It makes sense for Andre' to look at the various technologies on the market for his game console designs. I'd like to see him develop something ARM based.
Given the high-level of abstraction available on the more powerful ARMs, I don't know if a "complete" game console is really a good target. To a certain degree, you might as well just create your games on an iPod Touch. But some form of ARM-based educational package would be cool.
I'm always interested enough to listen to opinions and experiences etc about other micros, the XMOS included. Sometimes it's inappropriate to push it but this thread is all about XMOS vs P2 so I don't see the problem. Just as Phil says, so I agree, because as great a fan I am of the Prop it is not my only chip of choice, it could never be if you design commercially (though you have to really push to budge me from the Prop).
Leon, I've looked hard at the XMOS and I could lean toward toward it but I also know that with any micro (and more so XMOS) that there is a learning curve, which means time. So any tips, any advice, any pitfalls that you can pass on to the uninitiated would be appreciated. Thanks mate.
You first need to get a suitable board to play with, which depends to some extent on your interests; none of them are particularly expensive.
The $99 XC-1A is nice because it is self-contained with a USB interface, has a 1600 MIPS four-core XS1-G4 chip, LEDs, buttons, and a little speaker. It also has plenty of I/O. I'd get one of those if you are primarily interested in parallel programming techniques. You can either connect external PCBs to it with ribbon cables pugged onto headers soldered into one of the four sets of pads, or design a PCB that plugs into a connector soldered to the pads running across the board. I've used both techniques. The prototyping area could also be used. It could be used with a solderless breadboard. I actually use the earlier XC-1. They get rather hot: I can't keep my finger on the XMOS chip for more than a couple of seconds, but don't worry about it. The older XTAG interface built-in to the XC-1A uses an FTDI chip which has the nice feature of a UART interface to the host PC. The XTAG-2 used with most of the other boards doesn't have a similar interface, I think that XMOS will be adding something like it. I've just tried the UART interface with the latest version of the tools, and it works very well at 115112 baud. A module (module_uart.1v1) is provided which makes coding the UART function quite trivial:
#include <xs1.h>
#include <platform.h>
#include "uart.h"
on stdcore[0]: port p_rx = PORT_UART_RX;
on stdcore[0]: port p_tx = PORT_UART_TX;
/* This test runs the uart at the uart module's default bit rate of 115200 */
/* Echo any RX word on TX */
void test_uart(void)
{
chan data_ch;
par {
uart_tx(p_tx, data_ch);
{
// Send a message to Tx first
char msg[] = "\n\rHello World\n\rFrom the Simple Uart\n\r";
for (int i=0; i<sizeof(msg); i++) {
unsigned txData = (unsigned) (msg[i]);
data_ch <: txData;
}
// Then echo rx data
uart_rx(p_rx, data_ch);
}
}
}
int main(void) {
par {
on stdcore[0]: test_uart();
}
return 0;
}
It's written in XC, note the par construct. Another program is provided that uses one thread for transmission and another for reception, similar to using two cogs with the Propeller. A long time ago I wrote a very simple UART program, but it stopped working when a new version of the the compiler came out, and I couldn't be bothered to sort it out. It's almost working with the latest version of the compiler, so I'll see if I can fix it.
For motor control, robotics and that sort of thing, you need a $59 XK-1A which uses the single-core XS1-L1. It comes with an XTAG-2 debug/programming interface and a PS. If you need to control more joints, for instance, in your robot, you can simply daisy-chain them together. Individual XK-1As cost $49, and the XTAG-2 on its own costs $19.
Sparkfun sells a $41.95 board with an XS1-L1 on it. Don't buy the XTAG-2 from them, they are stuck with some they bought before XMOS reduced the price and are charging $50.95 for them. The board isn't all that well designed, but it works OK. It has a large prototyping area which makes it ideal for interfacing individual components to it. It's a pity that one can't solder a connector to it as one can with the XC-1A, the pads aren't suitably arranged for that. I suppose that one could be soldered to the prototyping area and wires run across to it, but it would be rather messy. The earlier version of the board (which I have) incorporated a UART function implemented with an FTDI chip, but it never worked properly. I might add a MAX202 to one of my boards and implement a UART in software using the aforesaid module.
