27-Apr-2024 14:39 GMT.		 UNDER CONSTRUCTION
Anonymous, there are 73 items in your selection (but only 23 shown due to limitation) [1 - 50] [51 - 73]
[News] Articia IANN.lu
Posted on 01-Jan-2004 20:24 GMT by ID4 - tHe SuRvIvOr -73 comments
View flat
View list
AGPx8, DDR2, MPEG4 ..... Are some of the new Articia I chipset features!!! Link here AGPx8, DDR2, MPEG4 ..... Are some of the new Articia I chipset features!!! Link here
Articia I : Comment 51 of 73ANN.lu
Posted by Amon_Re on 03-Jan-2004 16:18 GMT
In reply to Comment 47 (Alkis Tsapanidis):
He didn't said it was, well, atleast that's what i think.

Cheers
Articia I : Comment 52 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:32 GMT
In reply to Comment 14 (BrianK):
“Forward looking statements” is fine, but it’s quite different from actually selling the products.
Articia I : Comment 53 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:40 GMT
In reply to Comment 20 (Eric_Z):
>Just because Apple only sells CPUs running at 2.0 Ghz does not mean that IBM >can't manufacture faster CPUs,
Note that Apple is a tier one PC vendor.

>there are other things to consider like heat dissapation and the increased >bus speed as the 970 scales.
What about just a minor 0.2Ghz increase? Yields would be another issue.
Articia I : Comment 54 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:42 GMT
In reply to Comment 16 (BrianK):
I don’t operate on rumors i.e. I’m looking for engineering releases.
Articia I : Comment 55 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:45 GMT
In reply to Comment 16 (BrianK):
There are many PPC rumors in the past thus their less creditable rumor acceptance.
Articia I : Comment 56 of 73ANN.lu
Posted by Kjetil on 03-Jan-2004 23:45 GMT
In reply to Comment 48 (Matt Parsons):
I wonder how they are going to pull that one off, sound like an extra conversion to me, not practical inn side an computer anyway, I guess a lot of most change, like depending on new memory storage that can analyze the serial data at hi speed and convert it to address in memory and data to store inn an time frame that can compete whit the normal way of doing things,

Technology like that have it limits, guess they modulate the signal on different frequencies for data bus and address bus, and filter the different bus data out by middle band filter, the only big problem will be an AD/DA converter that need to work whit in speed that are exemptible on Serial BUS,

That technology will not go longer the USB memory sticks if I’m right.
Articia I : Comment 57 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:48 GMT
In reply to Comment 16 (BrianK):
>Use your favorite search engine for PowerPC970fx.

I’m well aware of rumors and certain “forward looking statements” from the said entities.
Articia I : Comment 58 of 73ANN.lu
Posted by Matt Parsons on 03-Jan-2004 23:49 GMT
In reply to Comment 56 (Kjetil):
Um... Think HyperTransport :-)
Articia I : Comment 59 of 73ANN.lu
Posted by hammer on 03-Jan-2004 23:58 GMT
In reply to Comment 14 (BrianK):
The indication of reach able speeds IF one over clock existing products to 2.6Ghz via liquid cooling.

This was one of the methods to find the potential reachable clock speed for the pre-1Ghz Athlon era. Via liquid cooling; the K8 FX-51was clocked to ~3.1Ghz.
Articia I : Comment 60 of 73ANN.lu
Posted by Eric_Z on 04-Jan-2004 10:59 GMT
In reply to Comment 53 (hammer):
You'd still get a bus speed increase of 100Mhz, this is no good if you have a controller chip that can only handle a 1Ghz Elastic-bus fsb.

My original statement stands though, why would IBM go and lie to chinese goverment officials that they can produce CPUs running @ up to 2.5Ghz for them, if they only can produce CPUs capable of running @2.0Ghz? It's not like a pdf is THAT hard to alter.
Articia I : Comment 61 of 73ANN.lu
Posted by Matt Parsons on 04-Jan-2004 11:24 GMT
In reply to Comment 60 (Eric_Z):
Have you never told a little white lie while dating?
Articia I : Comment 62 of 73ANN.lu
Posted by Kjetil on 04-Jan-2004 12:19 GMT
In reply to Comment 61 (Matt Parsons):
I predict optic computers, being the next great jump forward, you can have dozen optical fibers on the same space you will have copper weiring and it do not conduct heat, the only problem is that world do not have optical transistors yet, optical memory solved just look at CDRW, may be the same technology can be used for optical transistors some how, whit optical fibers you can send light bout ways at the same time on the bus whit out problems on the computer
Articia I : Comment 63 of 73ANN.lu
Posted by Eric_Z on 04-Jan-2004 13:38 GMT
In reply to Comment 61 (Matt Parsons):
Of course I think that IBM (lke all "good" buissneses) will bend the truth a little when they are marketing there chips. I just don't think it's propable that they would choose this as one of those things, as it's a relatively easy thing to check if they are telling the truth or not. I don't think that the Chinese officials would be very pleased if they placed an order for those 2.5Ghz 970 chips only to learn that either that those figures was sonething that IBM wiched that they could scaled the 970 to. Or that the yields of the 2.0Ghz+ CPUs where so low that the real limit is more like ... well ... 2.0Ghz. There are other, "fluffier", figures that they can bend to there advantage without lying, but while certanly pushing the limits of credabillety, like SPEC numbers for instance.
Articia I : Comment 64 of 73ANN.lu
Posted by hammer on 04-Jan-2004 20:31 GMT
In reply to Comment 60 (Eric_Z):
>You'd still get a bus speed increase of 100Mhz, this is no good if you have a >controller chip that can only handle a 1Ghz Elastic-bus fsb.

