flat assembler
Message board for the users of flat assembler.

Index > Heap > Intel Readies Massive Multicore Processors

Goto page Previous  1, 2, 3, 4, 5  Next
Author
Thread Post new topic Reply to topic
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
DustWolf wrote:
kohlrak wrote:
Which means our infinately powerful machines need a new way of making parts.


No doubt. They have researched that too.

There is no particular way to make transistors smaller than the atoms they are made of, you need a new type of transistor (or other data processing device to do so). I believe the effort was called quantum computing.


I've heard alot of promises of fiber optics, but i don't know much about the limitations. I know a bit about transistors and i doubt fiberoptics can redo transistors.
Post 19 Jun 2007, 20:30
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
DustWolf



Joined: 26 Jan 2006
Posts: 373
Location: Ljubljana, Slovenia
DustWolf
kohlrak wrote:
There is no particular way to make transistors smaller than the atoms they are made of, you need a new type of transistor (or other data processing device to do so). I believe the effort was called quantum computing.


I've heard alot of promises of fiber optics, but i don't know much about the limitations. I know a bit about transistors and i doubt fiberoptics can redo transistors.[/quote]

Yeah that's a good point, if one managed to put togather a computer purely optical they could go up to whatever frequency they liked, as light has none of the nasty properties of electricity e.g. conduit heating, circuit capacitance or inductance, etc. I guess the main problem is doing optical transistors.

The only succesfull things of the sort I have heard of so far were chemical-based but those are of course much slower than electronic transistors and hence not much of a solution there.

This is the basis for an interesting discussion BTW. One I'd love to participate in.
Post 19 Jun 2007, 20:43
View user's profile Send private message AIM Address Yahoo Messenger MSN Messenger Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
DustWolf wrote:
kohlrak wrote:
Quote:
There is no particular way to make transistors smaller than the atoms they are made of, you need a new type of transistor (or other data processing device to do so). I believe the effort was called quantum computing.


I've heard alot of promises of fiber optics, but i don't know much about the limitations. I know a bit about transistors and i doubt fiberoptics can redo transistors.


Yeah that's a good point, if one managed to put togather a computer purely optical they could go up to whatever frequency they liked, as light has none of the nasty properties of electricity e.g. conduit heating, circuit capacitance or inductance, etc. I guess the main problem is doing optical transistors.

The only succesfull things of the sort I have heard of so far were chemical-based but those are of course much slower than electronic transistors and hence not much of a solution there.

This is the basis for an interesting discussion BTW. One I'd love to participate in.


Diods would also be a problem... Irregardlessly, the fiber optic processors we've been promised have rumored to already be in place on some cellphones. Which dosn't surprise me. Some cellphones have some pretty amazing capability for little gadgets without a coolent system even as simple as a fan. I heard alot about fiber optics processors but i can't seperate the rumors from the truth. Fact is, too many companies would loose money to fiber optics. if they do have such ability, they woud loose money because of the major heat reduction and then ability to clock pretty damn high. Even if you can't make the transistors. There is alot of heat generated by other components as well. But i still say it's all in the ram. We access the ram alot, and with low ram speed we're screwed. Our super processor gets put into a wait state waiting for all the junk.
Post 19 Jun 2007, 21:58
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
DustWolf wrote:

Just pointing out that AMD has been using that approach in their multi-CPU server systems all this time so far. AMD uses NUMA, AMD uses fast interconnects, Intel does neither. Intel has been staying well away from anything the competition uses so far by principle, what makes you think they will switch sides now?

NUMA might be a solution, but requires at least OS support, and preferably application support, to be utilized decently. IMHO faster memory buses are preferable, but this does become sorta hard when steadily increasing the number of cores...

How much has those HT links helped AMD anyway? Smile

DustWolf wrote:

I have found dependance on old sources of information to be a bad idea when you're trying to get the best for your money. Update your data sources.

Old source of information?

The latest benchmarks I've seen (the K10 cinebench) weren't very favorable for AMD. Of course things might change, but it seemed pretty crappy considering K10 is supposed to be 'new stuff' rather than just a GHz bump.

