Firewire is designed as an external interface, and as such is limited by long cable speed etc. Serial ATA is an internal only interface, and can achieve much faster speeds. It's also a point to point system, so it doesn't suffer from Firewire's need for device identification etc.
Also, Firewire is designed to supply power over the interface - Serial ATA dedicates lines for this, which aren't optional.
The interconnect speeds used on current GPUs make connecting them together almost impossible. You just cannot move data at 1/2 GB/sec over a cable, unless you go serial (ala Serial ATA). Current GPU card designs are hyper critical of trace length between the RAMs and the GPU - at those speeds even a difference of a few millimeters can make or break a card.
Until someone comes up with a radically new scheme of processing, these physical limitations will always be with us. That's why the Voodoo (3Dfx) scheme of connecting stuff no longer works - the data rates have got too high, and the limits of light speed (electrons) start applying.
VRAM was a dual port RAM, with an output only port and a standard I/O port. The output port was used to refresh the screen, since it could produce data independantly from the I/O port.
However, with the bus widths being used by GPUs today (128,256), they really don't fit anymore. GPUs now manage the RAM accesses so that frame buffer access is shared with drawing etc. This means that the most important thing is RAM speed - with accesses for the frame buffer being sequential, the less time taken for that the more memory bandwidth left for drawing.
This will become even more important once we have the very high resolution monitors (LCD) on the horizon - 3K x 2K pixel displays will require a LOT of memory bandwidth to keep them refreshed.
Not all FPGAs lose their programming on power off. Some have an internal flash rom that backs up the ram (or even dispenses with it completely). Field Programmable Gate Array means exactly that - it CAN be programmed in the field, but doesn't have to be.
But, that "fair use" only applies to the original user who did the reverse engineering. AFAIK (IANAL) publishing or distributing such efforts bring you right into the DMCAs targets. Hopefully someone will point out that I'm wrong.
Funnily enough, all microsoft email addresses are xxxxxxxx@... (maximum 8 characters before the @). Seems that it's a legacy of the old Microsoft Mail system that used 8.3 filenames. Wonderful how their technology is stuck back in the middle ages. Contract/external people are also stuck with a "v-" prefix, leaving only 6 characters for their name/initials.
Not sure about the watering down, but in those times water was pretty dangerous stuff, full of impurities etc.
Wine and other fermented drinks were much safer to drink, since the process that produces alcohol also kills many of the harmful bacteria in the water.
So turning water into wine in effect purified it so the people could drink safely.
The system Apple used was the SASI standard, from which the SCSI standard was derived. In theory, SCSI drives can be used as SASI ones, but advances over the years have made this almost impossible.
Main differences are multiple masters (arbitration), messages and multiple device types. For anyone that cares to know, the SASI system was initially designed for 8" floppy drives by Shugart, before small hard drives even existed. Given that we are talking about losing the floppy, this is an interesting coincidence.
I'd like to support this too, but where I am (China) I can buy unprotected music CDs for $2, VCDs (video CDs) for $3, all legally. I can, of course, also buy from legitimate outlets pirate DVDs for $3 (like Die Another Day, with the oscar stuff over it).
There is a music/movie store here that sells real import DVDs and CDs, but at USA prices. Given that those prices exceed the monthly pay for most people here, they don't sell that many!
And, yes, I too was unemployed in the USA before I came here. I'd like to be back there, but stuff like the DMCA makes me think twice.
I think we were all very lucky...
on
Dealers of Lightning
·
· Score: 2, Interesting
that Xerox didn't take advantage of the many innovations that came from or were inspired by PARC.
A great many companies (Sun, Apollo etc.) benefited either directly or indirectly from their work, and over the years we have all benefited.
In the early 80s I worked on a couple of projects that certainly had their roots there. One was interfacing a laser printer (Seimens) to a batch processing mainframe (where the biggest problem was whether the unions would allow non union labor to use it). The other was the ICL (sort of defunct UK equiv. to IBM) port of Unix to the Three Rivers Perq. Fun times back then (must be getting old).
