Yes, already have those. The 'handle 100MHz' has caveats, as I said, capture an edge to 32 bits, and you find the XMOS part has only 16b port timers, or you find they are data-streaming captures, which choke the SW above a certain (rather low) frequency/edge rate.
I guess they focused on data-streaming apps, and forgot about more general timing problems.
Which is a shame, as all the blocks are there on silicon, just poorly conceived connections driven by a belief SW can solve anything...
Take timing - both XMOS and Prop have mutiple timers, but can they capture the time of a pin edge, to (eg) 32 bits, to the system clock precision ?
The Propeller can with one of its counters. It can also measure the width of a pulse or the coincident or anticoincident width of two pulses to that precision.
It's not like the other processors are going away is it? Nothing here is mutually exclusive. The same attributes that cost, in terms of the "software can't do everything" view expressed above, which is a perfectly rational view mind you, gain in other interesting ways.
Because with just a small amount of forethought, and essentially zero cost to the silicon, I could have the best of both worlds...
A software UART I can tolerate, because the ceiling is less an issue, but with timers there is really no excuse to think SW can replace HW.
The idea of threads is great, which is why I keep checking what these devices can do.
The collective lightbulb is slowly turning on in the industry, as timers do improve. (eg I see NXP have added Clear-On-Capture ).
Why does that detail matter ?
Imagine capture of a high edge rate, with COC you KNOW you have the most recent single pulse width (tho you may have missed some, in the SW read latency ), without that COC you have the issue of not being really sure how many edges occured since last capture.
The 'handle 100MHz' has caveats, as I said, capture an edge to 32 bits, and you find the XMOS part has only 16b port timers, or you find they are data-streaming captures, which choke the SW above a certain (rather low) frequency/edge rate.
This illustrates precisely why I so dislike Pollyanna recommendations from individuals who have never "been there" or "done that."
P2 is going to put a lot of that kind of power on the I/O. Based on Chip's past writings, I think your desire is well aligned with his overall design goals too. Things will look up, I am sure. Remember, this niche is still young, with a lot of boot-strapping done now. Many things are possible and practical, but there is a long way to go to realize real maturity, like we see in more traditional micros.
The Propeller can with one of its counters. It can also measure the width of a pulse or the coincident or anticoincident width of two pulses to that precision.
Do you mean time-interval, between two pins ?
Do you have any examples ? AN001 only mentions capture once, in a comment & the block diagrams in AN001 are vague.
We need to capture Period(same edges), or pulse width (opposite edges), into a 'highest precision' time-register, and also Divide/prescale an Input rate, before the capture, to give more SW headroom, where the signal is too fast.
I think the prop can do that prescale (PinA/N) => PinB, (fMAX.PinA = ? ) if we allocate two pins (tolerable) ?
Even nicer is the ability to capture BOTH the Edges count, and the time-delta, 'on the fly', (with no gaps), but even the best CPU timers are not there yet. We can do that in a CPLD, of course.
jmg,
Look at the counters' logic modes. The counting can be gated by any Boolean function of two inputs.
-Phil
Yes, I can see that - but that covers only a sub-set of timing, which is Pulse-width measurement.
Same-Edge capture is not gated, but a true capture, and ideally Boolean gates should be avoided, as they give edge-errors.
I can see POSEDGE and NEGEDGE which I think give one SysCLK pulse per IP edge, so effectively increments at the Pin Frequency, but there is no mention/example of this edge doing a time-capture ?
I'm trying to see the up side of this. Don't they still spell "color" wrong over there? :-)
Bill
Okay, you yanks have started something here. Let's see, you pronounce soLder like you would say soder, you pronounce vuLnerable like you would say vunerable. Seeing that you drop the L's then you should pronounce "color" as coor, correct?
Also aunt is incorrectly pronounced like ant (although you do pronounce aunty correctly) so you seem to drop the U's also, so no wonder colour became corrupted into color. Now entree as a main meal is just plain wrong wrong wrong.
There's no room on this forum for xenophobia! :frown:
@Leon, give it a rest will you, if you want to extol the virtues of XMOS do it somewhere else.
And before you say you were just answering Andre's question, why not just PM him instead?
