One mouse click, 2 PC's

On Mon, 14 Jun 2010 23:10:00 +0000, Tim Watts wibbled:

On an aside, too many interrupts can of course impair system performance. Some of the Intel gigabit ethernet chipsets employs an optional scheme called interrupt-coalescing which, instead of causing an interrupt on every packet received (for example), can instead buffer incoming packets and fire an interrupt once every n-packets. At least one of the linux ethernet device drivers even has an option to switch it into polled handling for high throughput (note, not low latency) devices as it is deemed to give a net gain in efficiency for some scenarios.

--
Tim Watts

Managers, politicians and environmentalists: Nature's carbon buffer.

Windows does this as well -- the interrupt service routine is *supposed* to basically "shut the device up" and then queue up a deferred procedure call (DPCs)... although there's nothing that actually forces you to do this, other than customers complaining that your device driver makes their mouse pointer unresponsive or whatever. :-)

I've been told that on heavily-loaded server systems (where the various execution quanta are pretty long relative to desktop systems) DPC latency is often easily in the hundreds of milliseconds range and can actually break one second if you're particularly unlucky.

---Joel

The Windows (NT and later) parallel port driver is purportedly like this: While a parallel port can generate interrupts, these days many printers will accept data bytes very rapidly (e.g., >1MBps) until their buffers are full, so getting an interrupt for each data byte consumes massive CPU time.

I've seen a fair amount of hardware that provides some option for, "generate an interrupt when the FIFO is half-full *OR* as soon as there's ANY data in it

*and a certain timeout has elapsed since the last byte was received*" as a means of trying to provide high-throughput without slamming the system. (Indeed, you see the same sort of thing with block-based protocols such as USB: In USB to serial converters, they'll usually wait a bit after receiving the first serial character to transmit, in the hopes that more will come since getting on the USB bus at all means they'll be sending, e.g., 64 bytes and it'll be largely wasted if there's only a few bytes of payload data.)

---Joel

cies

as

e

I disagree. I don't think you can really characterize either one (interrupt or polled) in too general terms. Much depends on how long the routine (or routines) will be before control is returned to the main program. For example, some software disables further interrupts until the existing interrupt is completed, etc... Some just do a quick flip of a few registers and return control. Others might commence length file operations, etc...

Not starting an argument. I mean, I do get the thrust of your statement. It's probably generally safe to say that polling systems are quite uniform, provided no interrupts are being serviced at the time. Whereas by contrast, something that is interrupt driven could be downright asynchronous! To anything.

That's a good point. I've always been of the mind that interrupt routines should do "relatively little" and return. For microcontroller-based hardware, often that's all that's needed -- but as was also mentioned, for high-falootin' OSes like Windows and Linux it often turns into a two-step approach with deferred procedure calls.

You're certainly correct that some programmers who write perfectly functional code spend a *very* long time in interrupt routines with interrupts disabled. I once wrote some bit-banged serial port routines that ran off of a timer interrupt and was a little annoyed to find that I had to drop the bit rate from my initially envisioned 9600bps down to 1200bps because the guy who wrote the USB stack did this...

Good point. Hopefully you don't have to program too many systems that deviate that far from what I think is the "intended" model of interrupt-driven vs. polled I/O. :-)

---Joel

Hey Greg...

What about those multi-track recorders that musicians use? I've seen several in the ~$200 -ish price range. They are pretty cheap these days, and most will offer more features than you will probably want or need!

I have a Fostek MR-8 Digital Multitracker and I'm reasonably sure it will do what you want. I use it for my keyboards (piano, mostly). The Fostek treats each track as it's own WAV file (might be an MP3, but I think only newer models handle those..?)

It's an older model, uses CF memory (yuk!), but you can probably find a used one on eBay for around $100 or so. On mine, you'd have to double up mono tracks 1 & 2, and also 3 & 4 to make stereo tracks. Or use tracks 5/6 and 7/8, which can be confiured as either mono or stereo.

Anyway, I don't want to suggest too strongly that something like my old Fostek unit is the ultimate way to go. There are almost certainly better models out there.

Try checking out the music stores in your area. Try:

or

You may find an even better (and possibly cheaper?) solution to play/ record multi-source stereo tracks (MP3 source audio).

Good luck.

-mpm

Simply drop a second audio card into your first PC.

There are set-ups out there with four.

OK, but the commercial audio editing software I use (CoolEdit) only generates a stereo signal. If I open a second instance of it, how do I force it to address the second soundcard?

