... Slightly off-topic but I was involved in a discussion about Java recently and it appears you cannot do this in Java :( Anyone disagree? volatile int...
Dave Cunningham
spark@...
Jun 1, 2006 3:26 pm
132
... There was confirmation of that a while ago on the concurrency- ... There was then a long discussion, after which it was concluded that there was no clever...
... ahhah ... For the ring buffer in question, you would use an AtomicIntegerArray, no? If you wanted to process 8 bit values you might have to waste a lot of...
Dave Cunningham
spark@...
Jun 1, 2006 7:24 pm
134
... I thought making the offsets volatile sufficed? Cheers, Chris...
Bob or Chris: How about posting the test source code somewhere (the yahoogroups file repository for lack of a better place)... then we can all check it out and...
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the liblf-dev group. File :...
liblf-dev@yahoogroups...
Jun 2, 2006 9:32 am
137
It has run for three days without an event. I will leave it running on my dual G5 (OSX) for a month. Bob ... -- AMSAT VP Engineering. Member: ARRL, AMSAT-DL,...
... On my dual G5 it dies instantly every time: [James-McCartneys-Computer:~/downloads/ring-buffer] jamesmcc% ./run wr0 0 queue-worker.h:39: failed assertion...
Hmm... So why isn't Bob's failing? Is it possible you two have different CPUs / memory implementations? (me knowing nothing about G5 Macintoshes) apparently...
This reply is a bit out of date, but could you explain the following advice. I've seen that line in kernel code and not yet understood why the "memory" part is...
It appears to be a "magic" instruction to GCC telling it not to reorder memory accesses across the instruction. (GCC doesn't try to understand assembler; it...
Woops, this has been in my outbox for a while.... ... Interestingly, Alphas have the worst ordering or any of the popular architectures[1], but it's not...
I am not surprised that the ring-buffer test fails to show any signs of store/load reordering. The code is pretty complex, so I may be mistaken, but it seems...
... Actually, the value written is incremented every time, making it trivial to spot data reordering. ... I don't believe so, though have no assurances on that...
... What I thought was that the same data was always written to the same memory location in the buffer, not that the same data was always pushed into the...
Hey all, I have a situation which is best described in pseudo-code: thread one (high performance not required): loop ... if( some unusual condition ) ...
Dear Abby: I am disgusted that yahoo has chosen to remove the "no-html" delivery option and has begun to cover us up with the crap. At least (temporarily?)...
Yeah, it's not so good.. looks like it might be time to look elsewhere for a list provider.. R. ... From: Robert McGwier To: softrock40@yahoogroups.com ;...
... votes++ gmail is probably worth considering -- joq...
Jack O'Quin
joq@...
Jun 13, 2006 3:06 pm
151
Richard and others: Thank you for pointing this out and I look forward to using it. http://groups.yahoo.com/mygroups is the page and in the upper right hand...
James and Robert, Did you guys discuss this privately? James: perhaps you could send your binary to Robert? Best wishes Ross. ... From: "James McCartney"...
We did not discuss this privately. The results are straightforward and have been reported elsewhere in this forum by another person who has run the code....
Hi Bob Ah, I must have missed your post saying you had it fail on your machine (I have been having some email problems lately)... I agree that it should be a...
One of the more helpful architectures I've come across actually provides explicit barrier instructions for every circumstance (read before read, read before...
Technique for passing data to other threads requiring only atomic write and atomic read. http://www.ddj.com/dept/cpp/189401457 Each pair of threads needs a...
Note that the article appears to be using "lock-free" to mean "without locks", not the technical definition of "guaranteed progress". The reader can be blocked...
... <chris.purcell.39@...> wrote: Hi there- new Mad Computer Scientist (Member in good standing of the Mad Scientist's Union, Local 42... :-) to the list...
... Not true. "Lock-free" means "there exists some bound n such that if any thread takes n steps, the system as a whole will always make progress". ...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