There's no room on this forum for xenophobia! :frown:
@Leon, give it a rest will you, if you want to extol the virtues of XMOS do it somewhere else.
And before you say you were just answering Andre's question, why not just PM him instead?
Coley
Samuel Johnson said that patriotism is the last resort of the scoundrel.:)
Why should I PM Andre? We wouldn't have got the useful discussion that we've had if I had done that.
So we should start answering posts with PM's? Andre' specifically asked about XMOS, nothing wrong with Leon defending it.
C.W.
Andre's question had to do with why XMOS isn't more widely used and talked about. Seriously, C.W., am I the only one here who can see that Leon is in the perfect position to enlighten us about that? Like many others, he was all fired up to do great things with XMOS. But it never happened. And he won't utter a word as to why. All he does is quote worthless sales literature. Never a mention of the difficulties he encountered or the fact that his fantasies never matched the hard realities.
To answer Andre's troll bait the reason we don't see Xmos game consoles or some sort of media box is because there are better chips than either the Prop or Xmos processors for that sort of application. Xmos knows this and they don't even bother targeting something like game boxes or set-tops. Get seriously into that territory then you butt heads with TI's OMAP and other SOC/GPU solutions by the big players. Even the P2 won't be in it's league.
As with all design decisions, the optimum solution depends on the application. The Propeller might be a better choice than a XMOS chip for one application, whereas an XMOS chip will be better suited to another application. Where XMOS scores is in applications using technologies such as Ethernet, high-speed USB and very high quality audio, they have several design wins in those areas. The Propeller can't compete with it in applications like those.
You should blame Andre for starting this thread, not me! You don't have to read my pearls of wisdom, of course.
I never cease to be amused by the agitation that Leon's even-tempered and factual responses seem to provoke in this forum. He has stated multiple times that the correct choice of controllers depends upon the application and that he has both used and recommended the Propeller on more than one occasion. I respect the fact that he is familiar enough with multiple options to make those choices rationally, rather than seeing every app as a nail because all he has at his disposal is a hammer. Promoting one processor or the other for its appropriate application is one thing. The kind of fanboyism, seen so often here, that promotes the Propeller at the expense of all others, regardless of the app, is quite another. Andre' asked a legitimate question. Despite being pushed, shoved, and taunted, Leon has attempted what I consider to be an even-handed answer. Get over it.
Did I say "oops" in my first post on this thread? The resulting outburst of intolerance is most predictable. Thank you Phil for the voice of reason.
With all due respect to Andre the original question was not so legitimate. Any processor from anywhere is clearly superior to the P2 simply by virtue of the fact that they exist and work and the P2 does not.
P.S. I did P.M. Andre pointing him to further sources of XMOS info rather than post it here. Knowing that there was a mob itching for a lynching.
The only similar fanboys I've encountered are the AVR Freaks. Many of them are still anti-Microchip, but most of them now accept that the various ARM MCUs are superior in many ways to the AVR and the AVR32. It's taken several years, though.
Phil, if you think Leon is "familiar enough with multiple options to make these choices rationally" then I wish you good luck with his advice. May you have as much success with XMOS (and Propeller) as he has had. 'Nuff said.
Comments
Yes, already have those. The 'handle 100MHz' has caveats, as I said, capture an edge to 32 bits, and you find the XMOS part has only 16b port timers, or you find they are data-streaming captures, which choke the SW above a certain (rather low) frequency/edge rate.
I guess they focused on data-streaming apps, and forgot about more general timing problems.
Which is a shame, as all the blocks are there on silicon, just poorly conceived connections driven by a belief SW can solve anything...
-Phil
It's not like the other processors are going away is it? Nothing here is mutually exclusive. The same attributes that cost, in terms of the "software can't do everything" view expressed above, which is a perfectly rational view mind you, gain in other interesting ways.
Because with just a small amount of forethought, and essentially zero cost to the silicon, I could have the best of both worlds...
A software UART I can tolerate, because the ceiling is less an issue, but with timers there is really no excuse to think SW can replace HW.
The idea of threads is great, which is why I keep checking what these devices can do.
