I'd forgotten that these processors are not made on the 3 micron processes like the chips I used to work on!
3 microns? Wow. That's huge! The top of the line chips these days are easily below 0.5 microns. (The PIV chips are 0.18 and 0.13 microns!) I know I was just shocked when I got my Spartan III FPGA kit. I couldn't believe how small the thing was in it's packaging. I can't even imagine how small the actual die must be!
besides, I wear a lead skull cap myself, keep my hair shaved so that I can be in constant contact with the metal of the cap.....and my cell phone reception has never been so good!
With SIMD instructions, you can execute all four of those adds in one instruction. I wish I knew SSE a bit better, then I could rewrite the above. Sadly, I haven't gotten around to learning the precise syntax.:-(
However, there's a fairly good (if not a bit dated) explanation of SIMD here.
You are wrong. These SIMD processors do loops just fine. There's a hefty hit for a mis-predicted branch, but the branch hint instruction works wonders for loops.
Um... I'm not sure that's what he's trying to say. SIMD by definition is Single Instruction, Multiple Data. i.e. You give it a couple of instructions and watch it perform them on every item in the stream of data. By definition, a loop is an iteration over each instruction, multiple times. a.k.a. Multiple Instruction Multiple Data (MIMD).
What's needed to take full advantage of SIMD is instructions that can be bundled together into one long stream. That way the entire stream can be processed without using a loop. In theory, this is the fastest possible way to process data. Especially with high-latency, high-bandwidth memory. Unfortunately, our programming models aren't designed around such concepts, leaving us relying on optimizing compilers and handcoded assembly.
Cell isn't a System-On-A-Chip. It's just a stripped-down, in-order power pc core coupled to 8 single-purpose in-order SIMD units, using an unconventional cache/local memory architecture
You know, I'm looking back at all these replies to the poor guy, and I can't help but think that he's sitting in front of his computer wondering, "Can't anyone explain it in ENGLISH?!?":-P
For instance, you have to unroll your "for" loops to start, since those SIMD co-processors can't do loops.
Actually, we need a new programming model. Instead of using FOR loops, we need a model under while you can say, "Perform these instructions X number of times." One could probably do a bit of guess-work in the compiler based on loops like "for(i=0;i<COUNT;i++)", but that doesn't help cases where the loop uses a more complex conditional statement (or where the test is affected by the loop itself). Thus the language needs to be changed to force the programmer to pre-compute the loop length for maximum performance. For example:
Um. That's kind of a weird statement. I think they mean to say that it encompasses much of the multiprocessing capabilities of a modern PC in a single chip. i.e. It's your CPU and GPU rolled into one.
Cell processors aren't really anything all that new per say. The multi-core design makes them superficially similar to GPUs (which are also vector processors) with the difference that GPUs use multiple pipelines for parallel processing whereas each cell is a self-contained pipeline capable of true multi-threaded execution. In theory, the interplay between these chips could accelerate a lot of the work currently done through a combination of software and hardware. e.g. All the work that graphics drivers do to process OpenGL commands into vector instructions could be done on one or two cells, thus allowing those cells to feed the other cells with data.
I guess you could say that the cell processor is the start of a general purpose vector processing design. I'm not really sure if it will take off, but unbroken thoroughput on these things is just incredible.
Ain't that just the typical MS-FUD running rampant ?
I think it's more of typcial sales-person FUD. Sales people can be extremely slimey critters, and will tell you anything to make sure you buy more stuff. This isn't unique to Microsoft, though it is amusing. (Especially after the whole Korn shell fiasco.)
It wouldn't surprise me at all if the AC's sale rep simply took two unrelated facts (the fact that OpenOffice contains GPL code, and the fact that OpenOffice implements the OpenDocument standard) and intentionally confused them. If he says it enough times, he might even believe it.
Reading the linked PDF, it strikes me that this law makes M and AO rated video games more of a controlled substance rather than outright banning them. That makes the case for a first amendment violation a tricky one.
That being said, I'm really not in favor of the government getting involved in these things. The rise in Ultra-Violent video games are a sign that the market is on the decline. Regulating such activities can have the opposite of the intended effect, as the government is basically okaying such products.
