak (at least in the Windows and >Macintosh versions) would make it attractive to multimedia designers. What do >you all think? This is a double-edged sword. It would compromise Squeak's ultra portability to depend on QT, since then a port to a bare chip would have to eschew everything built on QT. It is our goal to make Squeak nearly as nice without it, and hopefully better in a number of ways. However, there is no doubt that QT is a powerful facility and a growing standard. I've been hoping that one of these high-leverage external resources (QT, databases, 3D, native widgets) would galvanize some of our troops into doing a general interface from Squeak to such native resources. We of the Squeak team will probably do this if no one else does, but several other things are higher on our list right now. >P.S. Do any of you know when Squeak 2 will be available? To quote from my message of April 9... >A while ago I gave mid-April as a target date for Squeak 2.0. >It is clear now that it will be more like the middle of May instead. >Rather than wait, I thought I'd give you all some warning. That is still looking realistic. We plan to have a beta version out in the first week of May. - Dan

2 1

FFT
by Luciano Esteban Notarfrancesco 24 Apr '98

24 Apr '98

I found some odd behavior in FFT, and I think it's a bug. Inspect (FFT new nu: 8; realData: ((0 to: 255) collect: [ :each | (2 * Float pi * each / 255) cos]); transformForward: true. and take a look at realData. Since the original data is real, the real part of the Fourier transform must be symmetric, so realData first should equal realData last, but it does not happen, and instead of it realData second (!) equals realData last. Luciano.-

1 0

New open source paper by Eric S. Raymond
by John Dougan 24 Apr '98

24 Apr '98

'Homesteading the Noosphere' is a new paper by esr following up the regarded 'Cathedral and the Bazzar' paper. In it he discusses more closely the open source culture and the whys and hows of it's workings. It is located at: <URL:http://www.earthspace.net/~esr/writings/homesteading/> For Squeak I think that this paper has some valuable insight on project management, and how we'll want to develop Squeak for maximum progress. --john -------------------- John Dougan jdougan(a)kitsnet.vancouver.bc.ca Phone (Home): (604)253-5140 Owner, Researcher, Network Manager: Kitsilano Network Research. -- If it's not Boeing, I'm not going. --

1 0

← Newer
1
...
5
6
7
8
9
10
11
...
28
Older →

Jump to page: