The RAM attached to video cards is not only fast, but has terrible latency. How do you think they clock it at 1GHz+? It wouldn't do at all for the main memory in an AMD64 machine.
That's pretty much how HT works already. Even with a few cm worth of traces, the latency hit is about 20% for the local CPU, and 60% for non-local ones. When your memory latency is at least 50ns to begin with (and that's cutting-edge --- the best FSB-based platforms are up at 80-100ns), an extra 5ns wire dely is really not a whole lot.
Sorry to break it to you, but $3000 is the going rate for laptops in the Macbook's class. Dual-core 2.16 GHz processors with 1GB of DDR2-667, 100GB 7200RPM HDDs, and 17" screens all in 1" form-factors don't come in $1000 machines. Have you priced out a Dell XPS notebook lately? The cheapest one, when outfitted comparably to the Macbook, sets you back $3400. Sure, you get a faster GPU and a higher-res screen, but that's an extra $600! For a Dell!
Your description of the steps is not accurate. In the K8's LSU, the three cycle latency is divided into the following tasks:
1) Address computation in the AGU 2) Cache address decode and virtual -> physical translation (in parallel) 3) Data cache access
In any case, the LSU is usually not the bottleneck for something like this. There is a reason why Intel's Core microarchitecture has a 3-cycle 32KB cache despite the P4 had a 2-cycle cache. The LSU can kill your latency, sure, but you're really limited mostly by the latency of the SRAM itself. All the 32KB-64KB physically-addressed data caches I'm aware of (the Power6's, the Power5's, the K8's, and the Core's) are 3-cycle latency. Indeed, the documents on Fujitsu's SPARC64 V processor go into detail about the tradeoff between cache size and latency. It points out that there was original consideration of an 8KB, 2-cycle data cache, but ultimately, they decided to go with a 128KB 4-cycle data cache. Given the variety of processor architectures (and thus LSU designs) that have not been able to achieve 2-cycle latency for a large (32KB+) L1 in a 2GHz+ processor suggests that the physical limitations of the access latency of large SRAMs is the bottleneck, rather than LSU design.
Actually, the cache access latency is dominated by the size of the cache. Intel managed 2-cycle L1s in the P4 because the data cache was only 8KB. The 32KB data cache in the NGMA is 3 cycles.
Yes, it would be. RAM takes up a lot of area. Have you ever looked at a RAM module? It's made up of 8-16 seperate chips. The densest single RAM chips are on the order of 128MB. Moreover, RAM is manufactured on very different (and much cheaper) processes than CPUs are. Certain types of RAM are compatible with certain CPU processes (eDRAM, for example), but these are not cheap, nor particularly dense.
If Apple loses quality control by going open source, their product won't meet the standards of their users.
Are you serious? Safari has crashed on me more times in six months than Konqueror ever did in a couple of years of using KDE.
I am sure many "technical" people like you find great solice in such denial.
It has nothing to do with me being "technical". Despite being a "technical" person, I'll readily admit that people in philosophy or history or literature are "intellectuals". What distinguishes an "intellectual" is a person whose work is cerebral, requiring a great deal of thought and careful logic. Business is not an intellectual pursuit. Business is about strategy, planning, charisma, and not more than a little bit of luck.
Have more moronic words ever been written?
Very possibly. Exactly what are you objecting to here? The idea that a creature, once dead, is just meat, or the technical description of where stem cell lines come from. Both are absolutely true, so if you find the truth moronic, well, that's your brain-damage.
I don't really understand why everybody is asking this question. Bathrooms and bedrooms are completely different. There just needs to be enough bedrooms to sleep however many people will live in the house. There have to be enough bathrooms so that there is one conveniently accessible from any part of the house. Think about it --- you don't want to have to go to the other end of the house just to wash your hands or take a leak! If you look at the fact that this house is 66,000 square feet, you can see it has a bathroom covering about every 2750 square feet, which is actually pretty big. Most 2750 square foot houses will have at least a few bathrooms.