It's the same in the television industry. In the paper this morning they had a story about how 3/4 of TV shows today have strong sexual content in them. (Up from 56% in 2003.) The government is considering intervening. Again, I think the government should stay out. The ratings speak for themselves. The general populace won't buy their shock value crap for very long. Ratings will continue to decline until the shows are either fixed or television as we know it disappears. If the government involves themselves in it, they will only create controversy that will help the TV studios.
Are the mods huffing kittens or something? Parent is not a troll. Overly pessimistic, yes. Troll? No.
When there's a -1: Pessimistic option, then he should be modded down. In the meantime, reread the moderator rules.
As to the parent, I can't say I agree that this will happen. I agree that Microsoft will try (RTF, anyone?), but long term I think that Microsoft just has too many anti-trust watchers breathing down their necks at the moment. Everytime Microsoft attempts to rely on their old tactics (no matter how sneaky they are about it) someone is going to cry foul. It may seem silly, "Them: Microsoft has a tiny incompatibility in their support of the format! Microsoft: It's just a bug! No bigge!" but such attacks can really screw with Microsoft's time to market and keep them tied up in the courts for a very long time.
To hell with Microsoft then! They can use OpenOffice to draft any damn format they want! Cry Havoc, and let loose the dogs of war!
* tongue planted firmly in cheek:-)
Re:C++ is cross-platform, dont know what your smok
on
Write Portable Code
·
· Score: 1
Where reflection comes in handy is when you have a plain old data struct, with data members, and you want to automagically generate a UI to fill in the struct members.
That's one use (though you probably want to use the bean libraries for that). There are a few other good ones:
1. Since Java didn't implement variable arguments until recently, reflection could be used to dynamically call a method with an unknown number of arguments. An example of this is JHDL which looks at a static array to get the number of wires to pass to the constructor, then uses any additional ints or Strings as configuration parameters.
2. Somewhat related to serialization, reflection is perfect for creating Object databases by dynamically saving field values to disk, and reloading the state.
3. Reflection can be used to take a multidimensional array and quicky flatten it for storage on disk, or redimensionalize it after you pull the flat array from disk. This only works because an array of any dimension can be cast to an Object (and vica-versa) allowing you to have only one method for any size array.:-)
Well, you can write code explicitly for that processor, for one thing.
And double the cost of the Video Card. Your competitors will then use your computer's CPU to undercut your cost.
Instead of the many different added instructions that x86 processors have, and the timing differences between processors.
Much easier to simply recompile the drivers for another platform. Or if necessary, port the code. Software maintenece like that doesn't cost that much, and easily gets lost in the cost of normal operating expenses.
But if you have a low intensity of gamma radiation, i.e. a low number of photons, then you have ipso facto a low intensity of energy production. The number of gammas is directly proportional to the number of distintegrations. That sounds like it just won't do if you want a compact source of lots of energy.
I think you have a misunderstanding of radioisotopes. Sr90, for example, is another common isotope used in RTGs. It produces no Gamma radiation to speak of. Of course, it may produce gamma farther down the decay chain (radioisotopes usually go through being a few different elements before reaching a stable state), but in its initial decay it's perfectly safe. I wish I still had the link that showed all the decay states. According to the EPA website, however, Sr-90 decays next into Y-90 which (if I'm reading this page correctly) produces a fraction of 1% of its radiation as Gamma rays.
Er, I believe all energy from radioactive decay, other than what's in the kinetic energy of ejected particles, comes out as gammas. I don't think there are any radioactive decay schema that don't involve gamma radiation.
I didn't say find a type with no gammas. I said, find an isotope that only produces weak gamma. As long as you keep the gamma levels to far below background, you're not going to have a problem. All that RF radiation the phone is spewing would be a far greater health hazzard.
I don't see how shaping the isotope helps your shielding problem. You've got radiation leaving every surface of the isotope.
Simple. Every particle in the object may decay at any moment. If you shape the isotope correctly (I *think* you want a sphere, but I may be misremembering) then you get the maximum amount of material behind other materials. An ejected radioparticle would then have a higher change of hitting something and converting to thermal energy.:-)
I like (3), except that I'm pretty unconfident it can be made small. Problem is, there's a minimal thickness of shielding you need that doesn't shrink with the size of the generator. You need x mm of lead to shield you from gamma rays of a given frequency from a given radioactive decay, and x doesn't get smaller if your device does. So how are you going to get an atomic battery small enough to get into your cell phone?
