I agree, he spends most the article on how he set up his wireless card and how he set his clock?
For starters, here's a list of things it would of been nice if he covered if he were to fill us in on the media part(in no particular order): * Sound card? Is he using on board sound? Does it support hardware mixing, if not, what kind of software mixing is he using. Does he have a digital connection to his receiver for ac3 passthrough, and pcm output?
* More information on the video card. He says it supports multiple HD standards, but says it only has a composite and s-video output (i'm assuming in addition to the dvi/vga,) but no component out. How is he gonna output HD then? AFAIK, the nvidia support for HD out requires component out, else you have to set your custom modelines (as the latest nvidia driver 1.0.8178.) While hd output isn't necessary for a home media center I suppose, it increases your geek factor. Is the CPU gonna be fast enough to output 1080i?
* Power consumption. Would be nice for him to try and reduce the power on the unit, when it's not being used, considering it'll probably be idle most the time.
* Choice of case for something thats pleasing to the eye, and silent to the ear. Not a big problem if it's tucked away. But usually these things are in the living/family rooms alongside the entertainment center.
* Controllability: How is he controlling it? x11vnc? Mouse? Keyboard? lirc? What kinds of issues do you run into with these different choices.
* Software: He spends out two paragraphs on using totem and rhythmbox. How well did these integrate with his media in being able to play everything (other than the spew about OGG and mp3, flac anyone?), why is it important that gstream is uninstalled? Would be nice if he had tried some kind of fully-featured software that plays video and audio...He mentions a 6 in 1 card flash reader...is he using it? If so, with which software?
DVD sound is great - and I have to ask, how much better can the sound actually get with HD-DVD? We're back into CD vs SACD territory there.
Short Answer: Yes there is room for audio improvement.
DVD sound is great, because you are comparing them to vhs. DVD audio is very highly compressed. A typical ac3 5.1 channel track for a action thriller only takes up like 300-600 megs. That's at 5 channels for about 2 hours. Comparably a CD is (2 channel) 800 megs for about 80 minutes roughly. They don't get this tremendous decrease in audio size by lossless compression. Just find a dvd (usually live concerts) that has both PCM and ac3 audio tracks and compare them. The difference is quite significant.
Given that, only new movies (and old ones with all the tracks saved digitally and uncompressed would be able to benefit by having every channel in pcm. 7.1 channels in raw PCM probably take up about 7-8 gigs for a typical movie. This is all speculation, I'm not sure how the audio works on HD-DVD and BluRay and I didn't read the article.
Agreed. I was on a "field trip" at IBM's premier fab plant in East Fishkill, NY. When the Q/A portion came, someone asked what their yields were. They were allowed to say because it's IBM Confidential. But from what we gathered it's below 50% (this isn't read as being 49%, but the question was actually either below 50% or above 50%) for the 90nm process size.
But compared to the other big video store around here, Hollywood Video, this policy is much better. Typical rentals at blockbuster are for 7-8 days, so this is 8 days on top of that, which is 15-16 day rentals. I supposedly returned a few movies "late" to Hollywood video, and was shocked the next time when I had to pay about 9 dollars in late charges. The next time I rented a movie was with blockbuster.
Personally I used to like the 99 cent video stores run by asians. Great porn selection, 1 day rental (less lapse in memory in returning,) and only 99 cents. Want it for 2 days? Okay $1.98. The only problem here is with selection of the new movies, which they won't have 40 copies of. Selection could be spotty, but then again my local blockbuster didn't even have Deliverance.
There is also an article on him today on Day to Day and an obituary on All Things Considered. There are about a half dozen or so articles on him on NPR actually for those interested.
In other news about manufacting tolerances: Resistors spec'd at 5% tolerance are found to consistantly be 5% less than their stated value, in order to save money on raw materials.
You have any proof of this? While I'm not an expert on how resistors are made, I know they come in certain sizes because they are easier to make that way. Making them 5% off would be harder! That's why you have weird resistors values like 4.7k, because they are easier to make than 5k.
Unless the PLL in the motherboard is periodically recalibrated using (for instance) NIST, there's no way in hell that it's ever going to be accurate.