DustWolf wrote:

Biased view. AMD did make their own CPU. Just because it's compatible doesn't mean it's all the same. Copying intel's tech is for example Cyrix.

Afaik AMD started with pretty much 1:1 copy of intels processors in the 486 days, the first really competitive processor they did was the Athlon, which copied bus design from the Alpha architecture... 3DNow! wasn't much of a success, so instead SSE was copied (licensed of course). The first thing AMD really designed grounds-up was the 64bit stuff, and imho that was folly, and removing any hope of ever leaving x86 (as if all the legacy wasn't enough of a burden).

DustWolf wrote:

You wouldn't happen to have realized that if AMD designed the 64bit standards, that they might have designed them in such a way that they are more easily implementable on AMD's architecture?

And what do you mean by that?

A few posts ago, you it seemed that you implied intel didn't have x86-64 support yet, while they've had for several years...

PS: I run an amd64x2, so I'm not a biased fanboy, I pick the product that's best at the time of purchase. But currently, things don't look too well for AMD.
Post 19 Jun 2007, 23:40
View user's profile Send private message Visit poster's website Reply with quote
DustWolf



Joined: 26 Jan 2006
Posts: 373
Location: Ljubljana, Slovenia
DustWolf
f0dder wrote:
NUMA might be a solution, but requires at least OS support, and preferably application support, to be utilized decently. IMHO faster memory buses are preferable, but this does become sorta hard when steadily increasing the number of cores...

How much has those HT links helped AMD anyway? Smile


A little bit beside the point. We were talking about Intel's implementation not AMDs, but since you bring it up... NUMA works very well with Opteron server systems as I recall... at the very least there is the issue of efficiently managing 32 GB of RAM. As for the HT links... well not sure what you expect, benchmarks have been done for the practical side and the theoretical side is pretty straightforward too: A CPU can use such a thing as a fast bus, either to work with other CPUs or to work with any number of the other things that make a computer a computer.

Quote:
Old source of information?

The latest benchmarks I've seen (the K10 cinebench) weren't very favorable for AMD. Of course things might change, but it seemed pretty crappy considering K10 is supposed to be 'new stuff' rather than just a GHz bump.


There are new Athlon systems that bench faster than Cores. The prices are half as much anyway. Might just be the Slovenian market but that's the information I got.

Quote:
DustWolf wrote:

Biased view. AMD did make their own CPU. Just because it's compatible doesn't mean it's all the same. Copying intel's tech is for example Cyrix.

Afaik AMD started with pretty much 1:1 copy of intels processors in the 486 days, the first really competitive processor they did was the Athlon, which copied bus design from the Alpha architecture... 3DNow! wasn't much of a success, so instead SSE was copied (licensed of course). The first thing AMD really designed grounds-up was the 64bit stuff, and imho that was folly, and removing any hope of ever leaving x86 (as if all the legacy wasn't enough of a burden).


As I said: Irrelevant. Compatibility doesn't mean it's all the same. Sound blasters are all 100% compatible, since aincient times till now, doesn't mean every manufacturer uses the exact same implementation that they stole from somebody else.

Cyrix processors were made by reverse engineering Intel CPUs. Intel prooved that point in court, would you guess why Intel never went to proove that particular point to AMD?

3DNow!, as I recall was used just as much as SSE. The truth is most software simply doesn't use any of those extensions. Programs use MMX, but not SSE, SSE2 or SSE3 unless whoever wrote it specifically decides so. Most software that generically supports SSE, SSE2 and SSE3 also supports 3DNow! and 3DNow!+ (MP3 encoders, open source software, etc).

And if I understand AMD's statements correctly, part of the 3DNow! functionality is wired into the FPU, meaning the FPU preforms better even without a program specifically utilizing the 3DNow! instruction set. Hence making 3DNow! better utilized than any SSE implementation.

Quote:
DustWolf wrote:

You wouldn't happen to have realized that if AMD designed the 64bit standards, that they might have designed them in such a way that they are more easily implementable on AMD's architecture?