The XTAG-2 has an XS1-L1 on it, BTW, which runs the USB interface and the JTAG functions.
Download the development software, of course. It's Eclipse-based and new versions come out quite frequently. You need to be careful if you use the default Eclipse workspace, called "workspace", as other Eclipse-based tools might use it and can cause a conflict. I found that the Altera NIOS II compiler caused problems, so I renamed the workspace to XMOS. Installed applications seem to get corrupted sometimes and don't build; it's best to delete them, and the directory they are in, and download a new file from the XMOS web site. I haven't been able to work out the circumstances in which it happens, so that I can tell XMOS about it. Don't unzip downloaded applications before installing them, just import them into the XDE.
University staff can request $200 worth of free hardware. A friend of mine who runs the embedded systems lab at ESIEE near Paris is using his hardware for an interesting project - half the students will use XMOS kit and the other half will use real-time Java hardware and software. He wants to see which comes out best.
I'll come up with some more stuff as I think of it, but that should get you started.
In the google groops comp.lang.forth thread Intellasys vs XMOS Leon asked for it back in 2008, and was talking to the right bunch, but I never found anything.
In the google groops comp.lang.forth thread Intellasys vs XMOS Leon asked for it back in 2008, and was talking to the right bunch, but I never found anything.
What? Leon is evangelizing XMOS on other forums? That hurts! I thought the Parallax forum was the only one where he is trying to shift revenue from Parallax to XMOS. Now I find out that he's doing this with other companies as well! Googling "Leon XMOS" yields 17,400 hits. 2,970 contain the word Parallax. It's hard to believe he has the time to post comments on other forums. Leon, how much is XMOS paying you, or are you foolish enough to do all of this marketing for them for free? If you're not getting paid by them you should demand a salary equal to that of the VP of Marketing. You seem to be doing his job for him.
"leon parallax propeller" gets 216,000 hits. Perhaps Parallax should pay me something as well. I get PMs from people at Parallax thanking me for my evangelism on their behalf!
"leon parallax propeller" gets 216,000 hits. Perhaps Parallax should pay me something as well. I get PMs from people at Parallax thanking me for my evangelism on their behalf!
Who else is paying you?
Not that it's any of our business, but you do imply someone ....
Too bad. Zero bytes is a little bit too minimal of a kernal size for me. The Green Arrays board is way too expensive, so it looks like the prop is still the only option.
"leon parallax propeller" gets 216,000 hits. Perhaps Parallax should pay me something as well. I get PMs from people at Parallax thanking me for my evangelism on their behalf!
216,000 hits! How do find time to do anything else? I realize that most of those hits are referring to a different leon, parallax or propeller. I just find it odd that you seem to enjoy the role of being a contrarian on this forum -- almost to the point of exhibiting trollish behaviour.
Many people seem to like my contributions, which makes it all worthwhile.
Well, a lot of people like Raake Orret and Kim Chee, too. As with those preparations, perhaps the worst part of a troll is the smell. Perhaps my nose is more sensitive than the average.
@Coley - Total crack-up! Outstanding! Best ever!!!
"Durian - The edible flesh emits a distinctive odour, strong and penetrating even when the husk is intact. Some people regard the durian as fragrant; others find the aroma overpowering and offensive. The smell evokes reactions from deep appreciation to intense disgust, and has been described variously as almonds, rotten onions, turpentine and gym socks. The odour has led to the fruit's banishment from certain hotels and public transportation in southeast Asia."
The wife of a friend of mine comes from there and I asked her once if she could bring some back with her when she returned from a holiday, but it doesn't travel well.
And I guess that means you have lots of supporters? You're always good for a laugh.
Even the super salty popcorn is all sold out, and two members have ruptured their funny bones
I wouldn't pin too much reputation on the view count. There are a lot of reasons to read or return to a thread other than to be informed. After all, there's that other thread with more than 33K views.
BTW, the National Enquirer has nearly as large a circulation as the New York Times. What does that tell you?
The fact remains that anyone wishing to get an XMOS sales pitch would have already clickety-clacked "xmos" into their favorite search engine. Parallax.com is a dumb place to go for that information, and a dumb place to present that information. Recounting personal experiences with other products seems very legit to me. Evangelizing for them them doesn't.
