The entire show is taped a couple of days in advance of the actual air date. Gives them plenty of time to edit stuff out. They even do retakes when things don't quite work out.
Power 4 has thicker gate oxides than the 970. The 970 is a desktop chip, the Power 4 is a server chip.
Apparently, in order to increase the reliability of the Power4 for the high-end server market, IBM used much thicker gate oxides on the chip's transistors. The trade-off for this decreased failure rate and improved reliability was that the Power4's transistors have slower switching speeds, so even with process shrinks it's harder to push the design to higher clock speeds. Since the 970 is made for the desktop market, there's no need for such measures and therefore the new chip's clock speed will scale much higher than the Power4's. In sum, the 970 is made to be faster, cheaper, and significantly less reliable than the Power4."
"go with a sturdy, stable desktop for the four years"
Do you know how much faster processors get in four years, and how much cheaper memory and HDs get? Sadly I do. My Celeron 300a is still with me. I hate beeing a poor student.
On a positive note, by the second year your computer is so obsolete you can't play new games on it anymore. Gives you more time to study... unless you spend your time reading/.
Actually, the models don't have a super hi polycount. They start out as really hi-poly models but then to create ingame models they use the original to make really good bumpmaps and environment maps. The thing about Doom3 is the textures and the shadows. Not some super high polygon count.
In any comparison like this the P4 is the odd man out. The pipeline design of the P4 is like nothing else. Depending on how you count there are 20 stages, but there is an extra 8 stages in the frontend, the decode stage. The P4 saves decoded instructions in the L1 or Trace Cache, this includes branchpredicted instructions. There is no regular L1 Cache as in the PowerPC and Athlon.
Ars have an article comparing the Athlon to the G4, and there are more similarities than there are differences. The P4 works in a radically different way. Granted the PowerPC 970 is not the G4 but saying that the pipelines of the PowerPC and the P4 is similar would be a stretch. All you really can say is that both have a long execution pipeline. As the clocks ramp up, this is to be expected. The AMD Athlon 64 has two more stages than the Athlon.
While dual FPUs are useful this is what SIMD operations like Intels SSE and Motorolas Altivec are used for. The dual FPUs might be good for doing other work but DSP filters and Photoshop filters will use SIMD operations wherever they can.
That is very bandwidth intensive work, moving alot of floating point numbers from memory, and this is where the 970 will be superior to the G4e. But this is also the strong point of the Intel P4 running at super high frequencies. The AMD Athlon 64 will clock for clock be competitive with the P4 running genreal code, but doing SIMD operations they can all do 4 at a time. Then the higher clocked chip always wins. The Altivec unit of the 970 will have to be alot better than the SSE2 from Intel to beat it.
Oh, so your "secretary" would have problems with Linux... Right.
Hee hee, looks more like _you_ couldn't figure it out. With the cdrom in fstab you would just need/cdrom not the dev-part and btw, the command is 'umount':P
Hehe that makes me think of when I was in high-school. We were supposed to think up a moral problem, then play a little scenario in class like "you find pile of cash, do you hand it to the police or keep it?". You would say one thing was the right thing to do and debate with the class in terms of moralas and ethics and the other groups would have to take a stance, was this right or wrong. I figured - lets sit on our ass and don't do anything, then argue we still completed the assignment. We had a pretty good case going for us until some Miss Goody Two Shoes said we were just lazy and wanted to get away without doing the actual work, the other guys were rooting for us but the teacher eventually sided with the girls. Thus we completed the assignment:)
"The G4 spreads it all out over multiple pipes, but the pipes arent deep. The main work is figuring out which pipe is free to shove stuff into.... The 970, on the other hand, has more pipes than the G4 and the Pentium 4, but the pipes are deeper than the P4. So it can stuff a whole ton of stuff down and be very efficient. Wide and deep. Theres a bit of a tradeoff, but the chip is just engineered much better."
No the 970 is not as deeply pipelined as the P4, but it is pretty deep, alot more steps than the G4e. Deeper pipeline means you can clock it higher. But Intel will probably be able to reach much higher still.
The new Intel 875 mobos are at 800Mhz FSB today. The 970 is not released and no samples have been shown.
It is an interesting chip but I'll wait for real benchmarks before I get too excited. If good engineering was all it took x86 would be dead long ago, but it lives on and the x86-64 will only prolong it's life. Because lets face it: only performance matters.
No doom9 is a DVD backup resource. What matters is getting a DVD onto one or two CDs. Streaming and low bitrate might have their applications but ripping movies is not one of them.
"I'm not dissing his abilities or role in history. He has done shit that most of us could never come close to surpassing."