But surely the point is that 'Joe Public' will never see or hear about this case. Us on/. are aware of it, but let's face it, we're not exactly 'Joe Public'.
JP only notices things that are on national media (if they even watch the news), so unless this becomes the story of the day, they will never hear about it.
As far as Congress (and the corporations) are concerned, ignorance is wonderful, and to be exploited as far as possible.
Reading the article, it appears that these diseases are already in the major growing areas. And also, contrary to a previous post, these diseases mutate to resist traditional remedies so fast that treatment (pesticides) is not an option. Therefore there is a real urgency to this research - if not completed, there seems to be a real chance that the banana as we know it will simply be extinct.
A more important point is that although GM may well be a cure, it remains to be seen whether or not consumers would accept modified bananas.
Which may explain why SMS is not popular in the USA.
In almost every other country, SMS messaging is a huge business - the company I am working for in China was smart enough to be one of the providers. They make a lot of money even with a tiny percentage on each message.
Sklyarov wasn't acquitted - he agreed to be a witness for the prosecution in return for immunity. That is a BIG difference.
His company was acquitted, not him. Even if they were, he still spent a certain amount of time in a US jail, away from his family. He will never receive compensation for that time.
So, it seems that the USA and Norway are not equal. Especially as I don't remember a case where Norway attempted to use it's laws against someone not even in the country.
Please, please, can someone post this or send me a copy - where I am (China, temporarily) I can't access it. Any email to the above domain will get to me.
Like many here, I read Slashdot a lot (perhaps too much), and it's really nice to see Bruce getting involved with discussions etc.
Although we often see other luminaries mentioned, he is one of the few that takes the time to actually discuss matters he is concerned about. Only the other day, a discussion he was involved with brought comments from Jim Gettys and others that were very enlightening.
Thanks again, Bruce, for your articles and (unlike some of us) your unfailing ability to get to the point, instead of some of the superfluous stuff that appears here.
As a (now ex) H1-B worker from England (as in English etc.) one of the things companies must do (by law) is pay the equivalent rate for the job as they would pay an American
The fact that many don't do this is a government/oversight problem, not the workers problem. The law is already in place - it just gets abused by the employers.
Didn't hear too many complaints when there was full employment, either. Blame the companies, not the people they bring over. Of course, the fact that an H1-B has 10 days to get out of the country if they're laid off keeps the unemployment rate down.
Sorry, but get a bit pissed at the people who blame foreigners for the US economy.
Am I the only one that thinks it would be better to build a robot that does the fighting, and the soldier worries about the robot? It would seem a lot safer, since humans are 'supposed' to be more emotional than robots.
Since, in theory, the robot would never get anxious (even when half of it is blown away), it would make the soldiers job pretty easy.
Soldier - 'Sir, my robot got anxious and was blown to bits'.
Sergeant - 'Good work son, I'm sure your support helped - here's another robot for you to empathize with'
Interesting comment - as an English guy (who never made it to uni bacause of family issues), the dependance on College in the USA is much stronger than in the UK. It's almost as if you are useless if you didn't go to college (even though people like Bill Gates seem to have done OK). To get my H1-B visa, they had two get to college professors to evaluate my 20+ years experience and equate it to a degree.
I think the big difference is that in the UK you begin to specialize at 16, if not before. That, within the US High School system, is almost impossible. So in the USA, at 18, you don't know much more that you did at 16.
Believe me, you'd be just one of the 'college kids' in the US (and a year older). The difference is they would know that, straight from High School, you're pretty much useless (excepting, of course, those smart enough to overcome that handicap).
Of course, in the current economy, many graduates are working at McDonalds, but that's another story.
Slashdot Mirror
Try explaining that to DVD Jon in Norway - I'm fairly certain he would disagree with you.
For the rest of us out here in the rest of the world (outside the USA), the reach of American corporate interests is pretty amazing.
Maybe currently, but Microsoft is known to be working on including 3D elements on the desktop.
So in the future it may be useful to allow for a gender related setting.