And what do you mean by that?

A few posts ago, you it seemed that you implied intel didn't have x86-64 support yet, while they've had for several years...

PS: I run an amd64x2, so I'm not a biased fanboy, I pick the product that's best at the time of purchase. But currently, things don't look too well for AMD.


I have never said that Intel didn't have x86-64 support, I just said exactly what you later re-stated: Intel has acquired the standards specification for 64bit platform from AMD. Which means AMD has set those standards. Which means whatever standards AMD set, it is very likely they set them in a way that is easiest to implement on whatever wiring AMD processors use, which is probably different from the way Intels are put togather, in effect making it harder for Intel to implement 64bit support on their own hardware, compared to AMD. Hence in order for Intel to come up with a briliant plan, it is unlikely the 64bit technology would be fully compatible with whatever Intel had in mind, hence Intel would have to alter the 64bit standard to make it efficient on their CPUs.
Post 20 Jun 2007, 00:08
View user's profile Send private message AIM Address Yahoo Messenger MSN Messenger Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
DustWolf wrote:

at the very least there is the issue of efficiently managing 32 GB of RAM.

Managing the RAM isn't that hard - having a lot of cores all have fast access to RAM is.

DustWolf wrote:

As for the HT links... well not sure what you expect, benchmarks have been done for the practical side and the theoretical side is pretty straightforward too: A CPU can use such a thing as a fast bus, either to work with other CPUs or to work with any number of the other things that make a computer a computer.

What I meant is that I see a lot of AMD fanboys touting HT as being some fantastic and big advantage to whatever intel is doing, but it doesn't really seem to be much of a deal when looking at benchmarks.

Of course I haven't really looked into multi-socket systems and benchmarks, so the situation might be different there.

DustWolf wrote:

There are new Athlon systems that bench faster than Cores. The prices are half as much anyway. Might just be the Slovenian market but that's the information I got.

Price is an issue, sure, and I'll have to admit I haven't looked into how well the cheap-end core2duo does against the cheap-end amd64x2 (and I haven't bothered looking at Pentium4/D for quite a while). For what it's worth, in .dk a c2d E6600 is now priced roughly equal to and AMD64x2 5400+. The low-end c2d E4300 is priced roughly equal to an AMD64x2 4400+.

DustWolf wrote:

As I said: Irrelevant. Compatibility doesn't mean it's all the same. Sound blasters are all 100% compatible, since aincient times till now, doesn't mean every manufacturer uses the exact same implementation that they stole from somebody else.

Actually all soundblasters (and clones) aren't 100% compatible. Try getting a sblive to work with DOS games... without the emu software. Good luck.

DustWolf wrote:

Cyrix processors were made by reverse engineering Intel CPUs. Intel prooved that point in court, would you guess why Intel never went to proove that particular point to AMD?

Perhaps they did the smart thing and licensed the tech? Or did a settlement? I'll let historians solve that Smile

DustWolf wrote:

3DNow!, as I recall was used just as much as SSE. The truth is most software simply doesn't use any of those extensions. Programs use MMX, but not SSE, SSE2 or SSE3 unless whoever wrote it specifically decides so. Most software that generically supports SSE, SSE2 and SSE3 also supports 3DNow! and 3DNow!+ (MP3 encoders, open source software, etc).

3DNow! never got very widespread use, SSE sorta has. And no, it doesn't make sense to write SSE code for a word processor.

DustWolf wrote:

And if I understand AMD's statements correctly, part of the 3DNow! functionality is wired into the FPU, meaning the FPU preforms better even without a program specifically utilizing the 3DNow! instruction set. Hence making 3DNow! better utilized than any SSE implementation.

3DNow! registers were mapped on the MMX registers which were mapped on the x86 registers/stack... yeah, great design *cough*. See technote 21928 from AMD, unless there's a more recent version.

DustWolf wrote:

I have never said that Intel didn't have x86-64 support, I just said exactly what you later re-stated:

Lots of bullshit snipped... that was a complete pile of FUD crap.

