It just seems to me that multi-core processors are a waste, when that space could be better used by a single bigger processor core.
Anyone that's ever designed a circuit realizes that redundancy is redundant. 4 processors !/= 4x the power. Fractional improvements are fractional of what normal should be.
Because google can run on a single processor.
Yeah, parallel programming is useless. NSA dude. evil.
They backdoor encryption algorithms and limit your processor power.
Uh, yeah. You do realize that "faster" is not the same as powerful, right? I'm not talking about the fastest search engine known to man.
The same version of Maya that came out 5 years ago only runs twice as fast on the best machine that money can buy as it does on my piece of shit with a gig of memory. Do you comprehend what I'm saying? My computer is 5 years old. What, are you impaired or something? The technology FUCKING SUCKS! There's been no improvements. Did you see Beowulf? that shit had the same production time and quality as "Final Fantasy: The spirits within".
How can you say that's good, fagabilly?
Yay! Computers have stopped having significant improvements!
Your head, your rectum. Remove one from the other.
>>6
Shut your whore mouth, faggot.
I'm a fucking computer scientist, what the FUCK do you know?
Google uses ingenius algorithms, and powerful hardware.
The concept of time exists in algorithms (and partially in hardware; depending on the vendor there might be an explanation of a CPU cycle, bitrate, et al in the technical manual)
Now, what the FUCK do you know about von Neumann archs and parallelism? (MIMD, sounds familiar?)
The difference between what you're saying and the truth is that you're making excuses for an industry that that has been prevented from making considerable improvements by it's government because they can't beat anything above 90-bit encryption. All your "computer science" doesn't mean shit compared to what a free market is supposed to be. Obey more. The difference between what you're saying and reality is in application. Remember my comparison where the same version of Maya is run on my 5 year old piece of shit, and the exact same software and scene is made on the best, most modern, personal computer that money can buy, and the improvement is pitiful?
THAT'S ALL I NEED TO KNOW, YOU MASSIVE FLOATING PILE OF SUCK-UP.
Name:
Anonymous2008-03-05 19:08
Multiprocessors are overrated for very nearly all consumer purposes. They're just cheaper to make, so marketing.
Name:
Anonymous2008-03-05 20:41
>>9
Not to mention those multi processors have to all use the same memory. Yeah, multi-processors are massive amounts of FAIL.
Name:
4tran2008-03-06 0:33
Some tasks are inherently amenable to parallel processing. The GPU is a lot slower than the CPU, but because of how parallel it is, it is a lot better at rendering graphics.
The difference between what you're saying and the truth is that you're making excuses for an industry that that has been prevented from making considerable improvements by it's government because they can't beat anything above 90-bit encryption.
What the fuck are you talking about?
Remember my comparison where the same version of Maya is run on my 5 year old piece of shit, and the exact same software and scene is made on the best, most modern, personal computer that money can buy, and the improvement is pitiful?
Sigh..
Now I'll take a moment here and question myself
"How do you explain computers & programming to someone who has NO FUCKING IDEA abou the subject?"
You can't. I'll attempt, knowing from before my attempt will fall on deaf ears.
For starters, let's assume the program is a spell, and that the computer is a wizard "casting" the "spell".
Does the spell get more powerful if the wizard is more experienced?
No, ofcourse, the better wizard can cast it easier, but its strength is the same.
There are also some wirardz who don't have enough mana to cast the spell.
Holy fuck, I did a good job without getting into technical minutiae. Maybe you'll actually get it.
>>10
Do they?
That's all up to the kernel and the implementation. You have no fucking clue.
>>12 Hey, can I toss something in real quick?
Well, can you? I'm sure your great at your job, and that you worked hard to get your degree.
False assumptions, I don't work, I research, moreover it was really easy to get my PhD.
Also, you're*. But you sound like a huge dick,
I AM. so I can see why people don't want to listen to you.
Frankly I don't care.
If they were any wiser they'd take a moment to consider my words.
Name:
Anonymous2008-03-06 15:28
| | I'm sure your great at your job, and that you worked hard to get your degree.
| False assumptions, I don't work, I research, moreover it was really easy to get my PhD.
| Also, you're*.