Firewire is designed as an external interface, and as such is limited by long cable speed etc. Serial ATA is an internal only interface, and can achieve much faster speeds. It's also a point to point system, so it doesn't suffer from Firewire's need for device identification etc.
Also, Firewire is designed to supply power over the interface - Serial ATA dedicates lines for this, which aren't optional.
The interconnect speeds used on current GPUs make connecting them together almost impossible. You just cannot move data at 1/2 GB/sec over a cable, unless you go serial (ala Serial ATA). Current GPU card designs are hyper critical of trace length between the RAMs and the GPU - at those speeds even a difference of a few millimeters can make or break a card.
Until someone comes up with a radically new scheme of processing, these physical limitations will always be with us. That's why the Voodoo (3Dfx) scheme of connecting stuff no longer works - the data rates have got too high, and the limits of light speed (electrons) start applying.
VRAM was a dual port RAM, with an output only port and a standard I/O port. The output port was used to refresh the screen, since it could produce data independantly from the I/O port.
However, with the bus widths being used by GPUs today (128,256), they really don't fit anymore. GPUs now manage the RAM accesses so that frame buffer access is shared with drawing etc. This means that the most important thing is RAM speed - with accesses for the frame buffer being sequential, the less time taken for that the more memory bandwidth left for drawing.
This will become even more important once we have the very high resolution monitors (LCD) on the horizon - 3K x 2K pixel displays will require a LOT of memory bandwidth to keep them refreshed.
Somehow the words 'safe' and 'asbestos-lined' in the same sentence don't seem to fit together.
Not all FPGAs lose their programming on power off. Some have an internal flash rom that backs up the ram (or even dispenses with it completely). Field Programmable Gate Array means exactly that - it CAN be programmed in the field, but doesn't have to be.
But, that "fair use" only applies to the original user who did the reverse engineering. AFAIK (IANAL) publishing or distributing such efforts bring you right into the DMCAs targets. Hopefully someone will point out that I'm wrong.
You can PayPal it to bill.gates@microsoft.com
Funnily enough, all microsoft email addresses are xxxxxxxx@... (maximum 8 characters before the @). Seems that it's a legacy of the old Microsoft Mail system that used 8.3 filenames. Wonderful how their technology is stuck back in the middle ages. Contract/external people are also stuck with a "v-" prefix, leaving only 6 characters for their name/initials.
Not sure about the watering down, but in those times water was pretty dangerous stuff, full of impurities etc.
Wine and other fermented drinks were much safer to drink, since the process that produces alcohol also kills many of the harmful bacteria in the water.
So turning water into wine in effect purified it so the people could drink safely.
Whilst you were at that former employer, did you bring the license probems to the attention of the board or other responsible people?
If you did (and perhaps learnt they didn't care, or lost your job as a result), then good, but if you didn't then I have to wonder about your motives.
The system Apple used was the SASI standard, from which the SCSI standard was derived. In theory, SCSI drives can be used as SASI ones, but advances over the years have made this almost impossible.
Main differences are multiple masters (arbitration), messages and multiple device types. For anyone that cares to know, the SASI system was initially designed for 8" floppy drives by Shugart, before small hard drives even existed. Given that we are talking about losing the floppy, this is an interesting coincidence.
I'd like to support this too, but where I am (China) I can buy unprotected music CDs for $2, VCDs (video CDs) for $3, all legally. I can, of course, also buy from legitimate outlets pirate DVDs for $3 (like Die Another Day, with the oscar stuff over it).
There is a music/movie store here that sells real import DVDs and CDs, but at USA prices. Given that those prices exceed the monthly pay for most people here, they don't sell that many!
And, yes, I too was unemployed in the USA before I came here. I'd like to be back there, but stuff like the DMCA makes me think twice.
that Xerox didn't take advantage of the many innovations that came from or were inspired by PARC.
A great many companies (Sun, Apollo etc.) benefited either directly or indirectly from their work, and over the years we have all benefited.
In the early 80s I worked on a couple of projects that certainly had their roots there. One was interfacing a laser printer (Seimens) to a batch processing mainframe (where the biggest problem was whether the unions would allow non union labor to use it). The other was the ICL (sort of defunct UK equiv. to IBM) port of Unix to the Three Rivers Perq. Fun times back then (must be getting old).
