For embedded systems requiring low power, low chip count, linux and decent floating point performance, the PPC chips are better than any x86 alternatives. Ok, there are some x86 alternatives but the problem with x86 is that each chip design has a very short market life. You don't want to have to redesign your embedded motherboard every 2 years just to be able to ship products. Motorola has always been a great embedded PPC processor manufacturer.
The worst thing in the world is a manager who THINKS he knows how you should implement a specific feature, and doesn't realize that he knows bull.
Also, a manager who is clueless is more likely to accept a clueless employee who can't program and no one realizes it until one week before the drop-dead shipping date.
Most people I have interviewed for programming jobs who were just out of college or university have never designed their own C++ class hierarchy. But they say they 'know C++' because they double clicked on a button in Visual C++ and typed 'MessageBox("hello","test",MB_OK)'. Ask them about proper STL usage, exception safety, why would you use 'mutable', under which circumstances would you use or not use virtual multiple inheritance and you get dumb blank stares.
When Microsoft sucessfully lobbies the U.S. Government to crack down on the 'Communist GPL' and make it illegal to have or make or to use GPL'd software.
Don't think it is possible? Of course the GPL is not communist, and in a true free market the GPL would be free to compete with M$. No one is FORCING people to use GPL'd software.
With McCarthyism people were put in jail for less... 'She reads lots of books, I think she might be a communist! Arrest her!'.... 'Only evil hackers use linux! My neighbour uses linux! Arrest him!'
Luckily with IBM's support of the GPL, this won't happen soon.
What I want is to have caller-ID distinctive midi ring tones. All my friends could be assigned a MIDI riff that describes them. Then I'd know who it was BEFORE I looked at the phone!
YES! WHEN is there going to be a single multicasting standard that is usable over normal net connections? For streaming live video this is the only way to go - You just TRY to stream 8 live productions a day, each 1.5 hours long, to a total of 4000 people at a time, at a shitty 128 kbps. THEN try to pay the bandwidth bill at the end of the month. Coherent multicasting support in all routers will open up the world to easy, free live video.
I am currently using Linux in an embedded system based roughly on the sandpoint 8240... and it was a fair bit of a mess as sandpoint is not an 'official linux port'... I will definitely look into NetBSD/Sandpoint instead.
Is to uniquely mark up the paper that is used for 'anonymous' surveys. When you send back your response, it isn't anonymous anymore. They know who you are.
--jeff
The problem with doing it at home
on
Tokyo.Disney.Net
·
· Score: 1
Normal sound cards do not have PLLs (Phase Locked Loops) to synchronize the sample rate like cobranet does. So all your speakers will be delayed differently after you stream for a while. After you stream for a long time, you will probably get buffer underruns. When a sound card is at 44.1khz, it is never EXACTLY that.
A cobranet receiver extracts the clock from the timing of the received packets and runs it through a phase locked loop to make it stable. It works surprisingly well.
I got to work on a product used at TDS. The TDS system is massive.
Cobranet is pretty cool. Check out http://www.peakaudio.com/
The audio mixing is done by Level Control Systems. Check out http://www.lcsaudio.com/ - It transports audio over analog, cobranet, adat optical, redundant gigabit fibrechannel, and also has embedded linux for audio playback from scsi disks with raid. Scales to hundreds of audio channels in and out. Massive DSP processing, all done in floating point.
And that just comprises a PART of tokyo disney seas!
Isn't it funny that as technology gets better and better, the quality of 'available' multimedia gets lower and lower. Sure, quantity has gone up but now we have lots of garbage to watch and listen too.
Welcome to the mid-90's: V-Chips are a USAian invention, are standard because the US gov't deemed them nec. Canada just said sure, what the heck, we'll use 'em. Thus Canadian TV is now rated but it's not native technology, it's US stuff.
LIES!
The V-Chip was invented in Burnaby, British Columbia, CANADA, at the Simon Fraser University by Tim Collings.
Grab my test source code for measuring altivec performance.
http://www.jdkoftinoff.com/eqtest.tar.gz
works on the G4 with OS-X and Linux-PPC. Can compile and run on non-altivec processors, but not optimally since the algorithm is still focussed on altivec.