>>6
So, just to clarify:
(a) The version of Maya which was out 5 years ago doesn't support multithreading
or
(b) You're a fucking idiot and you're talking out your ass.
Thanks for playing.
PS: 90 bit, christ, I lol'd so hard.
Name:
Anonymous2008-03-07 20:06
Just a week ago I have installed POVRay and wanted to check how much faster it will run on all 4 cores compared to single-threaded on my new CPU. It turned out POVRay still doesn't support multithreading. Bummer.
>>23
There is no facepalm large enough to encompass your fail.
Name:
Anonymous2008-03-08 13:57
I have a Dell computer.
Name:
Anonymous2008-03-10 2:52
>>23
except quantum computers are useless to the home user. We dont need to do quantum physics calculations. FAIL Nanotubes are irrelevent to this conversation as of now, why would you use a tube to conduct electricity? hollow = smaller volume of metal = gets hotter = BAD
Name:
Anonymous2008-03-10 3:44
>>26
>quantum computers
>We dont need to do quantum physics calculations.
I live on a farm. The world is pretty much static to me.
Name:
Dariael!!WoiXApvOze+y6Vv2008-03-18 2:23
Oh goody, this ought to be fun....
Name:
Anonymous2008-03-18 3:15
why dont' they just make the circut chip 4x as big
Name:
Anonymous2008-03-18 4:05
We need to increase the data width, not other improvements. The first company that develops a 128bit cpu with a 256bit bus owns the world for the next 10 years.
Name:
Anonymous2008-03-18 5:19
>>24
stfu michio kakus says nano computer
and since you are not michio kaku you can suck my dick
Name:
Anonymous2008-03-18 6:57
>>33
Nano computer? Is that anything like that molecule shaped like a car(sort of) that runs on radio waves? Just to let you know, nobody has any plans to make any nano-computers anytime within the next 2 decades.
>>32
Your 128 bit CPU can calculate floating point numbers with higher precision than my 32 bit CPU, but you still cannot calculate 1+1 any faster than my 32 bit CPU. What matters now is speed, not precision. Unless you can somehow cram 2 64 bit instructions into a single 128 bit instruction and execute it in one go, your 128 bit CPU is not noticeably faster. Even if you do that, you'd have to drastically change computer architecture.
>>33
Who the hell is Michio Kaku? I think I've heard the name before.
>>36
Yeah. because computers -only- need to handle 1+1. Shit like the massive numbers involved 3d rendering will never be used by anyone. Fuckface.
Name:
Anonymous2008-03-21 5:35
Bakka bakka bakka! The reason we need architecture improvements is because even at the speed of light, a 32 bit processor with a 64 bit bus is -not- enough.
>>38
If I hear you insulting 4tran again, I'm going to gouge out your spleen with the base of an anvil, and then strip your body of enough tendons to make a few Genghis-era Asiatic shortbows.
Name:
4tran2008-03-21 23:16
>>38
With 32 bit float, you can get up to ~ +- 3*1038
With 64 bit float, you can get up to ~ +- 2*10308
Even if you convince me that 32 bit floats are insufficient, you can't convince me that 64 bit is insufficient; this latter range of numbers far exceeds the estimated # of particles in the universe.
Besides, 1038 + 1038 takes the same amount of time as 1 + 1: 1 clock cycle.
Good architectural design like efficient pipelining can dramatically speed up computation, but having extra bits won't.
The reason Crysis is slow on your shitty computer is because it's not fast enough, not because it's only 32/64 bit. If it's a lack of precision, then numerical errors will corrupt the game's physics in no time.
Most of the videogames I require to be made depend on calculating subatomic particles to create entire universes, ecosystems, and biological organisms that make up societies of people who stare at you and say "Outlander..."
>>41
meh, 2 bits are fine too, right? If we speed up the 2 bit processor, we should get it to run as good as a 64 bit, just by making it faster. We could make some awesome 2 bit computers these days. Yeah, that's the ticket.
Name:
Anonymous2008-03-22 20:26
if you can calculate 80,000 raytraces in the time it takes to calculate 1+1, that is no improvement. Come on, what's that good for?
