That's why it's called a non-disclosure agreement. He's not going to disclose actual perfromance as not only would that get him in trouble with Apple but his own employer to which I would think he has enough loyalty not to betray confidential information like that.</strong><hr></blockquote>
Assuming he has an NDA, which either he or his company would most certainly have if they are getting access to seed harware, then trust me on this one. He's already violated his NDA.
As for the actual report, some parts of it actually seem rather off to me (especially regarding the G4's), and of course, it doesn't change the fact that you aren't going to see a G5 in a Mac any time soon.
<strong>G5's have a 400 or 500MHz system bus that runs at a 16bit-wide data path. This is the RadipIO bus to a controller that controls most of the onboard peripherals such (...) Processor speeds are 1200/1250MHz to 1400/1500MHz. The processor has a memory controller included on die connected to the core by a 256bit memory bus running at one quarter the processor clock speed. The controller supports 2 memory bus but Apple only uses one so the internal width is cut to 128 bit at one quarter processor core speed. This provides more than enough bandwidth for PC2100 and the controller is designed to handle PC2700 memory ....</strong><hr></blockquote>
a memory bus speed 1/4 of the processor bus speed? on a 1200Mhz System that would be 300Mhz. i don't know how it works - therefor my question: can we use 266Mhz DDR Ram on a 300Mhz 128bit-wide memory bus? and can we use 333Mhz DDR RAM on a 350Mhz 128bit-wide memory bus? same thing as for the 312.5Mhz/375Mhz memory bus of the 1250Mhz/1500Mhz G5...
if there is an expert for this things. please reply to it.
400Mhz to 500Mhz RapidIO is correct. the 8540 has 500Mhz RapidIO...
Since moki keeps emphasizing that there wont be any G5 anytime soon (which fewer an fewer people believe), I'll further on reply to his threads with the following line:
You can run 133 Mhz memory with a 100 Mhz main bus, for example.
(replace 133 by 266 if you wish, but saying "266 Mhz" is misleading, as it is still running at 133 Mhz, with a bandwith theorically equivalent to 266 Mhz).
DDR 2700 is running at 166 Mhz ("equivalent to 333"), but can be used with current Athlon processors running on a 133 Mhz, while keeping the full memory speed. Just running async.
Now, I seriously doubt Apple would make the system bus run at 1/4th of the processor speed. Simply because it would be a nightmare to design such a motherboard. Especially if the bus has to be kept a relatevely even multiplier with PCI (33 Mhz) or AGP (66 Mhz).
A 400 Mhz main bus with DDR memory could meen two things:
- DDR 1600 memory, running by pairs (that's what is done with Pentium 4 + Rambus : 100 Mhz x 2 x 2). Running synchronously with main bus.
- DDR 2100 memory, 133 Mhz x 2, running asynchronously.
Since the G5 has its own memory controller, which is no longer located on the northbridge, itself located on the main bus, I will let you guess which answer is right.
<strong>A system bus can run asynch. from the memory bus.</strong><hr></blockquote>
right. but dorsals quote was, that the system bus of G5 will run at 400 or 500Mhz and the memory bus will run at 1/4 the processor speed. so will this work? 300Mhz memory bus (not system bus) with 266 DDR Ram(PC2100?) ...
it doesn't change the fact that you aren't going to see a G5 in a Mac any time soon.</strong><hr></blockquote>
Mr. Welch, you'll have to forgive us if we don't take your arguments seriously. There are a couple of possibilities:
- If you were under NDA and had prototype machines you wouldn't admit to knowing anything with your name attached to each message
- Ambrosia is hardly a large and important software developer (sorry), so correct operation of your products on the day that they ship machines isn't exactly critical. And on the flip side, I'm sure Apple employees love testing your products on all the new hardware. Its probably very extensive testing too. This means you aren't first in line for what is always a limited number of prototypes.
- Apple is running a tight ship these days so where you may have been aware of future processors in the past, that may no longer be the case because of their efforts to plug leaks. That you post here at all probably increases your "leak risk" in Apple's eyes.
- The G5 is purported to be fully compatible with the G4. You may have had advanced warning of the G4 because Apple was privately talking up AltiVec to developers well in advance of the chip's arrival. For the G5 there is no new software technology to prime developers with, so prototypes will only go to those affected by the new hardware -- PCI card developers, for example.
So, no offense, but we can't really take what you say here very seriously.
Personally if phil is right 1.2 ghz is enough for the high end. I am not very tech savvy but it seems with all the chip and motherboard advances that would be introduced with the g5 it would at least compete with intel and amd if not more.
<strong>Personally if phil is right 1.2 ghz is enough for the high end. I am not very tech savvy but it seems with all the chip and motherboard advances that would be introduced with the g5 it would at least compete with intel and amd if not more.
Hopefully we get it soon</strong><hr></blockquote>
Bah, that's the wrong attitude and I'm pretty sure its not the one Apple has. They should be aiming to leap as far ahead of the current WIntel machines as possible because they leap less often, and the competition usually catches up in a series of rapid steps.