I like that point. I would like to add my subjective view that 'Good art is always socially relevant'. Many people would disagree with me though.
Good programming can be an art. But I feel that the current 'state of the art programming' is equivalent to where painting was in medieval times when they finally figured out how to paint a christ child that didn't look just like a shrunken adult.
With the more complex processors, pipelining can get very complex. Try programming the TI 6701 DSP in assembly. It is a Very Long Instruction Word (VLIW) processor with software pipelining and 'flying registers'. For a non-trivial loop, the C compiler almost always generates better code than a human can.
In fact, on that DSP, they have a special 'optimizing linear assembler' so you don't have to think about parallel instructions and pipelining. And it STILL is very very complex.
The time of hand coded raw assembly is ending soon....
i can understand why open source software is usefull to me. i can just download it and run it. it's free and doesn't require any special physical equipment.
Wow cool, you can download and run open source software without any special equipment like a computer and monitor and ram and disk drive?
Open source VHDL and Verilog designs are a GREAT concept.
When enough people catch on, then we'll start seeing open source VHDL compilers.
If you design or change a modem algorithm, you are required to have it approved by the FCC before you are legally allowed to connect it to the telephone network.
Most of you wouldn't know what to do with MWAVE DSP source code anyways... You don't need it.
Slashdot Mirror
Good point, yes SHx, MIPS, and ARM as well as PPC are now all good options for embedded systems. Usually x86 is not.
--jeff
For embedded systems requiring low power, low chip count, linux and decent floating point performance, the PPC chips are better than any x86 alternatives. Ok, there are some x86 alternatives but the problem with x86 is that each chip design has a very short market life. You don't want to have to redesign your embedded motherboard every 2 years just to be able to ship products. Motorola has always been a great embedded PPC processor manufacturer.
--jeff
The worst thing in the world is a manager who THINKS he knows how you should implement a specific feature, and doesn't realize that he knows bull.
Also, a manager who is clueless is more likely to accept a clueless employee who can't program and no one realizes it until one week before the drop-dead shipping date.
Most people I have interviewed for programming jobs who were just out of college or university have never designed their own C++ class hierarchy. But they say they 'know C++' because they double clicked on a button in Visual C++ and typed 'MessageBox("hello","test",MB_OK)'. Ask them about proper STL usage, exception safety, why would you use 'mutable', under which circumstances would you use or not use virtual multiple inheritance and you get dumb blank stares.
Why is this?
--jeff
When Microsoft sucessfully lobbies the U.S. Government to crack down on the 'Communist GPL' and make it illegal to have or make or to use GPL'd software.
Don't think it is possible? Of course the GPL is not communist, and in a true free market the GPL would be free to compete with M$. No one is FORCING people to use GPL'd software.
With McCarthyism people were put in jail for less... 'She reads lots of books, I think she might be a communist! Arrest her!'.... 'Only evil hackers use linux! My neighbour uses linux! Arrest him!'
Luckily with IBM's support of the GPL, this won't happen soon.
--jeff
I wonder if it will take the C++ guys longer to rewrite it than the lisp guys took to write the original?
I use C++ all the time, and I am still amazed at the poor quality of most c++ projects out there.
--jeff
What I want is to have caller-ID distinctive midi ring tones. All my friends could be assigned a MIDI riff that describes them. Then I'd know who it was BEFORE I looked at the phone!
YES! WHEN is there going to be a single multicasting standard that is usable over normal net connections? For streaming live video this is the only way to go - You just TRY to stream 8 live productions a day, each 1.5 hours long, to a total of 4000 people at a time, at a shitty 128 kbps. THEN try to pay the bandwidth bill at the end of the month. Coherent multicasting support in all routers will open up the world to easy, free live video.
--jeff
This is quite awesome.
I am currently using Linux in an embedded system based roughly on the sandpoint 8240... and it was a fair bit of a mess as sandpoint is not an 'official linux port'... I will definitely look into NetBSD/Sandpoint instead.