Do you even know how a PLL works? It basically takes a reference clock, and synchronizes the output to this reference clock (usually multiplied using a divider). If you are interested here's a brief introduction. The reference clock is a crystal. Crystals are accurate to about to many decimel places. That's why you get crystals that are quoted to be 15.00000MHz. The nature of a PLL is a control feedback system of the 2nd order. Basically, if the clock is off, it will automatically correct itself (as long as it stays in the lock range.) Basically the accuracy is going to come down to the power supply causing jitter, but that isn't going to affect the frequency. Additionally, PLL's are most likely going to be all on one integrated circuit, causing any manufacturing related tolerances, to be uniform through all the components, and therefore transistor ratios in the IC (which is what is important) will remain the same. Not to mention that there won't be any resistors in it, they will use transistors with the gate tied to the drain.
Internal clocks on inside these things are very accurate. If they were off, there would be all kinds of problems, data wouldn't line up. Even an Intel chip has multiple PLLs inside to syncronize the local clocks to each other in the clock distribution network.
I didn't read the article, but, how is the computers real time clock affecting any of these measurements at all? The benchmarking time is measured from clock ticks, from the cpu, which is very accurate.
Maybe because the highway between LA and Las Vegas (I-15) is one of the most dangerous roads in the US?
Not only is it dangerous because of speeding, but
1) It's in the middle of the desert
2) Most people driving on it are sleep deprived from partying all night and
3) are still drunk
That's just to mention a few risks. With soaring gas prices, it is probably cheaper to fly to vegas as well. When you are in Vegas, there is no need for a car.
What really would of been nice is the super speed train from Vegas to Southern California, that I believe is now scrapped. It would take you to/from vegas in 75/90 minutes.
You're complaing about 5-CD installs? I guess you don't remember the days of the 15 3.5" disk installs. Not to mention it would probably come with 20 5.25" disks, in case you didn't have a state of the art 3.5" drive. After buying a few games, you're entire 100 disk capacity caddy would be full. Don't even get me started with games that used low density floppies...
That being said, when the first cd-rom games came out, a lot of people used the extra space to make really horrible games. I remember one game I received for free that was 3 CD's, which was unheard of, when most people had 1-2x speed cd-roms. The entire game was a movie, and every 2-3 minutes of script, you were able to make a choice of what to say next, to then watch another 2-3 minutes of script. Okay, if you're gonna make a movie and put it on cd's, thats one thing, but to make a movie and have me click every 2 minutes is not fun.
I don't remember the name of the game, but it was something along the lines of stopping a nuclear war or something.
Besides, once you are cooped up inside one of those natty suits that you have to wear in modern chip fabrication environments, believe me when I say that the lighting is *not* a major concern...
I was under the impression that modernchip fabrication environments were all automated and didn't require the full garb since all the wafers are enclosed and pushed around on air. As is explained in the 3rd paragraph of that article. Of course, I'll let you know in a few weeks, since we're taking a field trip there.
For us joe twelve packs (or just people that don't care about flashy boxes), the Retail BFG 6800 Ultra (and I think other models) come in a nice flashy white box.
The point of this long-windedness is basically to say that they won't have any problem finding emergency landing strips in the (fairly unlikely, IMO) event that they need one.
It probablly has a slightly higher chance of failing since it only has one engine.
Simple Barter system:
1-2 Hours == A few beers or so.
3-4 Hours == A few more beers + some munchies.
5-8 Hours == Scotch + a stripper.
9-16 Hours == Stimulants + a massage.
17+ hours == A lot more stimulants.
True. The threshold voltage (the voltage necessary to turn a transistor "on" at the gate) is bottoming out around 0.2V. Below this when a transistor is suppose to be off, there is a significant (~10E-9A) amount of current and this increases exponentially w/ decrease in threshold voltage. Not only do you have problems with power, but stacked transistors are no longer at 0 volts in the intermediate junction when they are suppose to be, resulting in slower pull ups and pull downs.
Size of the chip is not a problem. Currently, about 80-90% of the chip is L2 cache, the actual CPU only consists of a very small corner of the chip. Overall, speed is more likely to increase, due to the lower amount of charge that is needed to turn on/off a gate (RC Delays.)
Slashdot Mirror
You're right. It should be: "...computer scientists imagine..."
I agree, he spends most the article on how he set up his wireless card and how he set his clock?
For starters, here's a list of things it would of been nice if he covered if he were to fill us in on the media part(in no particular order):
* Sound card? Is he using on board sound? Does it support hardware mixing, if not, what kind of software mixing is he using. Does he have a digital connection to his receiver for ac3 passthrough, and pcm output?