He worked alone 24/7 for years writing a new operating system to give it away. That is, whenever he didn't sleep in the office, he was busy coding. His connection to MIT back then gave him the ability and resources to do so. No one could ever come close to that because this is a different day and age, the rules are not the same (cheap PCs, the WWW etc).
And the notion of copyleft, it is already "invented" so you can't just create a license and have it change an industry like the GPL and other free licenses have.
Sweden shipped steel and iron to both sides. This is how it works.
1. export to nations at war 2. stay neutral 3. profit
Sweden is a small country between the superpowers. Allies to the west, Germany south and Soviet to the east. It's not like Sweden were going out of their way to sell weapons to countries at war with each other on the other side of the globe... like say the US in the Iran-Iraq conflict.
There is one small problem with space elevators. The cable has to carry it's own weight. If you were to use a steel wire of a few millimeters in diameter at the surface, the diameter at geostationary orbit would be about the diameter of the solar system. This is not including the counter weight.
This might seem like complete nonsens but that is from a theoretical physicist and writer, Dr Hans-Uno Bengtsson. The original reply from a "ask the expert column" in swedens biggest newspaper. Fungerar en rymdhiss?.
BTW: If the wire were to be made from kevlar it would only have to be a few hundred meters in diameter. With more exotic materials you could shrink it to less than a meter.
Build. Aaaw, the memories... I still have a couple hundred laser printed pages. I would sneak into the school PC lab at night and print the docs. Build was a great entry into level making and mod creation. A friend of mine works as a level designer and content creator today, involved in some pretty well known titles. Started out with build as I guess many did. Build was fun to use. I had as much fun with build and the community around it as with the game itself. I can still find some of my old maps using google, that is pretty cool imho.
I find it amazing Linus and crew get any code done when they write a 5 line patch then spend 15 posts in lkml to debate it. And that's just Linus posts that kerneltrap found interesting.
What everyone on unix does is use multiple desktops. Together with a pager you get a kind of tabbed interface.
Some usability experts say that switching between desktops is disruptive and I would think this is the reason it's not default in Windows. Some would like to see interfaces based on zoom and pan instead. The fastest way to move around an area larger than your screen is not to pan around but to zoom out and back in on the area of interest. Combine this with a 2x2 or 3x3 virtual desktop and you get the virtual desktops without the disruptive switches. Like a snap to grid for zooming to desktops. You get the overview that helps your sence of location place objects in relation to each other.
I don't think the switches are disruptive. In fact I use virtual desktops like Windows-people use ALT+TAB. Only that ALT+TAB sucks when you have more than two applications running. I don't understand how people can stand having 15-20 Windows open at once, all overlapping. Using desktops you get fullscreen for your applications or place a few programs tiled on one desktop. Interesting to note is that explicitly tiled UIs reduce the time users spend managing windows. Up to 25% of the time to perform a task can be spent managing windows in the standard WIMP interface. Using 21" monitor I don't beleive in UIs purely based on fullscreen mode but similar effects using tiles and multiple desktops migth work better.
And if you actually read the article you would know they estimated the "exhaust" to be at most 3-4 degrees warmer than the incoming water.
BTW: Am I the only one who read about this new underwater jet and got disappointed. The idea is cool but it still runs on petrol or diesel. Back to the stone age...
"Some CPUs only allow one or the other of these. x86 uses both. (A short jump is an 8 bit signed jump, -128/+127 offset from your current location. A near jump is 16 bit. A far jump specifies a segment and offset, because x86 uses a segmented memory model.)......In x86 this is coupled with the CS register (code segment) and it is called the instruction pointer (IP or on 32 bit CPUs in 32 bit mode, EIP) and so you load the new instruction from CS:IP. As per my above paragraph a short or near jump updates IP or EIP, a far jump updates CS and [E]IP."
Now that kind of talk is exactly the reason you are just a geek and "Hannibal" gets to write articles explaining Microprocessors. You only know of one type and fail miserably to explain the concepts of it's innner workings.
I'm amazed you missed every part of the article that explained what it was about and what he tried to achieve with it. Skipped both intro and conclusion, eh?
I'm not so amazed you actually got modded up, this is a geek site and that was a whole lot of geekspeak, but the moderators clearly never read or understood the article.
"The stuff in this comment I'm writing now, on the other hand, is based on a class in x86 assembly, the final for which is on this coming Tuesday."
I really hope you teach that class, otherwise you just made a giant fool of yourself.
Hannibal if you read this - that was the best intro to microprocessors I ever read. Keep up the good work!
Slashdot Mirror
Torrent is good and all but you probably will get better speed with the direct downloads from the mirrors. 15 kb/sec vs 1500 kb/sec. Your choice.
But I live close to the arctic circle, you insensitive clod.
That was 580 kilometers per hour. Not miles per hour.