--jeff
Is to uniquely mark up the paper that is used for 'anonymous' surveys. When you send back your response, it isn't anonymous anymore. They know who you are.
--jeff
Normal sound cards do not have PLLs (Phase Locked Loops) to synchronize the sample rate like cobranet does. So all your speakers will be delayed differently after you stream for a while. After you stream for a long time, you will probably get buffer underruns. When a sound card is at 44.1khz, it is never EXACTLY that.
--jeff
A cobranet receiver extracts the clock from the timing of the received packets and runs it through a phase locked loop to make it stable. It works surprisingly well.
--jeff
I got to work on a product used at TDS. The TDS system is massive.
Cobranet is pretty cool. Check out http://www.peakaudio.com/
The audio mixing is done by Level Control Systems. Check out http://www.lcsaudio.com/ - It transports audio over analog, cobranet, adat optical, redundant gigabit fibrechannel, and also has embedded linux for audio playback from scsi disks with raid. Scales to hundreds of audio channels in and out. Massive DSP processing, all done in floating point.
And that just comprises a PART of tokyo disney seas!
--jeff
Awesome!
With cable before, I would get various results. Sometimes >4 Mbits, sometimes 1 Mbits.
ADSL is consistent for me.
jeff
In Vancouver, BC Canada
ADSL with 2.5 Mbit download, 640 Kbit upload. Static IP address. $80.00 CDN = $55.00 US
Works great. For a bit more a month they can give me 4.0 Mbit download.
Isn't it funny that as technology gets better and better, the quality of 'available' multimedia gets lower and lower. Sure, quantity has gone up but now we have lots of garbage to watch and listen too.
I find it painful to watch any web broadcast.
However it was the US that took the technology & ran with it. Had it remained Canadian it would have sank without a trace.
Yup. just like everything else Canadian. If it isn't sold to the Americans first, the Canadians don't want it.
LIES!
The V-Chip was invented in Burnaby, British Columbia, CANADA, at the Simon Fraser University by Tim Collings.
But they can't restrict anyone who paid the $15 from putting the files on another ftp mirror for free download to everyone else.
Of course!
That's why I said it wasn't optimized for non-altivec, silly!
Download the code and add the non-altivec optimized version then. It's GPL.
BTW You can compare the non-altivec tests on G4 as well. A 450 Mhz G4 without using altivec performs faster than a P-III 667.
See for yourself.
Grab my test source code for measuring altivec performance.
http://www.jdkoftinoff.com/eqtest.tar.gz
works on the G4 with OS-X and Linux-PPC. Can compile and run on non-altivec processors, but not optimally since the algorithm is still focussed on altivec.
Run it on all your boxes and see for yourself
--jeff
I like that point. I would like to add my subjective view that 'Good art is always socially relevant'. Many people would disagree with me though.
Good programming can be an art. But I feel that the current 'state of the art programming' is equivalent to where painting was in medieval times when they finally figured out how to paint a christ child that didn't look just like a shrunken adult.
We have a LONG way to go. Most code sucks.
With the more complex processors, pipelining can get very complex. Try programming the TI 6701 DSP in assembly. It is a Very Long Instruction Word (VLIW) processor with software pipelining and 'flying registers'. For a non-trivial loop, the C compiler almost always generates better code than a human can.
In fact, on that DSP, they have a special 'optimizing linear assembler' so you don't have to think about parallel instructions and pipelining. And it STILL is very very complex.
The time of hand coded raw assembly is ending soon....
Wow cool, you can download and run open source software without any special equipment like a computer and monitor and ram and disk drive?
Open source VHDL and Verilog designs are a GREAT concept.
When enough people catch on, then we'll start seeing open source VHDL compilers.
If you design or change a modem algorithm, you are required to have it approved by the FCC before you are legally allowed to connect it to the telephone network.
Most of you wouldn't know what to do with MWAVE DSP source code anyways... You don't need it.
Ha! good point!
Well I think they were scared and didn't understand exactly what it did. If I Love You was nastier they could have been required to do that.
Regardless, it cost them $$$$