Let the benchmarks speak. Not that 64bit benchmarks interest me much before there's a decent amount of software that actually utilizes 64bit and has noticable gains from it.

I do hope that AMD manage to pull their shit together and re-enter the competition, but I still don't think things look too good for them.
Post 20 Jun 2007, 00:47
View user's profile Send private message Visit poster's website Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8903
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
while reading this, somehow one question pops up in my head

is that possible when light beam strucks the mirror, it would move the mirror a bit?
Post 20 Jun 2007, 00:53
View user's profile Send private message Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
Who knows? Light has been known for being mysterious. Acts as a wave and a particle.
Post 20 Jun 2007, 01:31
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
DustWolf



Joined: 26 Jan 2006
Posts: 373
Location: Ljubljana, Slovenia
DustWolf
sleepsleep wrote:
while reading this, somehow one question pops up in my head

is that possible when light beam strucks the mirror, it would move the mirror a bit?


It would. In outer space.

An interesting idea, tho I'm not sure if it's possible in an earthly environment.
Post 20 Jun 2007, 10:30
View user's profile Send private message AIM Address Yahoo Messenger MSN Messenger Reply with quote
vid
Verbosity in development


Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia
vid
kohlrak: Not just light... everything is both wave and particle due to latest physical theories. So due to your thinking, that means everything is "mysterious" Wink
Post 20 Jun 2007, 13:55
View user's profile Send private message Visit poster's website AIM Address MSN Messenger ICQ Number Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
I don't keep myself up to date with the latest theories. I typically don't like theories that have been comming out in the past 100 years. They all seem to be propoganda. I've seeen too many theories based on other theories. They call the most recent theories scientifically based when it's not so scientific when they're based entirely on unproven theories, right? I personally suggest not taking science seriously until it's out of politics. Makes much more sence that way. Very Happy
Post 20 Jun 2007, 18:22
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
vid
Verbosity in development


Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia
vid
kohlrak: "unproved theory"? please... this demonstrates you don't even know what "theory" is.

i suggest you to first study something about science, and then comment it.

Excellent place for you to start: http://www.infidels.org/library/modern/richard_carrier/SciLit.html
Post 20 Jun 2007, 20:53
View user's profile Send private message Visit poster's website AIM Address MSN Messenger ICQ Number Reply with quote
tom tobias



Joined: 09 Sep 2003
Posts: 1320
Location: usa
tom tobias
sleepsleep wrote:
is that possible when light beam strucks the mirror, it would move the mirror a bit?

In my opinion, this is a profound question, since generally, light (i.e. photons,) with a very small mass, is regarded in most ordinary calculations as massless, though possessing momentum!!! Here is an excellent summary, including a few notes on Einstein's brilliant discovery in 1915 explaining how gravity affects light, NOT because of the photons' mass, but because gravitational fields "change the shape of space-time."
http://imagine.gsfc.nasa.gov/docs/ask_astro/answers/961102.html
So, the question then is whether or not a "massless" force, i.e. the light beam, can interact with an impenetrable solid, i.e. the mirror, itself traveling very fast, to change its velocity? My supposition is that the subtle change in velocity would be undetectable by our measurement capability. I would suppose that experiment has been carried out at CERN, using, instead of a mirror, a simple particle, like a neutron, accelerated to a known, measureable speed, moving circularly about the magnet. Such data ought to be available on the internet, I suppose...
Thanks sleepsleep, good question.... Smile
Post 20 Jun 2007, 22:18
View user's profile Send private message Reply with quote
sleepsleep



Joined: 05 Oct 2006
Posts: 8903
Location: ˛                             ⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣⁣Posts: 334455
sleepsleep
thanks & welcome, tom,
a friend of mine said "they slow down a satellite by turning the solar panels perpendicular to the sun etc" and light beam indeed could move mirror.

btw, and is that possible to trap light? eg like using a pyramid that internally full of mirrors that make light reflect inside (after we put a short beam into it) then close it, take it to dark mirror room and open the pyramid

if we could trap a short beam light inside then take it to cube mirror dark room then open it, probably we could know how and why the light vanishes.