There appears to be nothing in the program itself that will enable this. IOW if you change the default device in one instance, it changes it in all.

Any suggestions?

Another respondent mentioned 4 track recorders. My associate Ken (the OP) said he looked into these and, while they record 4 channels for mixing purposes, only 2 can be output, as a conventional stereo signal. At least for the lower cost models examined.

Greg Hanson

Not being a Cool Edit user, I don't know off-hand, but Cool Edit does claim to support multiple sound cards, so a search through the help files or just the menu options should be revealing.

Note that you probably should prefer a single audio card over several if your goal is to keep everything in lock-step sync: If you use multiple cards, they'll likely each have their own clock generators that will be ever so slightly different in frequency and hence the audio output will slowly get out of sync unless your software is very much on top of this and can insert or delete samples to "fix" it (I'd be surprised if most software had such an option...)

Most soundcards these days support 5.1 audio, so perhaps you could just use front L+R and rear L+R for your four outputs?

---Joel

Even if you had two identical make, model and configuration computers, with Windows on them CLONED to be identical, and somehow simultaneously triggered, there would still be lots of reasons the two machines would never be synchronized.

Even under Linux I would strongly suspect that two IDENTICAL computers could never be perfectly synchronized.

What, you thought that identical computers with identical software would give you IDENTICAL results?

LOL

If all this is for is a benchmark testing of software, then the low tech approach of using a double T bar to tap both left click buttons at the same time would be accurate enough.

Software benchmark testing is NOT down to the nanosecond.

The events timed are not short.

On a one minute process, looking for triggering within nanoseconds is not necessary.

On a 10 minute process, worrying about nanoseconds is absurd.

How long is the event being timed?

Is somebody seeking accuracy that is unrealistic?

Has anybody GENUINELY had a reason to want to fully synchronize two modern PC's?

In the olden days, Tandem computers ran synchronized computers always cross checking results for high reliability BANKING work.

The early space shuttle used to use dual computers checking each others work.

But the OP Ken Ingram is doing BENCH TESTING of software, right?

Make a wooden double T bar with bits of rubber to press the two left mouse buttons.

What other plans did you have?

How exactly are you TIMING this race, by the way? With an egg timer?

Did you REALLY think you had to run the two software packages side by side to do a speed test?

You could run one and watch the start and end time in tenths of a second on the system clock, and then do the same for the other program.

The "other Greg Hanson" in Cedar Rapids, Iowa

Run Linux and one of the many hugely advanced mixing studios that are saturating the community. Lots of FREEdom there.

Just about any live distro has some sound tools, but there are sound and composition specific distros out there where you boot it up, and you have 64 tracks at your disposal. Or whatever.

See above.

Note that switching latencies in such a scenario will be very short with respect to audio frequencies and events, so it should be fine for his intents.

Their not really synched 'lock-step-synch' is likely pretty damned close, like single digit milliseconds.

He said Cooledit does stereo.

Cluster arrangements have to be.

Where 'perfectly' has a definition.

One you failed to define and eluded to being a number which is not easily attainable in any system.

Your crap was pretty much a bunch of mumbling. Learn how to quote the person you are replying to.

You ain't 'all that'. You don't even come close.

I hear a faint whiny sound like a mosquito.

This sort of thing is my field of expertise. I assume you are using Windows since you mentioned Cool Edit.

1/ Are these WAV/MP3 files or does a signal need to be captured in real time? 2/ What version of Windows are you using? 3/ How old is your computer? Do you know if the sound card is 5.1 channel compatible?

--
- Mike

All gmail Usenet retards should be removed from the gene pool!

You are one such retard.

Yes, XP Pro and CoolEDit Pro. The only thing not "Pro" is obviously me :-)

These are files (test signals) produced in CoolEdit's audio tone generator, saved as MP3's and then played back, also from CoolEdit.

Please see above.

We have a number of PC's. The ones in use for this are matching IBM S51 Celerons, about 6-7 years old. But I have access to newer models.

The soundcards I bought for this purpose are Creative Audigy 2's with _5.1 outputs_. However, I can not find how to make CoolEdit feed out anything but 2 channel stereo.

If possible, I assume it involves cutting and pasting signals from the editor to the 4 channel multi-track window, but I see no way to assign each of these to a different 5.1 channel of the soundcard.

Any advice would be much appreciated.

Greg Hanson