The entire show is taped a couple of days in advance of the actual air date. Gives them plenty of time to edit stuff out. They even do retakes when things don't quite work out.
Power 4 has thicker gate oxides than the 970. The 970 is a desktop chip, the Power 4 is a server chip.
Apparently, in order to increase the reliability of the Power4 for the high-end server market, IBM used much thicker gate oxides on the chip's transistors. The trade-off for this decreased failure rate and improved reliability was that the Power4's transistors have slower switching speeds, so even with process shrinks it's harder to push the design to higher clock speeds. Since the 970 is made for the desktop market, there's no need for such measures and therefore the new chip's clock speed will scale much higher than the Power4's. In sum, the 970 is made to be faster, cheaper, and significantly less reliable than the Power4."
Inside the IBM PowerPC 970
"go with a sturdy, stable desktop for the four years"
/.
Do you know how much faster processors get in four years, and how much cheaper memory and HDs get? Sadly I do. My Celeron 300a is still with me. I hate beeing a poor student.
On a positive note, by the second year your computer is so obsolete you can't play new games on it anymore. Gives you more time to study... unless you spend your time reading
digital camera
Actually, the models don't have a super hi polycount. They start out as really hi-poly models but then to create ingame models they use the original to make really good bumpmaps and environment maps. The thing about Doom3 is the textures and the shadows. Not some super high polygon count.
In any comparison like this the P4 is the odd man out. The pipeline design of the P4 is like nothing else. Depending on how you count there are 20 stages, but there is an extra 8 stages in the frontend, the decode stage. The P4 saves decoded instructions in the L1 or Trace Cache, this includes branchpredicted instructions. There is no regular L1 Cache as in the PowerPC and Athlon.
Ars have an article comparing the Athlon to the G4, and there are more similarities than there are differences. The P4 works in a radically different way. Granted the PowerPC 970 is not the G4 but saying that the pipelines of the PowerPC and the P4 is similar would be a stretch. All you really can say is that both have a long execution pipeline. As the clocks ramp up, this is to be expected. The AMD Athlon 64 has two more stages than the Athlon.
While dual FPUs are useful this is what SIMD operations like Intels SSE and Motorolas Altivec are used for. The dual FPUs might be good for doing other work but DSP filters and Photoshop filters will use SIMD operations wherever they can.
That is very bandwidth intensive work, moving alot of floating point numbers from memory, and this is where the 970 will be superior to the G4e. But this is also the strong point of the Intel P4 running at super high frequencies. The AMD Athlon 64 will clock for clock be competitive with the P4 running genreal code, but doing SIMD operations they can all do 4 at a time. Then the higher clocked chip always wins. The Altivec unit of the 970 will have to be alot better than the SSE2 from Intel to beat it.
"unmount /dev/hdd /cdrom"
/cdrom not the dev-part and btw, the command is 'umount' :P
Oh, so your "secretary" would have problems with Linux... Right.
Hee hee, looks more like _you_ couldn't figure it out. With the cdrom in fstab you would just need
Hehe that makes me think of when I was in high-school. We were supposed to think up a moral problem, then play a little scenario in class like "you find pile of cash, do you hand it to the police or keep it?". You would say one thing was the right thing to do and debate with the class in terms of moralas and ethics and the other groups would have to take a stance, was this right or wrong. I figured - lets sit on our ass and don't do anything, then argue we still completed the assignment. We had a pretty good case going for us until some Miss Goody Two Shoes said we were just lazy and wanted to get away without doing the actual work, the other guys were rooting for us but the teacher eventually sided with the girls. Thus we completed the assignment :)
"The G4 spreads it all out over multiple pipes, but the pipes arent deep. The main work is figuring out which pipe is free to shove stuff into. ...
The 970, on the other hand, has more pipes than the G4 and the Pentium 4, but the pipes are deeper than the P4. So it can stuff a whole ton of stuff down and be very efficient. Wide and deep. Theres a bit of a tradeoff, but the chip is just engineered much better."
No the 970 is not as deeply pipelined as the P4, but it is pretty deep, alot more steps than the G4e. Deeper pipeline means you can clock it higher. But Intel will probably be able to reach much higher still.
The new Intel 875 mobos are at 800Mhz FSB today. The 970 is not released and no samples have been shown.
It is an interesting chip but I'll wait for real benchmarks before I get too excited. If good engineering was all it took x86 would be dead long ago, but it lives on and the x86-64 will only prolong it's life. Because lets face it: only performance matters.
No doom9 is a DVD backup resource. What matters is getting a DVD onto one or two CDs. Streaming and low bitrate might have their applications but ripping movies is not one of them.
"I'm not dissing his abilities or role in history. He has done shit that most of us could never come close to surpassing."