1) China is far from being in the same category as governments like those of Angola or other African dicatorships. They're not great, but they're in a completely different league, these days, then many other countries you can list.
2) It is very rare that there are people who have no qualities worth admiration. Saying so is an emotional and irrational reaction.
That is a phenomenally stupid argument. Heads of state live and die by symbolic gestures. They do everything based on how the act will be perceived. The President of China isn't going to risk something like this being perceived the wrong way just to save a bit of gas or a few hours on a plane!
Gates, Harvard undergraduate dropout. GDub, Yale BA, Harvard MBA. Who is the intellectual?
MBA's aren't intellectuals. The intelligence required to get an MBA, even somewhere like Harvard, is a fraction of the intelligence required to get into Harvard as a technical major. Given that Gates has shown both far more intellectual capacity than Bush, technically, as well as having been orders of magnitude more successful as a businessman, I find your point to be positively silly.
If you call opposing the creation of a race of subhumans bread only for their stem cells to be anti-intellectual then I hope we have more of it. You are intellectually dishonest.
Who exactly is proposing creating a race of subhumans breed for their stem cells? Do you have a good grasp on how the technology works? The stem cells come from disposed fetuses (which are about as human as a piece of steak), and it is the cells that are cultured, not the fetuses.
The scientists conclude that "the surrounding area indicated the specimens were forest dwellers". Using the same logic, I suppose the millions of seashells in the Himalayas were also sea organisms that were forest dwellers.
How exactly do you reach that conclusion?
Beliefs on origins, regardless of what they are, never have been, nor will ever be, based on science, nor can they be.
This isn't a "belief about origin". This is evidence for something that happened *after* creation. Its a peculiar arrogance of man to believe that he was created with the universe --- the universe has existed (or was created) for billions of years, and it will continue to exist for billions of years after man is gone. Science cannot, yet, explain the act of creation, or event, but it can certainly explain how humans became the way they are.
Regardless of what facts and figures you conjure up, it cannot explain the eternal matter and acting force that started it all.
I love the use of "conjure" and "facts" in the same sentence. Scientists need to "magic" to arrive at their explanations. Their evidence is the world around us that can be plainly seen. Anyway, this is not a case regarding "eternal matter", its a theory regarding our very non-eternal selves.
Those necessarily have to be, without debate or variation, relegated to an article of faith. The only thing this news article says definitively is that these scientists have chosen to adopt evolution as their article of faith.
Bullshit. The varacity of evolution can be supported just as can any other branch of science. It cannot be "proven", but then again, quantum physics hasn't been "proven" either, yet it is still the underpinning of some of our biggest industries. Faith is belief in something --- evidence is not a prerequisite for faith. Science doesn't operate on faith, their "beliefs" are of a very different nature.
If Nintendo loses money on the Revolution, somebody needs to be fired. The damn thing is a die-shrunk Gamecube, it should be even cheaper to manufacture than the original.
What if that one game that's awesome is in a genre you couldn't care less about? Super Mario 64 is a great platformer, but I hate platformers. Metroid and Zelda are great action-adventure titles, and guess what, I hate action-adventure games. There is no depth in any particular genre in Nintendo's library. You've got maybe a handful of good titles in any genre. Unless you're the kind of person you can pick up any type of game and enjoy playing it, or are the type of person who can play the same game over and over again, you'll quickly run out of stuff to play on a Nintendo console.
As for variety --- you do realize that out of the five games you mentioned, one isn't a Nintendo game, and the other three feature pretty much the same characters?
My point was to compare the PS* libraries to the N64 and Gamecube libraries. With the N64 and beyond, Nintendo's mantra "was quality not quantity". For me, that meant a bunch of highly rated games (the Zelda series, the Mario series and its spin-offs, the Metroid series), that I had zero interest in playing. The PS*'s library is like the SNES's. You like arcade flight simulators? You've got a dozen to choose from. You like Japanese-style RPG? You have several dozen. You like first person shooters? Line them up. Sure, a lot of them are mediocre, and some are just bad, but there are a lot of good ones too, and I'd rather play a mediocre game in a genre I like then a steller game in a genre I don't.