This speculation is irrelevent, however, because the speeds that arrive are determined by processor yields and costs. Apple will ship the fastest machine that they can manage and still be price competitive and have sufficient supply. They will not ship a 1.2 instead of a 1.5 because it is "fast enough", they will do it because they can't get enough fast chips.
Man you guys are tough n Dorsal. He simply posts what's he had the chance to play on, the machines and what he believes are in these machines. He makes no predictions, he doesn't use the "Confirmed" word, he doesn't say expect the G5 by next week, and yet a few of you are calling for the death of his dog simply for sharing his own personal experiences.
And you wonder why there's no real inside info here, the insiders are probably scared to get lynched by posting info that won't be believed not to mention their posting history or creditabilty.
He was right that we would stay at ATA/66 when we all thought it had to be ATA/100. Everyone was shocked. Yes, he has been wrong too.</strong><hr></blockquote>
I did not mean to take an attitude I more was asking a question in a statement form. I know I make no sense sorry. I agree apple should put out their best because they don't update as often but I really wanted to know if I was right. Would a g5 compete at 1.2 ghz
Eugene is fundamentally correct. Dorsal has not said a single thing in his post that hasn't already been speculated to death in here over the last three months. He's combined and regurgitated all the most reasonable predictions / rumors in his own writing style knowing that 80% of you will bow at his feet, thanking him for his "wisdom."
Maybe he's seen prototypes recently and maybe he hasn't, but he himself admits he doesn't know what's coming or when - how is that of any benefit to anyone? To know that there might be a G5 prototype out there helps you, even though you may not see one for a year or more? Based on his post, there's just as much chance we'll see one a year and a week from now, as there is that we'll see one in a week. He basically came right out and said "I have no idea what's next" and yet you thank him as if he's graced us with some amazing piece of information. Use your heads, people.
Remember also that with enough reasonable guesses the law of averages pretty much dictates a person will hit the proverbial bullseye sooner or later. He hit his sooner and has pretty much been wrong ever since as far as I can tell.
on the other hand, does no one else see the post by dorsal m as a post that should not be criticized in the infantile way that eugenechan.com continually does? it used to be entertaining, eugene, now it is isnt. just stop it, kid. it doesnt make you big in any of our estimations.
Comments
<strong>
That's why it's called a non-disclosure agreement. He's not going to disclose actual perfromance as not only would that get him in trouble with Apple but his own employer to which I would think he has enough loyalty not to betray confidential information like that.</strong><hr></blockquote>
Assuming he has an NDA, which either he or his company would most certainly have if they are getting access to seed harware, then trust me on this one. He's already violated his NDA.
As for the actual report, some parts of it actually seem rather off to me (especially regarding the G4's), and of course, it doesn't change the fact that you aren't going to see a G5 in a Mac any time soon.
So are you really the President of Ambrosia Software, Inc.?
- Mark
<strong>G5's have a 400 or 500MHz system bus that runs at a 16bit-wide data path. This is the RadipIO bus to a controller that controls most of the onboard peripherals such (...) Processor speeds are 1200/1250MHz to 1400/1500MHz. The processor has a memory controller included on die connected to the core by a 256bit memory bus running at one quarter the processor clock speed. The controller supports 2 memory bus but Apple only uses one so the internal width is cut to 128 bit at one quarter processor core speed. This provides more than enough bandwidth for PC2100 and the controller is designed to handle PC2700 memory ....</strong><hr></blockquote>
a memory bus speed 1/4 of the processor bus speed? on a 1200Mhz System that would be 300Mhz. i don't know how it works - therefor my question: can we use 266Mhz DDR Ram on a 300Mhz 128bit-wide memory bus? and can we use 333Mhz DDR RAM on a 350Mhz 128bit-wide memory bus? same thing as for the 312.5Mhz/375Mhz memory bus of the 1250Mhz/1500Mhz G5...
if there is an expert for this things. please reply to it.
400Mhz to 500Mhz RapidIO is correct. the 8540 has 500Mhz RapidIO...
greets,
krassy
Thanks Eugene.
"Carry On"
CPU: Either a 7450 variant from 866MHz to over 1GHz or an 8500 at 1.2GHz to 1.4GHz.
AGP: Either 4X or 8X
USB: Some have 2 ports, some more.
IDE: Present
RAM: G4s-SDRAM G5s-Either PC2700 or PC2100
Dorsal's Guess: "We may see one or the other, but I think we might see both G4s and G5s".
What's up now?
The G5 and EV Nova will be released the same day.
G-news
<strong>
No, Nostradamus = PowerMac G4/codename from the old AI.</strong><hr></blockquote>
Yeah we all knew that, I was just thinking maybe he added another multiple personality to the mix.
BTW: Welcome back, you came out from your igloo!
Most Athlon motherboards to do this.
You can run 133 Mhz memory with a 100 Mhz main bus, for example.
(replace 133 by 266 if you wish, but saying "266 Mhz" is misleading, as it is still running at 133 Mhz, with a bandwith theorically equivalent to 266 Mhz).
