Get an American Express for students credit card. They'll give it to you. Use it responsibly (i.e. don't forget to make a payment every month), and watch your credit rating rise... I've done the same, and got to a point were every week I was getting ~2 to 3 new card applications.
There were multiple causes, but religious dogma certainly played a part. For example, could you get a proper education outside the church system (as in - without being a cleric/monk)? And what was the literacy rate, say, compared to Roman times?
BTW, in the parent post I was referring primarily to the West, not the Eastern Empire.
Yeap, if you have a server app (web/app/db server) with a lot of inherent parallelism, that's great! Niagra kicks ass. As you were saying - that won't parallelize single-threaded code though. You could, I guess, do some speculative parallelization (there are a few very good research proposals here). OTOH you could also try various tricks to increase your out-of-order instruction window (there are some good proposals here as well).
BTW, I wasn't trying to say that stacked architectures are not worth it, I just didn't like the blank statement from the story: "stacked architectures are better than register-based ones". In architecture one size doesn't fit all (which can be applied the other way around, too, you wouldn't want out-of-order execution for an embedded app)
I didn't even try to torrent-download the thing, but I can tell you why stack machines aren't better than register-based ones. The main reason is that it's much much much harder to do renaming of a stack than of"regular" registers. Without renaming you can't do out-of-order execution... Currently, there are two "mainstream" architectures that include stacked register files: Itanium and Sparc. Neither of them have out-of-order implementations.
But why do you need out-of-order execution? Well, misses to memory are very expensive these days - it can easily take from 200 to 400 cycles to service a load that misses all the way to main memory. This can have a significant effect on performance. What out-of-order execution does is to allow independent instructions that are younger than the load to execute in parallel with it. Quite often these parallely-executed instruction will generate other misses to main memory, overlapping their latencies. So - latency of loads that miss is still very high, but at the very least the processor is not idle while servicing them (for a good read see "MLP Yes! ILP no!" by Andy Glew)
Itanium and Sparc compensate for the fact that they don't do stuff out-of-order by putting sh*tloads of L2/3 cache on-chip. The cost of a miss is still very high, but it happens much less often. The manufacturing cost of a chip is also much higher.
Note that what NASA is sending into space is "old" tech. The reason - well, cosmic rays are much stronger in outer space, and the smaller the gate, the easier it is for them to flip its state.
... even though your post implies that it's the thing that will "save us" from the evils of a represive state. Especially when the kind of "opportunities" you're talking about may simply be illegal. I'm not a tax lawyer, but what you're talking about seems to be tax evasion. You also seem to forget that the state generally has a lot more power than market forces. Market forces can be "repressive" in that they can drive a business to bankruptcy. The state, on the other hand, can imprison you and even kill you if it so desires. Let me give you an example. At some point in Nazi Germany prices were controlled very tightly, and there was also an oversupply of money. This should have led to a black "free" market, but it didn't. The reason? Well, the penalties for doing black market activities were incredibly high (like, death).
I also find it interesting that you seem so upset about losing your right to smoke in a restaurant, but, OTOH, you don't seem to consider another person's right to breathe clean air (especially in a restaurant, when one is supposed to enjoy him/herself). When you smoke you're not only affecting negatively your own health (which, as far as I'm concerned, is your absolute right), but also the health of the people around you. Free market doesn't really work here, since most owners will not put "no smoking" signs, considering that it limits their clientelle. So a person who hates smoking, like me, only has two choices: eat out and inhale very annoying and toxic fumes, or eat home.
I've heard that the first Core Duo is based on the PIII core (with many modification though). Not sure about Conroe though. Happen to have an "authoritative" reference?
Also, what do you mean by "excellent on-die cache control"? Cache control is always on-die (it's the cache data that can sometimes be off-chip).
Intel has been betting on two "mediocre" horses for too long. The first - Pentium 4 - has been successful initially, with its high clock rates, but in the end people figured out that it doesn't deliever performance (regardless, they kept it at their flagship for ~5years).
The second - Itanium - as far as I'm concerned, it's simply a step backwards. A processor these days is most of the times limited by the slow memory (it can easily take 200-300 cycles to service a request from memory, as opposed to 2-3 cycles from the L1 cache or 6-20 cycles from the L2). Out-of-order execution (Pentium Pro and after) alleviates the problem to a certain extent, by allowing other instructions that do not depend on the result of the instruction that missed to execute. So the processor can still do something while servicing the miss (quite often it executes other loads that miss, effectively increasing the memory-level parallelism of the processor). Because Itanium executes instructions in order, it simply can't do that. Furthermore the compiler can't tell which instructions are going to miss (it needs a profiler to figure frequently-missing instructions, and only then it can generate prefetches). Intel's solution - let's throw shitloads of caches on the Itanium, to reduce the occurence of the misses. Of course, that makes the chip huge, considerably more expensive, etc.
Nevertheless, Conroe seems to be an awesome chip. Time to buy some INTC shares...
You buy a license to play the content bound to the medium. When the medium dies... well tough luck. Even though the current copyright law allows you to make a backup copy, you apparently wave that right when accepting the license...