My statements have nothing to do with FSB speed. IF that was the case I'm not quite impresses due to competition starting from the letter "A" (not Apple Corp).

>My original statement stands though, why would IBM go and lie to chinese >goverment officials that they can produce CPUs running @ up to 2.5Ghz for >them, if they only can produce CPUs capable of running @2.0Ghz? It's not like >a pdf is THAT hard to alter.

The real capability is when they actual deliver the products at reasonable quantity. Note that AMD/Intel/'rich overclocker fanatics' can "cherry pick" CPUs for maximum speed. Switching to different processes for clock speed increase is a no brainier when competition also see this as an avenue for clock speed increases.
Articia I : Comment 65 of 73ANN.lu
Posted by Eric_Z on 04-Jan-2004 23:05 GMT
In reply to Comment 64 (hammer):
Perhaps I should have worded it a bit differently:

Just because Apple only sells systems with CPUs running at 2.0 Ghz does not mean that IBM can't manufacture faster CPUs, there are other things for Apple to consider like heat dissapation and the increased bus speed as the 970 scales.
Articia I : Comment 66 of 73ANN.lu
Posted by BrianK on 05-Jan-2004 00:45 GMT
In reply to Comment 57 (hammer):
@ Hammer...

Not sure why you had to post the same message worded in 3 different ways as a response but whatever.

Once again January brings the next Apple conference. So, we'll see if/what rumors are confirmed within the next week as far as Apple's PPC use.
Articia I : Comment 67 of 73ANN.lu
Posted by hammer on 05-Jan-2004 06:27 GMT
In reply to Comment 65 (Eric_Z):
“Can manufacture” and “actually delivering the product” can be a different things.
Articia I : Comment 68 of 73ANN.lu
Posted by hammer on 05-Jan-2004 06:32 GMT
In reply to Comment 65 (Eric_Z):
IF want you to see the potential or obtaining clues for clock speed outcomes is to over clock it with liquid cooling (testing the design of the CPU core with the current processes).
Articia I : Comment 69 of 73ANN.lu
Posted by Olegil on 05-Jan-2004 07:12 GMT
In reply to Comment 68 (hammer):
Uhm, this is a very special CPU we're talking about.

The FSB is quite high, and the multiplier is fixed (to 2).

How would you begin overclocking it?
Since you can't change multiplier like on the G3/G4, you would need to change the bus frequency. That means you need support chips able to overclock as many percent as the CPU.
Articia I : Comment 70 of 73ANN.lu
Posted by Anonymous on 05-Jan-2004 09:40 GMT
In reply to Comment 36 (Neko):
Neko, you are talking out of your arse, the Pegasos AGP Slot is
clearly coded 3.3V and PCI-X which this "AGP" is based on is 3.3V,
too.
Articia I : Comment 71 of 73ANN.lu
Posted by hammer on 06-Jan-2004 01:02 GMT
In reply to Comment 69 (Olegil):
> How would you begin overclocking it?
> Since you can't change multiplier like on the G3/G4, you would need to
> change the bus frequency. That means you need support chips able to over
> clock as many percent as the CPU.

Well,
- Purchase some fast PC4000 and PC4500 DDRSDRAM.
- AMD K8's FSB is at 1600 Mhz and it didn’t quite stop those 3DMarks2001/2003 point scorers in over clocking it (e.g. multiplier locked ‘Athlon 64’). Apple’s FSB is at 1000Mhz* as I recall. Both are Hypertransport based technologies as I recall.

- In over clocking the .13 micron Athlon XPs (T-Bred/Barton) (i.e. beyond 2.4 Ghz), some of extra cooling was applied for motherboard’s power regulators and Northbridge.

- purchasing a good power brick with plenty of Amps output on 3.3 volt rail (ATX).

- Purchasing of liquid cooled heat sink solution. A Dry Ice/Heat Pipe/LN2 could also work.

*Effective.

I don’t know IF Apple’s hyper transport has the ability lock down AGP and PCI speeds. This factor is critical for the hard disk. Principles from overclocking X86 processors maybe applied for the said PPC systems.
Articia I : Comment 72 of 73ANN.lu
Posted by JoannaK on 06-Jan-2004 08:27 GMT
In reply to Comment 71 (hammer):
Hammer... Elastibus is a way different from Hypertransport.. But those overclocking Ideas look nice..
Articia I : Comment 73 of 73ANN.lu
Posted by BrianK on 08-Jan-2004 16:06 GMT
In reply to Comment 59 (hammer):
@Hammer,

Did you follow the events at Apple this week?

Apple's new Xserve G5 systems use 0.09-micron G5 chips from IBM Thus, it appears Apple/IBM have been working on a smaller process for the G5. Which, in turn, better sets itself up for faster clock speeds. With the tracking of .09 micron G5's it's only a matter of time until the desktop ramps up and the G5 Powerbook becomes a product.

We'll see in July how Jobs prediction of 3Ghz G5 machines within a year pans out.
Anonymous, there are 73 items in your selection (but only 23 shown due to limitation) [1 - 50] [51 - 73]
Back to Top