DDR 2700 is running at 166 Mhz ("equivalent to 333"), but can be used with current Athlon processors running on a 133 Mhz, while keeping the full memory speed. Just running async.
Now, I seriously doubt Apple would make the system bus run at 1/4th of the processor speed. Simply because it would be a nightmare to design such a motherboard. Especially if the bus has to be kept a relatevely even multiplier with PCI (33 Mhz) or AGP (66 Mhz).
A 400 Mhz main bus with DDR memory could meen two things:
- DDR 1600 memory, running by pairs (that's what is done with Pentium 4 + Rambus : 100 Mhz x 2 x 2). Running synchronously with main bus.
- DDR 2100 memory, 133 Mhz x 2, running asynchronously.
Since the G5 has its own memory controller, which is no longer located on the northbridge, itself located on the main bus, I will let you guess which answer is right.
Bruno
(G5 IS coming)
<strong>A system bus can run asynch. from the memory bus.</strong><hr></blockquote>
right. but dorsals quote was, that the system bus of G5 will run at 400 or 500Mhz and the memory bus will run at 1/4 the processor speed. so will this work? 300Mhz memory bus (not system bus) with 266 DDR Ram(PC2100?) ...
thx for info...
Processor speeds are 1200/1250MHz to 1400/1500MHz. <hr></blockquote>
It could be that Some test boxes were 1200-1400 MHz but these were definately overclocked prototypes.
You will get G5's but only 800/1000/1200 for the time being.
<strong>
it doesn't change the fact that you aren't going to see a G5 in a Mac any time soon.</strong><hr></blockquote>
Mr. Welch, you'll have to forgive us if we don't take your arguments seriously. There are a couple of possibilities:
- If you were under NDA and had prototype machines you wouldn't admit to knowing anything with your name attached to each message
- Ambrosia is hardly a large and important software developer (sorry), so correct operation of your products on the day that they ship machines isn't exactly critical. And on the flip side, I'm sure Apple employees love testing your products on all the new hardware. Its probably very extensive testing too. This means you aren't first in line for what is always a limited number of prototypes.
- Apple is running a tight ship these days so where you may have been aware of future processors in the past, that may no longer be the case because of their efforts to plug leaks. That you post here at all probably increases your "leak risk" in Apple's eyes.
- The G5 is purported to be fully compatible with the G4. You may have had advanced warning of the G4 because Apple was privately talking up AltiVec to developers well in advance of the chip's arrival. For the G5 there is no new software technology to prime developers with, so prototypes will only go to those affected by the new hardware -- PCI card developers, for example.
So, no offense, but we can't really take what you say here very seriously.
Hopefully we get it soon
<strong>Personally if phil is right 1.2 ghz is enough for the high end. I am not very tech savvy but it seems with all the chip and motherboard advances that would be introduced with the g5 it would at least compete with intel and amd if not more.
Hopefully we get it soon</strong><hr></blockquote>
Bah, that's the wrong attitude and I'm pretty sure its not the one Apple has. They should be aiming to leap as far ahead of the current WIntel machines as possible because they leap less often, and the competition usually catches up in a series of rapid steps.
This speculation is irrelevent, however, because the speeds that arrive are determined by processor yields and costs. Apple will ship the fastest machine that they can manage and still be price competitive and have sufficient supply. They will not ship a 1.2 instead of a 1.5 because it is "fast enough", they will do it because they can't get enough fast chips.
And you wonder why there's no real inside info here, the insiders are probably scared to get lynched by posting info that won't be believed not to mention their posting history or creditabilty.
<strong>
He was right that we would stay at ATA/66 when we all thought it had to be ATA/100. Everyone was shocked. Yes, he has been wrong too.</strong><hr></blockquote>
Who was shocked?
No delusion like self-delusion.
Eugene is fundamentally correct. Dorsal has not said a single thing in his post that hasn't already been speculated to death in here over the last three months. He's combined and regurgitated all the most reasonable predictions / rumors in his own writing style knowing that 80% of you will bow at his feet, thanking him for his "wisdom."
Maybe he's seen prototypes recently and maybe he hasn't, but he himself admits he doesn't know what's coming or when - how is that of any benefit to anyone? To know that there might be a G5 prototype out there helps you, even though you may not see one for a year or more? Based on his post, there's just as much chance we'll see one a year and a week from now, as there is that we'll see one in a week. He basically came right out and said "I have no idea what's next" and yet you thank him as if he's graced us with some amazing piece of information. Use your heads, people.
Remember also that with enough reasonable guesses the law of averages pretty much dictates a person will hit the proverbial bullseye sooner or later. He hit his sooner and has pretty much been wrong ever since as far as I can tell.
[ 01-27-2002: Message edited by: Moogs ? ]</p>
[ 01-27-2002: Message edited by: Belle ]</p>
<strong>
Yeah we all knew that, I was just thinking maybe he added another multiple personality to the mix.
</strong><hr></blockquote>
Different ip addresses, different parts of the country (Houston, NY/NJ)
[quote]
BTW: Welcome back, you came out from your igloo!
<hr></blockquote>
err, thanks, never was really gone.