In your average laptop, the power consumed by a CPU when running something (i.e. not just idling around) is about half the total power. The other half, roughly, is consumed by the screen.
Current superscalars still fetch instructions in order, and squash everything after a mis-predicted branch. The cost of branch mis-speculation is in fact getting higher, because deeper pipelines means longer times between the mis-prediction of the branch and the execution (where the correction takes place). In other words, it means longer times on the wrong path.
The purpose of "good dispatching" (i.e. out-of-order execution) is to hide the latencies of misses to main memory (it takes between 200 and 400 cycles these days to get something from memory, assuming that the memory bus isn't saturated), by executing instructions following the miss but not dependent on it. Out-of-order execution has been around Pentium Pro, btw.
West Philadelphia to be more precise. I've seen cars broken for much less valuable stuff... (like a pack of softdrinks, if you can believe it). Living your purse in a car in a conspicuous place is simply dumb around here.
What's the main selling point of Oracle to begin with? From what I've seen so far, most IT people dislike it (pain to install on Linux, huge memory requirements, optimizer that needs many hints, etc).
Your project was really cool, but it's just a very simple in-order pipeline. Doing the same thing on a complex, ~20 stages out-of-order pipeline is very different. For instance, verifying such a design is considerably more difficult than for a clocked design. With verification accounting for about half the design cycles these days, I believe that asynchronicity won't make it in high-perf processors in the near future.
The alternative proposed by the research community is GALS - globally asynchronous, locally synchronous. You get some of the benefits of fully-asynchronous designs (e.g. greatly reduced clock skew), while keeping verification complexity low.
that the iPod had no real competitor. I didn't say "Japanese consumers don't buy foreign stuff" but that, given the choice of comparable products (PS2 and XBOX are clearly in the same category), they will pick the Japanese one.
The real problem with the original XBOX was very simple - Japanese consumers preferred giving their money to Japanese companies (Sony & Nintendo). I think MS is gonna have a really hard time breaking that adoption barrier in Japan.
"It's entirely possible that OOOE could beat out the execution scheme that AMD has going but I wouldn't know enough to comment on it. I remember that there used to be a lot of buzz about IA-64's OOOE [wikipedia.org] processing used on Itanium. But I'm not sure that was too popular among programmers."
Let's clarify a few things: A processor executes instructions either in-order (including VLIW processors), or out-of-order. The former is much simpler to implement, but the latter is much more powerful. Why? because when you miss in the L2 cache and it takes ~200 cycles to bring the thing from memory, it's good if you can do something else in the meanwhile. This is only a performance thing, btw, the programmer doesn't see it. For the same instruction-set architecture you can have both in-order (e.g. Transmeta) or out-of-order implementations (AMD since K6 and Intel since Pentium Pro)
Most high-performance processors these days use OOO (the processors that 99% of the/. crowd currently use to read and post are out-of-order). The notable exceptions are: Sun SPARC (a few things in their instruction set makes OOO very difficult to implement, like instruction windows), the Itanium (designed specifically not to need OOO, but see where it got it), Sun Niagra (the idea there is that a server works better with many simpler, in-order execution contexts than with one bigger out-of-order contexts; don't try to run other stuff though, as you'll probably be unimpressed), and the console chips. I'd add Transmeta, but not sure if they're still alive.
Quality of UMD is actually comparable with DVD. Resolution is 480x272 progressive, so for a normal TV you probably wouldn't be able to tell the difference (if you could play it on a TV, that is). Capacity is "only" 1.8G, on the other hand the encoding is H264 (considerably more powerful than MPEG2). The "low quality" perception comes from the fact that you can only play it on the PSP.
You're right about the other aspects, but I think the main problem is that you can only play it on a PSP (the Universal part is a euphemism)
Slashdot Mirror
Get an American Express for students credit card. They'll give it to you. Use it responsibly (i.e. don't forget to make a payment every month), and watch your credit rating rise ... I've done the same, and got to a point were every week I was getting ~2 to 3 new card applications.
BTW, in the parent post I was referring primarily to the West, not the Eastern Empire.
size-challenged planet?
BTW, I wasn't trying to say that stacked architectures are not worth it, I just didn't like the blank statement from the story: "stacked architectures are better than register-based ones". In architecture one size doesn't fit all (which can be applied the other way around, too, you wouldn't want out-of-order execution for an embedded app)
But why do you need out-of-order execution? Well, misses to memory are very expensive these days - it can easily take from 200 to 400 cycles to service a load that misses all the way to main memory. This can have a significant effect on performance. What out-of-order execution does is to allow independent instructions that are younger than the load to execute in parallel with it. Quite often these parallely-executed instruction will generate other misses to main memory, overlapping their latencies. So - latency of loads that miss is still very high, but at the very least the processor is not idle while servicing them (for a good read see "MLP Yes! ILP no!" by Andy Glew)
Itanium and Sparc compensate for the fact that they don't do stuff out-of-order by putting sh*tloads of L2/3 cache on-chip. The cost of a miss is still very high, but it happens much less often. The manufacturing cost of a chip is also much higher.