I think there's at least 20 games still on my "definately want to play" list for the PS2 that I haven't got to yet.
Exactly. I don't game much, but I just finished up Xenosaga II over the break. XSII wasn't a AAA game, but I'm a fan of the series and I enjoyed playing it quite a bit. Right now, I'm looking forward to playing some of the Shadow Hearts titles, of which I have two to catch up on before I get to the new one. Games like these don't come out for Nintendo consoles. Its as if everything is either a Kingdom Hearts II (super highly rated AAA title), or a Quest 64 (crap).
I agree with you. I bought a PSP earlier this year, and didn't use it for awhile because of the drought of good games (some of which is natural to a new console, of course). Recently, I bought Syphon Filter and Grand Theft Auto, then loaded up a couple of emulators, and am having a blast. Hopefully the next wave of titles coming takes a cue from games like SF.
I hate the "quality not quantity" mantra. That's the main reason why post-SNES Nintendo game libraries have sucked so hard. It totally eliminates the variety of titles that make a system worth owning.
The fundemental problem is that the compiler doesn't know at runtime exactly what the dependencies will be. Branches and memory operations, which are extremely common in most software, create dependencies that the compiler cannot analyze at compile-time, but the processor can analyze at run-time. In the real-world, in-order versus out-of-order isn't just a matter of code scheduling, but fundementally limits the types of code you can run at high speed.
Because Jews are really smart. No, seriously. Why do you think they're so rich? Studies have actually shown that there is a sub-population of Jews that gets Nobel Prizes vastly out of proportion with their numbers.
Slashdot Mirror
The RAM attached to video cards is not only fast, but has terrible latency. How do you think they clock it at 1GHz+? It wouldn't do at all for the main memory in an AMD64 machine.
That's pretty much how HT works already. Even with a few cm worth of traces, the latency hit is about 20% for the local CPU, and 60% for non-local ones. When your memory latency is at least 50ns to begin with (and that's cutting-edge --- the best FSB-based platforms are up at 80-100ns), an extra 5ns wire dely is really not a whole lot.
The 15.4" MacBook Pro is a 5.6lb Intel Mac. Indeed, it was the first one to come out, and has been out since February.
Sorry to break it to you, but $3000 is the going rate for laptops in the Macbook's class. Dual-core 2.16 GHz processors with 1GB of DDR2-667, 100GB 7200RPM HDDs, and 17" screens all in 1" form-factors don't come in $1000 machines. Have you priced out a Dell XPS notebook lately? The cheapest one, when outfitted comparably to the Macbook, sets you back $3400. Sure, you get a faster GPU and a higher-res screen, but that's an extra $600! For a Dell!
No, I wasn't autistic as a child...
Your description of the steps is not accurate. In the K8's LSU, the three cycle latency is divided into the following tasks:
1) Address computation in the AGU
2) Cache address decode and virtual -> physical translation (in parallel)
3) Data cache access
In any case, the LSU is usually not the bottleneck for something like this. There is a reason why Intel's Core microarchitecture has a 3-cycle 32KB cache despite the P4 had a 2-cycle cache. The LSU can kill your latency, sure, but you're really limited mostly by the latency of the SRAM itself. All the 32KB-64KB physically-addressed data caches I'm aware of (the Power6's, the Power5's, the K8's, and the Core's) are 3-cycle latency. Indeed, the documents on Fujitsu's SPARC64 V processor go into detail about the tradeoff between cache size and latency. It points out that there was original consideration of an 8KB, 2-cycle data cache, but ultimately, they decided to go with a 128KB 4-cycle data cache. Given the variety of processor architectures (and thus LSU designs) that have not been able to achieve 2-cycle latency for a large (32KB+) L1 in a 2GHz+ processor suggests that the physical limitations of the access latency of large SRAMs is the bottleneck, rather than LSU design.