??? Don't chose a material that produces significant amounts of gamma throughout it's decay life? It's been awhile since I last checked, but I believe there are a couple of options for this. (Pu-238?)
Also, you'll want to shape the isotope for maximum heat anyway, which will cause it to partly shield itself.
I doubt anyone will be writing a JVM for the thing. Nonetheless, this is an example of a processor that is in widespread use.
1) You're twisting my words. I never stated that there was a JVM for every processor. I said there was one for every platform. A PIC is not a platform no matter how many cutesy hacks people make with them. (That PONG game on a PIC was pretty darn cool, though.)
2) If you really want Java on a PIC, you have to be a little more creative. Write the code in Java (preferable targetting something like JavaCard), then translate the instructions to the native set before the load. This is a pretty common practice, and has been done with a lot of embedded devices. If you do it right, you can even fit most of the Java features, just without the normal class format.
Oh, I almost fotgot. IIRC, Java allows you to snag shared memory blocks in byte buffers via memory mapping. That's probably the type of IPC you're thinking of.
Now you're just being weird. There are Java VMs that target devices of a similar size (e.g. JavaCard), but why oh why do you need a JVM for an application specific microprocessor?
Hrumph. Now someone will hear us and go write one. Way to go chief.;-)
What's amazing is that the Mods keep modding him up.
MODS: His post is disinformation! I realize that he probably wasn't aware of the various options, but that doesn't make his post informative or insightful!
Slashdot Mirror
I'd forgotten that these processors are not made on the 3 micron processes like the chips I used to work on!
3 microns? Wow. That's huge! The top of the line chips these days are easily below 0.5 microns. (The PIV chips are 0.18 and 0.13 microns!) I know I was just shocked when I got my Spartan III FPGA kit. I couldn't believe how small the thing was in it's packaging. I can't even imagine how small the actual die must be!
besides, I wear a lead skull cap myself, keep my hair shaved so that I can be in constant contact with the metal of the cap.. ...and my cell phone reception has never been so good!
I don't have a ruler with me at the moment, but that looks pretty close to right for my thumbnail. What can I say? I'm a big guy. :-)
All CPUs are the size of your fingernail. It's the packaging that makes them appear large. :-)
Thanks. That's one of those I keep getting wrong. Keep reminding me and I'll remember at some point. :-)
mov ecx,b
:-(
shr ecx,2
loop:
add eax,[ebx]
add eax,[ebx+4]
add eax,[ebx+8]
add eax,[ebx+12]
add ebx,16
dec ecx
jnz loop
With SIMD instructions, you can execute all four of those adds in one instruction. I wish I knew SSE a bit better, then I could rewrite the above. Sadly, I haven't gotten around to learning the precise syntax.
However, there's a fairly good (if not a bit dated) explanation of SIMD here.
a.k.a. Multiple Instruction Multiple Data (MIMD).
Minor correction. That's supposed to be Single Instruction, Single Data. (SISD) My bad.
Excellent! Now all we need are SIMD optimized LISP Compilers.
:-P
(Must (resist (temptation (to (joke (about (syntax)))))))
You are wrong. These SIMD processors do loops just fine. There's a hefty hit for a mis-predicted branch, but the branch hint instruction works wonders for loops.
Um... I'm not sure that's what he's trying to say. SIMD by definition is Single Instruction, Multiple Data. i.e. You give it a couple of instructions and watch it perform them on every item in the stream of data. By definition, a loop is an iteration over each instruction, multiple times. a.k.a. Multiple Instruction Multiple Data (MIMD).
What's needed to take full advantage of SIMD is instructions that can be bundled together into one long stream. That way the entire stream can be processed without using a loop. In theory, this is the fastest possible way to process data. Especially with high-latency, high-bandwidth memory. Unfortunately, our programming models aren't designed around such concepts, leaving us relying on optimizing compilers and handcoded assembly.