The collective lightbulb is slowly turning on in the industry, as timers do improve. (eg I see NXP have added Clear-On-Capture ).
Why does that detail matter ?
Imagine capture of a high edge rate, with COC you KNOW you have the most recent single pulse width (tho you may have missed some, in the SW read latency ), without that COC you have the issue of not being really sure how many edges occured since last capture.
This illustrates precisely why I so dislike Pollyanna recommendations from individuals who have never "been there" or "done that."
P2 is going to put a lot of that kind of power on the I/O. Based on Chip's past writings, I think your desire is well aligned with his overall design goals too. Things will look up, I am sure. Remember, this niche is still young, with a lot of boot-strapping done now. Many things are possible and practical, but there is a long way to go to realize real maturity, like we see in more traditional micros.
Do you have any examples ? AN001 only mentions capture once, in a comment & the block diagrams in AN001 are vague.
We need to capture Period(same edges), or pulse width (opposite edges), into a 'highest precision' time-register, and also Divide/prescale an Input rate, before the capture, to give more SW headroom, where the signal is too fast.
I think the prop can do that prescale (PinA/N) => PinB, (fMAX.PinA = ? ) if we allocate two pins (tolerable) ?
Even nicer is the ability to capture BOTH the Edges count, and the time-delta, 'on the fly', (with no gaps), but even the best CPU timers are not there yet. We can do that in a CPLD, of course.
...or were there, couldn't do it, and are now telling people how great it is.
Look at the counters' logic modes. The counting can be gated by any Boolean function of two inputs.
-Phil
Yes, I can see that - but that covers only a sub-set of timing, which is Pulse-width measurement.
Same-Edge capture is not gated, but a true capture, and ideally Boolean gates should be avoided, as they give edge-errors.
I can see POSEDGE and NEGEDGE which I think give one SysCLK pulse per IP edge, so effectively increments at the Pin Frequency, but there is no mention/example of this edge doing a time-capture ?
What you want to do is possible, and the subject of a separate thread, so we don't hijack this one. You can start it.
-Phil
I'm trying to see the up side of this. Don't they still spell "color" wrong over there? :-)
Bill
Also aunt is incorrectly pronounced like ant (although you do pronounce aunty correctly) so you seem to drop the U's also, so no wonder colour became corrupted into color. Now entree as a main meal is just plain wrong wrong wrong.
Your turn
There's no room on this forum for xenophobia! :frown:
@Leon, give it a rest will you, if you want to extol the virtues of XMOS do it somewhere else.
And before you say you were just answering Andre's question, why not just PM him instead?
Coley
Maybe that's the "Yank Dialect" used in Boston and parts of New York.
We don't all "paak da caa in da yaad" here.
Sure could use some of that popcorn (vegetable oil only) for this thread.
So we should start answering posts with PM's? Andre' specifically asked about XMOS, nothing wrong with Leon defending it.
C.W.
Samuel Johnson said that patriotism is the last resort of the scoundrel.:)
Why should I PM Andre? We wouldn't have got the useful discussion that we've had if I had done that.
Andre's question had to do with why XMOS isn't more widely used and talked about. Seriously, C.W., am I the only one here who can see that Leon is in the perfect position to enlighten us about that? Like many others, he was all fired up to do great things with XMOS. But it never happened. And he won't utter a word as to why. All he does is quote worthless sales literature. Never a mention of the difficulties he encountered or the fact that his fantasies never matched the hard realities.
Yeah, why PM Andre' when you can just spout off about XMOS like you usually do, I am sure you just do it for the wind up.
At every opportunity you tell us how much better the XMOS is than Propeller.
If it's so good then why not actually go and do something with it instead of wasting your time over here.
Am I the only one here who is sick and tired of it?
You should blame Andre for starting this thread, not me! You don't have to read my pearls of wisdom, of course.
http://answers.yahoo.com/question/index?qid=20070909073711AAyBQY6
-Phil
With all due respect to Andre the original question was not so legitimate. Any processor from anywhere is clearly superior to the P2 simply by virtue of the fact that they exist and work and the P2 does not.
P.S. I did P.M. Andre pointing him to further sources of XMOS info rather than post it here. Knowing that there was a mob itching for a lynching.