Actually, the cache access latency is dominated by the size of the cache. Intel managed 2-cycle L1s in the P4 because the data cache was only 8KB. The 32KB data cache in the NGMA is 3 cycles.
Yes, it would be. RAM takes up a lot of area. Have you ever looked at a RAM module? It's made up of 8-16 seperate chips. The densest single RAM chips are on the order of 128MB. Moreover, RAM is manufactured on very different (and much cheaper) processes than CPUs are. Certain types of RAM are compatible with certain CPU processes (eDRAM, for example), but these are not cheap, nor particularly dense.
If Apple loses quality control by going open source, their product won't meet the standards of their users. Are you serious? Safari has crashed on me more times in six months than Konqueror ever did in a couple of years of using KDE.
I am sure many "technical" people like you find great solice in such denial.
It has nothing to do with me being "technical". Despite being a "technical" person, I'll readily admit that people in philosophy or history or literature are "intellectuals". What distinguishes an "intellectual" is a person whose work is cerebral, requiring a great deal of thought and careful logic. Business is not an intellectual pursuit. Business is about strategy, planning, charisma, and not more than a little bit of luck.
Have more moronic words ever been written?
Very possibly. Exactly what are you objecting to here? The idea that a creature, once dead, is just meat, or the technical description of where stem cell lines come from. Both are absolutely true, so if you find the truth moronic, well, that's your brain-damage.
I don't really understand why everybody is asking this question. Bathrooms and bedrooms are completely different. There just needs to be enough bedrooms to sleep however many people will live in the house. There have to be enough bathrooms so that there is one conveniently accessible from any part of the house. Think about it --- you don't want to have to go to the other end of the house just to wash your hands or take a leak! If you look at the fact that this house is 66,000 square feet, you can see it has a bathroom covering about every 2750 square feet, which is actually pretty big. Most 2750 square foot houses will have at least a few bathrooms.
1) China is far from being in the same category as governments like those of Angola or other African dicatorships. They're not great, but they're in a completely different league, these days, then many other countries you can list.
2) It is very rare that there are people who have no qualities worth admiration. Saying so is an emotional and irrational reaction.
That is a phenomenally stupid argument. Heads of state live and die by symbolic gestures. They do everything based on how the act will be perceived. The President of China isn't going to risk something like this being perceived the wrong way just to save a bit of gas or a few hours on a plane!
Gates, Harvard undergraduate dropout. GDub, Yale BA, Harvard MBA. Who is the intellectual?
MBA's aren't intellectuals. The intelligence required to get an MBA, even somewhere like Harvard, is a fraction of the intelligence required to get into Harvard as a technical major. Given that Gates has shown both far more intellectual capacity than Bush, technically, as well as having been orders of magnitude more successful as a businessman, I find your point to be positively silly.
If you call opposing the creation of a race of subhumans bread only for their stem cells to be anti-intellectual then I hope we have more of it. You are intellectually dishonest.
Who exactly is proposing creating a race of subhumans breed for their stem cells? Do you have a good grasp on how the technology works? The stem cells come from disposed fetuses (which are about as human as a piece of steak), and it is the cells that are cultured, not the fetuses.
The scientists conclude that "the surrounding area indicated the specimens were forest dwellers". Using the same logic, I suppose the millions of seashells in the Himalayas were also sea organisms that were forest dwellers.
How exactly do you reach that conclusion?
Beliefs on origins, regardless of what they are, never have been, nor will ever be, based on science, nor can they be.
This isn't a "belief about origin". This is evidence for something that happened *after* creation. Its a peculiar arrogance of man to believe that he was created with the universe --- the universe has existed (or was created) for billions of years, and it will continue to exist for billions of years after man is gone. Science cannot, yet, explain the act of creation, or event, but it can certainly explain how humans became the way they are.