You know, I'm looking back at all these replies to the poor guy, and I can't help but think that he's sitting in front of his computer wondering, "Can't anyone explain it in ENGLISH?!?"
For instance, you have to unroll your "for" loops to start, since those SIMD co-processors can't do loops.
Actually, we need a new programming model. Instead of using FOR loops, we need a model under while you can say, "Perform these instructions X number of times." One could probably do a bit of guess-work in the compiler based on loops like "for(i=0;i<COUNT;i++)", but that doesn't help cases where the loop uses a more complex conditional statement (or where the test is affected by the loop itself). Thus the language needs to be changed to force the programmer to pre-compute the loop length for maximum performance. For example:
Um. That's kind of a weird statement. I think they mean to say that it encompasses much of the multiprocessing capabilities of a modern PC in a single chip. i.e. It's your CPU and GPU rolled into one.
Cell processors aren't really anything all that new per say. The multi-core design makes them superficially similar to GPUs (which are also vector processors) with the difference that GPUs use multiple pipelines for parallel processing whereas each cell is a self-contained pipeline capable of true multi-threaded execution. In theory, the interplay between these chips could accelerate a lot of the work currently done through a combination of software and hardware. e.g. All the work that graphics drivers do to process OpenGL commands into vector instructions could be done on one or two cells, thus allowing those cells to feed the other cells with data.
I guess you could say that the cell processor is the start of a general purpose vector processing design. I'm not really sure if it will take off, but unbroken thoroughput on these things is just incredible.
Ain't that just the typical MS-FUD running rampant ?
I think it's more of typcial sales-person FUD. Sales people can be extremely slimey critters, and will tell you anything to make sure you buy more stuff. This isn't unique to Microsoft, though it is amusing. (Especially after the whole Korn shell fiasco.)
It wouldn't surprise me at all if the AC's sale rep simply took two unrelated facts (the fact that OpenOffice contains GPL code, and the fact that OpenOffice implements the OpenDocument standard) and intentionally confused them. If he says it enough times, he might even believe it.
Reading the linked PDF, it strikes me that this law makes M and AO rated video games more of a controlled substance rather than outright banning them. That makes the case for a first amendment violation a tricky one.
That being said, I'm really not in favor of the government getting involved in these things. The rise in Ultra-Violent video games are a sign that the market is on the decline. Regulating such activities can have the opposite of the intended effect, as the government is basically okaying such products.
It's the same in the television industry. In the paper this morning they had a story about how 3/4 of TV shows today have strong sexual content in them. (Up from 56% in 2003.) The government is considering intervening. Again, I think the government should stay out. The ratings speak for themselves. The general populace won't buy their shock value crap for very long. Ratings will continue to decline until the shows are either fixed or television as we know it disappears. If the government involves themselves in it, they will only create controversy that will help the TV studios.
Are the mods huffing kittens or something? Parent is not a troll. Overly pessimistic, yes. Troll? No.
When there's a -1: Pessimistic option, then he should be modded down. In the meantime, reread the moderator rules.
As to the parent, I can't say I agree that this will happen. I agree that Microsoft will try (RTF, anyone?), but long term I think that Microsoft just has too many anti-trust watchers breathing down their necks at the moment. Everytime Microsoft attempts to rely on their old tactics (no matter how sneaky they are about it) someone is going to cry foul. It may seem silly, "Them: Microsoft has a tiny incompatibility in their support of the format! Microsoft: It's just a bug! No bigge!" but such attacks can really screw with Microsoft's time to market and keep them tied up in the courts for a very long time.
That's right. Viva la ASCII! ;-)
To hell with Microsoft then! They can use OpenOffice to draft any damn format they want! Cry Havoc, and let loose the dogs of war!
:-)
* tongue planted firmly in cheek
Where reflection comes in handy is when you have a plain old data struct, with data members, and you want to automagically generate a UI to fill in the struct members.
:-)
That's one use (though you probably want to use the bean libraries for that). There are a few other good ones:
1. Since Java didn't implement variable arguments until recently, reflection could be used to dynamically call a method with an unknown number of arguments. An example of this is JHDL which looks at a static array to get the number of wires to pass to the constructor, then uses any additional ints or Strings as configuration parameters.