weird, if we r in a dark cube mirror room, and we take a laser gun and fire one shot on the mirror wall, will the laser beam forever stuck inside that cube mirror?
Post 20 Jun 2007, 23:40
View user's profile Send private message Reply with quote
LocoDelAssembly
Your code has a bug


Joined: 06 May 2005
Posts: 4633
Location: Argentina
LocoDelAssembly
About light moving things http://en.wikipedia.org/wiki/Crookes_radiometer

I have one of these that my aunt bought to me when she went to Australia Very Happy

Also look at this http://en.wikipedia.org/wiki/Solar_sail
No, I don't have one of these Razz
Post 21 Jun 2007, 00:43
View user's profile Send private message Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
Quote:
kohlrak: "unproved theory"? please... this demonstrates you don't even know what "theory" is.


A theory is something that is thought to be a certain way, but never proven to be that way. At least, that's what i've been told most of the time. Though, for some reson, some science books say that theories are the same as facts. Though, i have a bad habit of using adjectives for emphasis.


LocoDelAssembly wrote:
About light moving things http://en.wikipedia.org/wiki/Crookes_radiometer

I have one of these that my aunt bought to me when she went to Australia Very Happy

Also look at this http://en.wikipedia.org/wiki/Solar_sail
No, I don't have one of these Razz


Let us not forget solar panels. I'm sure that to convert the "light energy" to "chemical energy" there has to be some movment of the electrons in the atoms, right?
Post 21 Jun 2007, 07:14
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
vid
Verbosity in development


Joined: 05 Sep 2003
Posts: 7105
Location: Slovakia
vid
Quote:
A theory is something that is thought to be a certain way, but never proven to be that way. At least, that's what i've been told most of the time. Though, for some reson, some science books say that theories are the same as facts. Though, i have a bad habit of using adjectives for emphasis.
you was talking about scientific theories, so you would better stick to scientific definition of theory. That's how the science and theory authors meant it.

by the way, not "some" science books, all science books (now we can argue what is science Wink )
Post 21 Jun 2007, 08:05
View user's profile Send private message Visit poster's website AIM Address MSN Messenger ICQ Number Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
Quote:
you was talking about scientific theories, so you would better stick to scientific definition of theory. That's how the science and theory authors meant it.


It's not necessarily a change of deffinition, but addition of an adjective to stress a particular part of it.

Quote:
by the way, not "some" science books, all science books (now we can argue what is science )


Are you saying that theories and facts are the same things?
Post 21 Jun 2007, 08:37
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
f0dder



Joined: 19 Feb 2004
Posts: 3170
Location: Denmark
f0dder
Post 21 Jun 2007, 10:07
View user's profile Send private message Visit poster's website Reply with quote
kohlrak



Joined: 21 Jul 2006
Posts: 1421
Location: Uncle Sam's Pad
kohlrak
f0dder: http://dictionary.reference.com/browse/theory .

Aside from the mocking i must propose the referance to the same site with the [url="http://dictionary.reference.com/browse/Scientific%20Theory"]scientific theory link[url]. 1 entry... Infact, even that entry only vaguely supports the wiki entry. Lemme put it this way. If mis-information is widely believed, it's probable (as i see it) that the mis-information will end up on wiki... By the way, what does the word tunami/tsunami/つなみ mean in japanese? I found a few pages on there that said it means "big tidal wave" in japanese. Perhaps in english, but i'm curious on what you feel it means in japanese. I will admit wiki does have some good information, but i don't see everything there as true (my opinion of wiki has changed since the last argument where wiki was used).
Post 21 Jun 2007, 10:34
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger Reply with quote
Display posts from previous:
Post new topic Reply to topic

Jump to:  
Goto page Previous  1, 2, 3, 4, 5  Next

< Last Thread | Next Thread >
Forum Rules:
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You can attach files in this forum
You can download files in this forum


Copyright © 1999-2020, Tomasz Grysztar. Also on YouTube, Twitter.

Website powered by rwasa.