* More information on the video card. He says it supports multiple HD standards, but says it only has a composite and s-video output (i'm assuming in addition to the dvi/vga,) but no component out. How is he gonna output HD then? AFAIK, the nvidia support for HD out requires component out, else you have to set your custom modelines (as the latest nvidia driver 1.0.8178.) While hd output isn't necessary for a home media center I suppose, it increases your geek factor. Is the CPU gonna be fast enough to output 1080i?
* Power consumption. Would be nice for him to try and reduce the power on the unit, when it's not being used, considering it'll probably be idle most the time.
* Choice of case for something thats pleasing to the eye, and silent to the ear. Not a big problem if it's tucked away. But usually these things are in the living/family rooms alongside the entertainment center.
* Controllability: How is he controlling it? x11vnc? Mouse? Keyboard? lirc? What kinds of issues do you run into with these different choices.
* Software: He spends out two paragraphs on using totem and rhythmbox. How well did these integrate with his media in being able to play everything (other than the spew about OGG and mp3, flac anyone?), why is it important that gstream is uninstalled? Would be nice if he had tried some kind of fully-featured software that plays video and audio...He mentions a 6 in 1 card flash reader...is he using it? If so, with which software?
Worst LMHC article ever.
DVD sound is great - and I have to ask, how much better can the sound actually get with HD-DVD? We're back into CD vs SACD territory there.
Short Answer: Yes there is room for audio improvement.
DVD sound is great, because you are comparing them to vhs. DVD audio is very highly compressed. A typical ac3 5.1 channel track for a action thriller only takes up like 300-600 megs. That's at 5 channels for about 2 hours. Comparably a CD is (2 channel) 800 megs for about 80 minutes roughly. They don't get this tremendous decrease in audio size by lossless compression. Just find a dvd (usually live concerts) that has both PCM and ac3 audio tracks and compare them. The difference is quite significant.
Given that, only new movies (and old ones with all the tracks saved digitally and uncompressed would be able to benefit by having every channel in pcm. 7.1 channels in raw PCM probably take up about 7-8 gigs for a typical movie. This is all speculation, I'm not sure how the audio works on HD-DVD and BluRay and I didn't read the article.
This One actually loads. Since the page is obviously slashdotted.
Agreed. I was on a "field trip" at IBM's premier fab plant in East Fishkill, NY. When the Q/A portion came, someone asked what their yields were. They were allowed to say because it's IBM Confidential. But from what we gathered it's below 50% (this isn't read as being 49%, but the question was actually either below 50% or above 50%) for the 90nm process size.
Also, where would I go if I just want X or XX? Sometimes I'm not really in the mood for XXX, but X just hits the spot.
Ahh you have too many steps involved (and in the incorrect order.) Simply middle click.
Good idea, don't want to get kicked by punkbuster for having aimbots.
Yeah...well I'm gonna build my own Vista, with blackjack and hookers. In fact, forget the Vista.
But compared to the other big video store around here, Hollywood Video, this policy is much better. Typical rentals at blockbuster are for 7-8 days, so this is 8 days on top of that, which is 15-16 day rentals. I supposedly returned a few movies "late" to Hollywood video, and was shocked the next time when I had to pay about 9 dollars in late charges. The next time I rented a movie was with blockbuster.
Personally I used to like the 99 cent video stores run by asians. Great porn selection, 1 day rental (less lapse in memory in returning,) and only 99 cents. Want it for 2 days? Okay $1.98. The only problem here is with selection of the new movies, which they won't have 40 copies of. Selection could be spotty, but then again my local blockbuster didn't even have Deliverance.
There is also an article on him today on Day to Day and an obituary on All Things Considered. There are about a half dozen or so articles on him on NPR actually for those interested.
In other news about manufacting tolerances: Resistors spec'd at 5% tolerance are found to consistantly be 5% less than their stated value, in order to save money on raw materials.
You have any proof of this? While I'm not an expert on how resistors are made, I know they come in certain sizes because they are easier to make that way. Making them 5% off would be harder! That's why you have weird resistors values like 4.7k, because they are easier to make than 5k.
Unless the PLL in the motherboard is periodically recalibrated using (for instance) NIST, there's no way in hell that it's ever going to be accurate.