He worked alone 24/7 for years writing a new operating system to give it away. That is, whenever he didn't sleep in the office, he was busy coding. His connection to MIT back then gave him the ability and resources to do so. No one could ever come close to that because this is a different day and age, the rules are not the same (cheap PCs, the WWW etc).
And the notion of copyleft, it is already "invented" so you can't just create a license and have it change an industry like the GPL and other free licenses have.
So no my friend, none of "you" could come close.
Sweden shipped steel and iron to both sides. This is how it works.
1. export to nations at war
2. stay neutral
3. profit
Sweden is a small country between the superpowers. Allies to the west, Germany south and Soviet to the east. It's not like Sweden were going out of their way to sell weapons to countries at war with each other on the other side of the globe... like say the US in the Iran-Iraq conflict.
There is one small problem with space elevators. The cable has to carry it's own weight. If you were to use a steel wire of a few millimeters in diameter at the surface, the diameter at geostationary orbit would be about the diameter of the solar system. This is not including the counter weight.
This might seem like complete nonsens but that is from a theoretical physicist and writer, Dr Hans-Uno Bengtsson. The original reply from a "ask the expert column" in swedens biggest newspaper. Fungerar en rymdhiss?.
BTW: If the wire were to be made from kevlar it would only have to be a few hundred meters in diameter. With more exotic materials you could shrink it to less than a meter.
Build. Aaaw, the memories... I still have a couple hundred laser printed pages. I would sneak into the school PC lab at night and print the docs. Build was a great entry into level making and mod creation. A friend of mine works as a level designer and content creator today, involved in some pretty well known titles. Started out with build as I guess many did. Build was fun to use. I had as much fun with build and the community around it as with the game itself. I can still find some of my old maps using google, that is pretty cool imho.
I find it amazing Linus and crew get any code done when they write a 5 line patch then spend 15 posts in lkml to debate it. And that's just Linus posts that kerneltrap found interesting.
That's GLAME to you mister.
One word: consistency
What everyone on unix does is use multiple desktops. Together with a pager you get a kind of tabbed interface.
Some usability experts say that switching between desktops is disruptive and I would think this is the reason it's not default in Windows. Some would like to see interfaces based on zoom and pan instead. The fastest way to move around an area larger than your screen is not to pan around but to zoom out and back in on the area of interest. Combine this with a 2x2 or 3x3 virtual desktop and you get the virtual desktops without the disruptive switches. Like a snap to grid for zooming to desktops. You get the overview that helps your sence of location place objects in relation to each other.
I don't think the switches are disruptive. In fact I use virtual desktops like Windows-people use ALT+TAB. Only that ALT+TAB sucks when you have more than two applications running. I don't understand how people can stand having 15-20 Windows open at once, all overlapping. Using desktops you get fullscreen for your applications or place a few programs tiled on one desktop. Interesting to note is that explicitly tiled UIs reduce the time users spend managing windows. Up to 25% of the time to perform a task can be spent managing windows in the standard WIMP interface. Using 21" monitor I don't beleive in UIs purely based on fullscreen mode but similar effects using tiles and multiple desktops migth work better.
And if you actually read the article you would know they estimated the "exhaust" to be at most 3-4 degrees warmer than the incoming water.
BTW: Am I the only one who read about this new underwater jet and got disappointed. The idea is cool but it still runs on petrol or diesel. Back to the stone age...
"so please cut the bullshit and post real news stories"
You are new to slashdot, right?
"Some CPUs only allow one or the other of these. x86 uses both. (A short jump is an 8 bit signed jump, -128/+127 offset from your current location. A near jump is 16 bit. A far jump specifies a segment and offset, because x86 uses a segmented memory model.)... ...In x86 this is coupled with the CS register (code segment) and it is called the instruction pointer (IP or on 32 bit CPUs in 32 bit mode, EIP) and so you load the new instruction from CS:IP. As per my above paragraph a short or near jump updates IP or EIP, a far jump updates CS and [E]IP."
Now that kind of talk is exactly the reason you are just a geek and "Hannibal" gets to write articles explaining Microprocessors. You only know of one type and fail miserably to explain the concepts of it's innner workings.
I'm amazed you missed every part of the article that explained what it was about and what he tried to achieve with it. Skipped both intro and conclusion, eh?
I'm not so amazed you actually got modded up, this is a geek site and that was a whole lot of geekspeak, but the moderators clearly never read or understood the article.
"The stuff in this comment I'm writing now, on the other hand, is based on a class in x86 assembly, the final for which is on this coming Tuesday."
I really hope you teach that class, otherwise you just made a giant fool of yourself.
Hannibal if you read this - that was the best intro to microprocessors I ever read. Keep up the good work!