2. Somewhat related to serialization, reflection is perfect for creating Object databases by dynamically saving field values to disk, and reloading the state.
3. Reflection can be used to take a multidimensional array and quicky flatten it for storage on disk, or redimensionalize it after you pull the flat array from disk. This only works because an array of any dimension can be cast to an Object (and vica-versa) allowing you to have only one method for any size array.
Well, you can write code explicitly for that processor, for one thing.
And double the cost of the Video Card. Your competitors will then use your computer's CPU to undercut your cost.
Instead of the many different added instructions that x86 processors have, and the timing differences between processors.
Much easier to simply recompile the drivers for another platform. Or if necessary, port the code. Software maintenece like that doesn't cost that much, and easily gets lost in the cost of normal operating expenses.
But if you have a low intensity of gamma radiation, i.e. a low number of photons, then you have ipso facto a low intensity of energy production. The number of gammas is directly proportional to the number of distintegrations. That sounds like it just won't do if you want a compact source of lots of energy.
I think you have a misunderstanding of radioisotopes. Sr90, for example, is another common isotope used in RTGs. It produces no Gamma radiation to speak of. Of course, it may produce gamma farther down the decay chain (radioisotopes usually go through being a few different elements before reaching a stable state), but in its initial decay it's perfectly safe. I wish I still had the link that showed all the decay states. According to the EPA website, however, Sr-90 decays next into Y-90 which (if I'm reading this page correctly) produces a fraction of 1% of its radiation as Gamma rays.
Er, I believe all energy from radioactive decay, other than what's in the kinetic energy of ejected particles, comes out as gammas. I don't think there are any radioactive decay schema that don't involve gamma radiation.
:-)
I didn't say find a type with no gammas. I said, find an isotope that only produces weak gamma. As long as you keep the gamma levels to far below background, you're not going to have a problem. All that RF radiation the phone is spewing would be a far greater health hazzard.
I don't see how shaping the isotope helps your shielding problem. You've got radiation leaving every surface of the isotope.
Simple. Every particle in the object may decay at any moment. If you shape the isotope correctly (I *think* you want a sphere, but I may be misremembering) then you get the maximum amount of material behind other materials. An ejected radioparticle would then have a higher change of hitting something and converting to thermal energy.
I like (3), except that I'm pretty unconfident it can be made small. Problem is, there's a minimal thickness of shielding you need that doesn't shrink with the size of the generator. You need x mm of lead to shield you from gamma rays of a given frequency from a given radioactive decay, and x doesn't get smaller if your device does. So how are you going to get an atomic battery small enough to get into your cell phone?
??? Don't chose a material that produces significant amounts of gamma throughout it's decay life? It's been awhile since I last checked, but I believe there are a couple of options for this. (Pu-238?)
Also, you'll want to shape the isotope for maximum heat anyway, which will cause it to partly shield itself.
I doubt anyone will be writing a JVM for the thing. Nonetheless, this is an example of a processor that is in widespread use.
1) You're twisting my words. I never stated that there was a JVM for every processor. I said there was one for every platform. A PIC is not a platform no matter how many cutesy hacks people make with them. (That PONG game on a PIC was pretty darn cool, though.)
2) If you really want Java on a PIC, you have to be a little more creative. Write the code in Java (preferable targetting something like JavaCard), then translate the instructions to the native set before the load. This is a pretty common practice, and has been done with a lot of embedded devices. If you do it right, you can even fit most of the Java features, just without the normal class format.
But none of those are inter process communications.
:-)
They're not? Guess I should have my Geek Card revoked then.
IPC is a pretty broad term.
Oh, I almost fotgot. IIRC, Java allows you to snag shared memory blocks in byte buffers via memory mapping. That's probably the type of IPC you're thinking of.
Now you're just being weird. There are Java VMs that target devices of a similar size (e.g. JavaCard), but why oh why do you need a JVM for an application specific microprocessor?
;-)
Hrumph. Now someone will hear us and go write one. Way to go chief.
What's amazing is that the Mods keep modding him up.
MODS: His post is disinformation! I realize that he probably wasn't aware of the various options, but that doesn't make his post informative or insightful!