It was rather stupid of Andre to start the thread in the first place. He must have known what was likely to happen, especially as the P2 doesn't exist yet. I wonder why he hasn't chimed in, perhaps we've frightened him off.
For many applications an ARM is much better than an XMOS chip, or a Propeller!
An XMOS chip is used in this reincarnation of the Amiga ... It's obviously cost a lot of money to develop it, but I can't see it being all that successful.
I tend to use whatever device best suits the application I'm working on, It might be an XMOS chip, a Propeller, or any one of a multitude of different devices.
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.
These are not the words of an XMOS-databook-thumping evangelical zealot. Like it or not, the subject of this thread is XMOS, and it was asked by a respected Parallax development partner. Leon has experience with the XMOS chip and is trying to be balanced and informative, in response to Andre's original question. I find it frankly delusional to infer from anything he's written here that he's trying to promote the XMOS chip to the detriment of the Propeller.
As a side note, the original question related to the XMOS vs. the P2, not the P1. So I'm not sure any valid comparison can yet be made, since the P2's specs have not been finalized and we have none in hand.
Comments
Given the high-level of abstraction available on the more powerful ARMs, I don't know if a "complete" game console is really a good target. To a certain degree, you might as well just create your games on an iPod Touch. But some form of ARM-based educational package would be cool.
Leon, I've looked hard at the XMOS and I could lean toward toward it but I also know that with any micro (and more so XMOS) that there is a learning curve, which means time. So any tips, any advice, any pitfalls that you can pass on to the uninitiated would be appreciated. Thanks mate.
The $99 XC-1A is nice because it is self-contained with a USB interface, has a 1600 MIPS four-core XS1-G4 chip, LEDs, buttons, and a little speaker. It also has plenty of I/O. I'd get one of those if you are primarily interested in parallel programming techniques. You can either connect external PCBs to it with ribbon cables pugged onto headers soldered into one of the four sets of pads, or design a PCB that plugs into a connector soldered to the pads running across the board. I've used both techniques. The prototyping area could also be used. It could be used with a solderless breadboard. I actually use the earlier XC-1. They get rather hot: I can't keep my finger on the XMOS chip for more than a couple of seconds, but don't worry about it. The older XTAG interface built-in to the XC-1A uses an FTDI chip which has the nice feature of a UART interface to the host PC. The XTAG-2 used with most of the other boards doesn't have a similar interface, I think that XMOS will be adding something like it. I've just tried the UART interface with the latest version of the tools, and it works very well at 115112 baud. A module (module_uart.1v1) is provided which makes coding the UART function quite trivial:
#include <xs1.h> #include <platform.h> #include "uart.h" on stdcore[0]: port p_rx = PORT_UART_RX; on stdcore[0]: port p_tx = PORT_UART_TX; /* This test runs the uart at the uart module's default bit rate of 115200 */ /* Echo any RX word on TX */ void test_uart(void) { chan data_ch; par { uart_tx(p_tx, data_ch); { // Send a message to Tx first char msg[] = "\n\rHello World\n\rFrom the Simple Uart\n\r"; for (int i=0; i<sizeof(msg); i++) { unsigned txData = (unsigned) (msg[i]); data_ch <: txData; } // Then echo rx data uart_rx(p_rx, data_ch); } } } int main(void) { par { on stdcore[0]: test_uart(); } return 0; }It's written in XC, note the par construct. Another program is provided that uses one thread for transmission and another for reception, similar to using two cogs with the Propeller. A long time ago I wrote a very simple UART program, but it stopped working when a new version of the the compiler came out, and I couldn't be bothered to sort it out. It's almost working with the latest version of the compiler, so I'll see if I can fix it.
For motor control, robotics and that sort of thing, you need a $59 XK-1A which uses the single-core XS1-L1. It comes with an XTAG-2 debug/programming interface and a PS. If you need to control more joints, for instance, in your robot, you can simply daisy-chain them together. Individual XK-1As cost $49, and the XTAG-2 on its own costs $19.