Note that what NASA is sending into space is "old" tech. The reason - well, cosmic rays are much stronger in outer space, and the smaller the gate, the easier it is for them to flip its state.
P.S. I'm a computer architect.
True, doesn't work for everything (e.g. 3D acceleration), but otherwise does wonders with respect to compatibility.
If it's older than the earliest Egyptian civilization (around 3000BCE), well ... Sahara wasn't a desert at the time: Sahara.
I also find it interesting that you seem so upset about losing your right to smoke in a restaurant, but, OTOH, you don't seem to consider another person's right to breathe clean air (especially in a restaurant, when one is supposed to enjoy him/herself). When you smoke you're not only affecting negatively your own health (which, as far as I'm concerned, is your absolute right), but also the health of the people around you. Free market doesn't really work here, since most owners will not put "no smoking" signs, considering that it limits their clientelle. So a person who hates smoking, like me, only has two choices: eat out and inhale very annoying and toxic fumes, or eat home.
Also, what do you mean by "excellent on-die cache control"? Cache control is always on-die (it's the cache data that can sometimes be off-chip).
The second - Itanium - as far as I'm concerned, it's simply a step backwards. A processor these days is most of the times limited by the slow memory (it can easily take 200-300 cycles to service a request from memory, as opposed to 2-3 cycles from the L1 cache or 6-20 cycles from the L2). Out-of-order execution (Pentium Pro and after) alleviates the problem to a certain extent, by allowing other instructions that do not depend on the result of the instruction that missed to execute. So the processor can still do something while servicing the miss (quite often it executes other loads that miss, effectively increasing the memory-level parallelism of the processor). Because Itanium executes instructions in order, it simply can't do that. Furthermore the compiler can't tell which instructions are going to miss (it needs a profiler to figure frequently-missing instructions, and only then it can generate prefetches). Intel's solution - let's throw shitloads of caches on the Itanium, to reduce the occurence of the misses. Of course, that makes the chip huge, considerably more expensive, etc.
Nevertheless, Conroe seems to be an awesome chip. Time to buy some INTC shares ...
You buy a license to play the content bound to the medium. When the medium dies ... well tough luck. Even though the current copyright law allows you to make a backup copy, you apparently wave that right when accepting the license...
OTOH the other two moons are small enough to be called moons.
When I said "roughly" I meant ~10->20% is consumed by other components.
In your average laptop, the power consumed by a CPU when running something (i.e. not just idling around) is about half the total power. The other half, roughly, is consumed by the screen.
The purpose of "good dispatching" (i.e. out-of-order execution) is to hide the latencies of misses to main memory (it takes between 200 and 400 cycles these days to get something from memory, assuming that the memory bus isn't saturated), by executing instructions following the miss but not dependent on it. Out-of-order execution has been around Pentium Pro, btw.
... you insensitive clod! :)
West Philadelphia to be more precise. I've seen cars broken for much less valuable stuff ... (like a pack of softdrinks, if you can believe it). Living your purse in a car in a conspicuous place is simply dumb around here.
What's the main selling point of Oracle to begin with? From what I've seen so far, most IT people dislike it (pain to install on Linux, huge memory requirements, optimizer that needs many hints, etc).
The alternative proposed by the research community is GALS - globally asynchronous, locally synchronous. You get some of the benefits of fully-asynchronous designs (e.g. greatly reduced clock skew), while keeping verification complexity low.
that the iPod had no real competitor. I didn't say "Japanese consumers don't buy foreign stuff" but that, given the choice of comparable products (PS2 and XBOX are clearly in the same category), they will pick the Japanese one.
The best argument against democracy is a five-minute conversation with the average voter. Winston Churchill
The real problem with the original XBOX was very simple - Japanese consumers preferred giving their money to Japanese companies (Sony & Nintendo). I think MS is gonna have a really hard time breaking that adoption barrier in Japan.
Let's clarify a few things: A processor executes instructions either in-order (including VLIW processors), or out-of-order. The former is much simpler to implement, but the latter is much more powerful. Why? because when you miss in the L2 cache and it takes ~200 cycles to bring the thing from memory, it's good if you can do something else in the meanwhile. This is only a performance thing, btw, the programmer doesn't see it. For the same instruction-set architecture you can have both in-order (e.g. Transmeta) or out-of-order implementations (AMD since K6 and Intel since Pentium Pro)
Most high-performance processors these days use OOO (the processors that 99% of the /. crowd currently use to read and post are out-of-order). The notable exceptions are: Sun SPARC (a few things in their instruction set makes OOO very difficult to implement, like instruction windows), the Itanium (designed specifically not to need OOO, but see where it got it), Sun Niagra (the idea there is that a server works better with many simpler, in-order execution contexts than with one bigger out-of-order contexts; don't try to run other stuff though, as you'll probably be unimpressed), and the console chips. I'd add Transmeta, but not sure if they're still alive.
compress that :)
You're right about the other aspects, but I think the main problem is that you can only play it on a PSP (the Universal part is a euphemism)