Do you even know how a PLL works? It basically takes a reference clock, and synchronizes the output to this reference clock (usually multiplied using a divider). If you are interested here's a brief introduction. The reference clock is a crystal. Crystals are accurate to about to many decimel places. That's why you get crystals that are quoted to be 15.00000MHz. The nature of a PLL is a control feedback system of the 2nd order. Basically, if the clock is off, it will automatically correct itself (as long as it stays in the lock range.) Basically the accuracy is going to come down to the power supply causing jitter, but that isn't going to affect the frequency. Additionally, PLL's are most likely going to be all on one integrated circuit, causing any manufacturing related tolerances, to be uniform through all the components, and therefore transistor ratios in the IC (which is what is important) will remain the same. Not to mention that there won't be any resistors in it, they will use transistors with the gate tied to the drain.
Internal clocks on inside these things are very accurate. If they were off, there would be all kinds of problems, data wouldn't line up. Even an Intel chip has multiple PLLs inside to syncronize the local clocks to each other in the clock distribution network.
I didn't read the article, but, how is the computers real time clock affecting any of these measurements at all? The benchmarking time is measured from clock ticks, from the cpu, which is very accurate.
Maybe because the highway between LA and Las Vegas (I-15) is one of the most dangerous roads in the US?
Not only is it dangerous because of speeding, but 1) It's in the middle of the desert 2) Most people driving on it are sleep deprived from partying all night and 3) are still drunk
That's just to mention a few risks. With soaring gas prices, it is probably cheaper to fly to vegas as well. When you are in Vegas, there is no need for a car. What really would of been nice is the super speed train from Vegas to Southern California, that I believe is now scrapped. It would take you to/from vegas in 75/90 minutes.
You're complaing about 5-CD installs? I guess you don't remember the days of the 15 3.5" disk installs. Not to mention it would probably come with 20 5.25" disks, in case you didn't have a state of the art 3.5" drive. After buying a few games, you're entire 100 disk capacity caddy would be full. Don't even get me started with games that used low density floppies...
That being said, when the first cd-rom games came out, a lot of people used the extra space to make really horrible games. I remember one game I received for free that was 3 CD's, which was unheard of, when most people had 1-2x speed cd-roms. The entire game was a movie, and every 2-3 minutes of script, you were able to make a choice of what to say next, to then watch another 2-3 minutes of script. Okay, if you're gonna make a movie and put it on cd's, thats one thing, but to make a movie and have me click every 2 minutes is not fun.
I don't remember the name of the game, but it was something along the lines of stopping a nuclear war or something.
Besides, once you are cooped up inside one of those natty suits that you have to wear in modern chip fabrication environments, believe me when I say that the lighting is *not* a major concern...
I was under the impression that modern chip fabrication environments were all automated and didn't require the full garb since all the wafers are enclosed and pushed around on air. As is explained in the 3rd paragraph of that article. Of course, I'll let you know in a few weeks, since we're taking a field trip there.
For us joe twelve packs (or just people that don't care about flashy boxes), the Retail BFG 6800 Ultra (and I think other models) come in a nice flashy white box.
Indeed, 468s on ebay are bit more than Pentium IIs. But isn't this like comparing cars to cpus?
Uhh...so are you trying to tell me the movie, Armageddon isn't realistic?
Naw, we're mostly harmless...
The point of this long-windedness is basically to say that they won't have any problem finding emergency landing strips in the (fairly unlikely, IMO) event that they need one.
It probablly has a slightly higher chance of failing since it only has one engine.
Simple Barter system:
1-2 Hours == A few beers or so.
3-4 Hours == A few more beers + some munchies.
5-8 Hours == Scotch + a stripper.
9-16 Hours == Stimulants + a massage.
17+ hours == A lot more stimulants.
True. The threshold voltage (the voltage necessary to turn a transistor "on" at the gate) is bottoming out around 0.2V. Below this when a transistor is suppose to be off, there is a significant (~10E-9A) amount of current and this increases exponentially w/ decrease in threshold voltage. Not only do you have problems with power, but stacked transistors are no longer at 0 volts in the intermediate junction when they are suppose to be, resulting in slower pull ups and pull downs.
Size of the chip is not a problem. Currently, about 80-90% of the chip is L2 cache, the actual CPU only consists of a very small corner of the chip. Overall, speed is more likely to increase, due to the lower amount of charge that is needed to turn on/off a gate (RC Delays.)