Regardless of what facts and figures you conjure up, it cannot explain the eternal matter and acting force that started it all.
I love the use of "conjure" and "facts" in the same sentence. Scientists need to "magic" to arrive at their explanations. Their evidence is the world around us that can be plainly seen. Anyway, this is not a case regarding "eternal matter", its a theory regarding our very non-eternal selves.
Those necessarily have to be, without debate or variation, relegated to an article of faith. The only thing this news article says definitively is that these scientists have chosen to adopt evolution as their article of faith.
Bullshit. The varacity of evolution can be supported just as can any other branch of science. It cannot be "proven", but then again, quantum physics hasn't been "proven" either, yet it is still the underpinning of some of our biggest industries. Faith is belief in something --- evidence is not a prerequisite for faith. Science doesn't operate on faith, their "beliefs" are of a very different nature.
If Nintendo loses money on the Revolution, somebody needs to be fired. The damn thing is a die-shrunk Gamecube, it should be even cheaper to manufacture than the original.
To be fair, the GB isn't a console, its a handheld, its priced that way.
What if that one game that's awesome is in a genre you couldn't care less about? Super Mario 64 is a great platformer, but I hate platformers. Metroid and Zelda are great action-adventure titles, and guess what, I hate action-adventure games. There is no depth in any particular genre in Nintendo's library. You've got maybe a handful of good titles in any genre. Unless you're the kind of person you can pick up any type of game and enjoy playing it, or are the type of person who can play the same game over and over again, you'll quickly run out of stuff to play on a Nintendo console.
As for variety --- you do realize that out of the five games you mentioned, one isn't a Nintendo game, and the other three feature pretty much the same characters?
My point was to compare the PS* libraries to the N64 and Gamecube libraries. With the N64 and beyond, Nintendo's mantra "was quality not quantity". For me, that meant a bunch of highly rated games (the Zelda series, the Mario series and its spin-offs, the Metroid series), that I had zero interest in playing. The PS*'s library is like the SNES's. You like arcade flight simulators? You've got a dozen to choose from. You like Japanese-style RPG? You have several dozen. You like first person shooters? Line them up. Sure, a lot of them are mediocre, and some are just bad, but there are a lot of good ones too, and I'd rather play a mediocre game in a genre I like then a steller game in a genre I don't.
I think there's at least 20 games still on my "definately want to play" list for the PS2 that I haven't got to yet.
Exactly. I don't game much, but I just finished up Xenosaga II over the break. XSII wasn't a AAA game, but I'm a fan of the series and I enjoyed playing it quite a bit. Right now, I'm looking forward to playing some of the Shadow Hearts titles, of which I have two to catch up on before I get to the new one. Games like these don't come out for Nintendo consoles. Its as if everything is either a Kingdom Hearts II (super highly rated AAA title), or a Quest 64 (crap).
I agree with you. I bought a PSP earlier this year, and didn't use it for awhile because of the drought of good games (some of which is natural to a new console, of course). Recently, I bought Syphon Filter and Grand Theft Auto, then loaded up a couple of emulators, and am having a blast. Hopefully the next wave of titles coming takes a cue from games like SF.
I hate the "quality not quantity" mantra. That's the main reason why post-SNES Nintendo game libraries have sucked so hard. It totally eliminates the variety of titles that make a system worth owning.
The fundemental problem is that the compiler doesn't know at runtime exactly what the dependencies will be. Branches and memory operations, which are extremely common in most software, create dependencies that the compiler cannot analyze at compile-time, but the processor can analyze at run-time. In the real-world, in-order versus out-of-order isn't just a matter of code scheduling, but fundementally limits the types of code you can run at high speed.
The T1 doesn't do branch processing on multiple cores. Its just a giant array of fairly standard (and slow), in order UltraSPARC II processors.
Because Jews are really smart. No, seriously. Why do you think they're so rich? Studies have actually shown that there is a sub-population of Jews that gets Nobel Prizes vastly out of proportion with their numbers.