Name:
Anonymous2008-03-23 0:46
>>44
So just to clarify, you genuinely think that lack of precision, rather than speed, is a serious issue in most computing? If only we had 128bit floating point operations done in one clock cycle, we'd have realtime raytracing or something? Of course a 2 bit processor would be awful: because 2 bits is clearly not sufficiently precise for any real number calculations. 64 bits is plenty for all non-scientific purposes. Hell, 32 bits is most likely enough for all non-scientific purposes. You have a 10-year-old's understanding of this matter if you truly believe that 128-bit or 256-bit or 1024-bit computers are going to be radically faster than 32- or 64-bit computers are.
>>45
And how is this supposed 128bit CPU going to compute 80,000 "raytraces" (I assume you mean ray-triangle intersections; the definition of a "raytrace" would clearly depend on what you are ray tracing) in the same time it takes to calculate 1+1? At best a 128-bit CPU could perform operations on 4-vectors of 32 bit floating point numbers (16 bits is generally considered insufficient for graphics purposes nowadays); this would hypothetically offer a 4x speed increase. However, a 4 core 32 bit CPU would offer that same 4x speed increase and be more flexible at the same time.
Name:
4tran2008-03-23 4:40
>>42
If that's the case (calculating subatomic particles to create entire universes), then you're already screwed. Lrn 2 chaos theory. For very large systems, small errors tend to grow exponentially with time (primary reason N body problem is intractable). In other words, even 10000 bits won't last you more than a few yrs, assuming of course, that you can even simulate a yr before the end of the real world.
Basically, you just said why the other guy was right, and you are wrong.
Name:
Anonymous2008-03-23 16:27
Soon we will have quantum CPUs and solid state drives made of crystal that can store terabytes in a single beam of light. I read WiReD so it has to be true.
>>48 >>49
First off, I like how you always make two posts in rapid succession in an attempt to make it look like you're two people. You actually had me going for a minute there. Anyways, your entire understanding of CPU architecture is apparently based on commercials for the Nintendo 64. I can't come up with any other reason for how you could possibly be this stupid. Further, try reading my fucking post; I specifically said a 2 bit processor would NOT be sufficient due to the obvious lack of precision. 32 bit and 64 bit processors do not suffer from a lack of precision in virtually any non-scientific applications. 128+ bit processors bring virtually nothing to the table except the possibility of use as vector processors, which tends to be usable only in niche applications, unlike other improvements (more cores). Indeed, most of the driving cause behind the movement to 64 bit was memory addressing, not precision of floating point numbers.
Name:
Anonymous2008-03-24 5:37
>>52
I make 2 posts because I can't delete the first one, shitface.
Name:
Anonymous2008-03-24 5:40
Besides, it should be obvious from the language style. I've seen your posts all over the place, and I never assumed you were 2 people.
Name:
Anonymous2008-03-24 5:45
and finally, if you can't figure out how to create a bit of machine code that allows a 2 bit processor to act like a 64 bit, irregardless of how long it takes for the 2 bit processor to go through the instructions, that's on you.
Name:
Anonymous2008-03-24 5:53
as far as that goes, machine code is like clockwork, and programs simply count fractions. The trick is to get the clockwork to perform usable or fun functions while it progresses. You're boring.
Name:
Anonymous2008-03-24 12:31
>>53 >>54 >>55 >>56
Christ, try and fucking think about your post before making it. Also, you see my posts "all over the place"? I've made a grand total of maybe 20 posts on /sci/ and none on the other textboards, and none of them were double (or quadruple) posts.
Yes, of course you can emulate a 32 bit processor - or 64 bit processor - with a 2 bit processor; however it comes at a massive loss of performance. However, this DOES NOT WORK IN REVERSE. Just because a 32 bit processor is considerably slower doing 64 bit computations does not mean that a 128 bit processor will be considerably faster doing 64 bit computations. At BEST you can double performance by having it function as a vector processor, but this is very limiting compared to having two 64 bit cores.
Also, with regards to the last of your four post monstrosity - jesus christ I lol'd hard. Did you learn about computer science from a Neal Stephenson novel?
Finally, here's a simple question: How is a 128 bit CPU going to significantly improve performance (not precision; PERFORMANCE) over a 64 bit CPU, other than in the very limited fashion which I have already described? Until you can answer this clearly and logically (ie, not your usual "DUH! MORE BITS HUR HUR"), I will not reply again.