Sparkfun sells a $41.95 board with an XS1-L1 on it. Don't buy the XTAG-2 from them, they are stuck with some they bought before XMOS reduced the price and are charging $50.95 for them. The board isn't all that well designed, but it works OK. It has a large prototyping area which makes it ideal for interfacing individual components to it. It's a pity that one can't solder a connector to it as one can with the XC-1A, the pads aren't suitably arranged for that. I suppose that one could be soldered to the prototyping area and wires run across to it, but it would be rather messy. The earlier version of the board (which I have) incorporated a UART function implemented with an FTDI chip, but it never worked properly. I might add a MAX202 to one of my boards and implement a UART in software using the aforesaid module.
The XTAG-2 has an XS1-L1 on it, BTW, which runs the USB interface and the JTAG functions.
Download the development software, of course. It's Eclipse-based and new versions come out quite frequently. You need to be careful if you use the default Eclipse workspace, called "workspace", as other Eclipse-based tools might use it and can cause a conflict. I found that the Altera NIOS II compiler caused problems, so I renamed the workspace to XMOS. Installed applications seem to get corrupted sometimes and don't build; it's best to delete them, and the directory they are in, and download a new file from the XMOS web site. I haven't been able to work out the circumstances in which it happens, so that I can tell XMOS about it. Don't unzip downloaded applications before installing them, just import them into the XDE.
University staff can request $200 worth of free hardware. A friend of mine who runs the embedded systems lab at ESIEE near Paris is using his hardware for an interesting project - half the students will use XMOS kit and the other half will use real-time Java hardware and software. He wants to see which comes out best.
I'll come up with some more stuff as I think of it, but that should get you started.
In the google groops comp.lang.forth thread Intellasys vs XMOS Leon asked for it back in 2008, and was talking to the right bunch, but I never found anything.
Did you ever start one?
https://www.xcore.com/projects/working-title-xs1-forth
I lent Stephen Pelc at MPE a spare XC-1 board of mine, but I don't think that that he did anything with it.
It should be quite easy to port one of the Forths written in C to it.
Not that it's any of our business, but you do imply someone ....
I get the occasional freeby from NXP for running the LPC2000 Users Group, and never have any problems getting samples of new devices.
XMOS has some interesting board members and advisors:
http://www.xmos.com/company/team
They include Herman Hauser, Steve Furber and Sir Robin Saxby, who put ARM on the map, as well as Ivan Sutherland.
Too bad. Zero bytes is a little bit too minimal of a kernal size for me. The Green Arrays board is way too expensive, so it looks like the prop is still the only option.
Simple. He doesn't do anything else. Sharing pellets of wisdom is now Leon's life.
Well, a lot of people like Raake Orret and Kim Chee, too. As with those preparations, perhaps the worst part of a troll is the smell. Perhaps my nose is more sensitive than the average.
@Coley - Total crack-up! Outstanding! Best ever!!!
"Durian - The edible flesh emits a distinctive odour, strong and penetrating even when the husk is intact. Some people regard the durian as fragrant; others find the aroma overpowering and offensive. The smell evokes reactions from deep appreciation to intense disgust, and has been described variously as almonds, rotten onions, turpentine and gym socks. The odour has led to the fruit's banishment from certain hotels and public transportation in southeast Asia."
I thought that was a great summary!
http://www.chinatravel.com/kunming/food/smoked-smelly-bean-curd/
The wife of a friend of mine comes from there and I asked her once if she could bring some back with her when she returned from a holiday, but it doesn't travel well.
Even the super salty popcorn is all sold out, and two members have ruptured their funny bones
BTW, the National Enquirer has nearly as large a circulation as the New York Times. What does that tell you?
-Phil
These are not the words of an XMOS-databook-thumping evangelical zealot. Like it or not, the subject of this thread is XMOS, and it was asked by a respected Parallax development partner. Leon has experience with the XMOS chip and is trying to be balanced and informative, in response to Andre's original question. I find it frankly delusional to infer from anything he's written here that he's trying to promote the XMOS chip to the detriment of the Propeller.
As a side note, the original question related to the XMOS vs. the P2, not the P1. So I'm not sure any valid comparison can yet be made, since the P2's specs have not been finalized and we have none in hand.
-Phil