So instead of losing a mars probe because of miscalculations, we now lose a percentage of the data?
Of course, if it's measured in units USAians understand, it will be 'those other people' who will get it wrong, which is probably alright, yes?
But surely the point is that 'Joe Public' will never see or hear about this case. Us on /. are aware of it, but let's face it, we're not exactly 'Joe Public'.
JP only notices things that are on national media (if they even watch the news), so unless this becomes the story of the day, they will never hear about it.
As far as Congress (and the corporations) are concerned, ignorance is wonderful, and to be exploited as far as possible.
Reading the article, it appears that these diseases are already in the major growing areas. And also, contrary to a previous post, these diseases mutate to resist traditional remedies so fast that treatment (pesticides) is not an option. Therefore there is a real urgency to this research - if not completed, there seems to be a real chance that the banana as we know it will simply be extinct.
A more important point is that although GM may well be a cure, it remains to be seen whether or not consumers would accept modified bananas.
Which may explain why SMS is not popular in the USA.
In almost every other country, SMS messaging is a huge business - the company I am working for in China was smart enough to be one of the providers. They make a lot of money even with a tiny percentage on each message.
That's strange - so all these huge trade deficits are just rumours then?
That explains why no one seems to care about them.
Sklyarov wasn't acquitted - he agreed to be a witness for the prosecution in return for immunity. That is a BIG difference.
His company was acquitted, not him. Even if they were, he still spent a certain amount of time in a US jail, away from his family. He will never receive compensation for that time.
So, it seems that the USA and Norway are not equal. Especially as I don't remember a case where Norway attempted to use it's laws against someone not even in the country.
Please, please, can someone post this or send me a copy - where I am (China, temporarily) I can't access it. Any email to the above domain will get to me.
Like many here, I read Slashdot a lot (perhaps too much), and it's really nice to see Bruce getting involved with discussions etc.
Although we often see other luminaries mentioned, he is one of the few that takes the time to actually discuss matters he is concerned about. Only the other day, a discussion he was involved with brought comments from Jim Gettys and others that were very enlightening.
Thanks again, Bruce, for your articles and (unlike some of us) your unfailing ability to get to the point, instead of some of the superfluous stuff that appears here.
As a (now ex) H1-B worker from England (as in English etc.) one of the things companies must do (by law) is pay the equivalent rate for the job as they would pay an American
The fact that many don't do this is a government/oversight problem, not the workers problem. The law is already in place - it just gets abused by the employers.
Didn't hear too many complaints when there was full employment, either. Blame the companies, not the people they bring over. Of course, the fact that an H1-B has 10 days to get out of the country if they're laid off keeps the unemployment rate down.
Sorry, but get a bit pissed at the people who blame foreigners for the US economy.
Am I the only one that thinks it would be better to build a robot that does the fighting, and the soldier worries about the robot? It would seem a lot safer, since humans are 'supposed' to be more emotional than robots.
Since, in theory, the robot would never get anxious (even when half of it is blown away), it would make the soldiers job pretty easy.
Soldier - 'Sir, my robot got anxious and was blown to bits'.
Sergeant - 'Good work son, I'm sure your support helped - here's another robot for you to empathize with'
Interesting comment - as an English guy (who never made it to uni bacause of family issues), the dependance on College in the USA is much stronger than in the UK. It's almost as if you are useless if you didn't go to college (even though people like Bill Gates seem to have done OK). To get my H1-B visa, they had two get to college professors to evaluate my 20+ years experience and equate it to a degree.
I think the big difference is that in the UK you begin to specialize at 16, if not before. That, within the US High School system, is almost impossible. So in the USA, at 18, you don't know much more that you did at 16.
Believe me, you'd be just one of the 'college kids' in the US (and a year older). The difference is they would know that, straight from High School, you're pretty much useless (excepting, of course, those smart enough to overcome that handicap).
Of course, in the current economy, many graduates are working at McDonalds, but that's another story.