Is there evidence of an approaching G5? - Page 2

[quote]Originally posted by orb24:
<strong>The facts,
(...)
the g4 supports only two chips
(...)
</strong><hr></blockquote>

This is most definitely not true, as there are numerous quad G4 boards in the embedded market.

Bye,
RazzFazz

[quote]Originally posted by Scott H.:
<strong>What ever happened to that Matrox (sp?) "Real Time" video editing computers that even Apple said was coming but never came?

Maybe Apple did get G4/Alti-Vec lust and just dropped it?</strong><hr></blockquote>

It's about as useless as the g3/ apple video combo. Last I heard they were packaging it with final cut and a contour controller for 12 or 13 hundred (us$). haven't seen it since. I can't imagine matrox is too happy but maybe they made a very limited run.

FCP3 doesn't need a render card anymore it uses the cpu. So the only reason you need a card is interface the computer to your beta sp, digi beta, whatever to your computer.

[ 01-08-2002: Message edited by: orb24 ]</p>

Another reason to expect the G5 soon:

MWSF was a consumer event. iBooks, iMacs both were upgraded.

So at the next event it will be the pro machines that get upgraded. Apollo G4 for the Titanium, to push it up to 1 GHz or so, and G5 for the Powermacs, to push them up to 1.4 -1.6 GHz or so.

This makes too much sense for it not to happen!

[quote]Originally posted by RazzFazz:
<strong>

This is most definitely not true, as there are numerous quad G4 boards in the embedded market.

Bye,
RazzFazz</strong><hr></blockquote>


How can you compare embedded market to desktop market? and if you can then I can confirm that G5's are being used in embedded devices and should be seen in towers soon. It's not a question as to wether or not they exist it's just wether or not apple is using them now or later.

I too have seen the quad chip linux rendering briq's etc. so thanks for the reminder but I don't think that they represent the same thing at all. Although I just use it, don't build it.

[ 01-08-2002: Message edited by: orb24 ]

[ 01-08-2002: Message edited by: orb24 ]</p>

whether or not the g5 is around the corner, I am very intrigued about Apple's next design for their "towers"... I'm predicting front to back depth of just enough for full size pci cards, probably 2/3rds the depth of current towers... maybe round? titanium colored?? or silver, like a classic car.. Apple has perfected small computer devices: cube, imac II... I think we may be in for some surprise..

I think it was RickAg that asked me when I expected a G5 to be released by Motorola and I still stand by the answer I gave him, MWSF '03 (jan '03).

[quote]Originally posted by RazzFazz:
<strong>

There are dual AMD chipsets right now, but they are hard to come by, pretty expensive and need a special power supply.

Bye,
RazzFazz</strong><hr></blockquote>

That was only true for the Tyan Thunder K7 motherboard. Tyan shortly thereafter released the TigerMP motherboard that cost $200 less and used a standard ATX power supply. Recently a new wave of AMD MPX chipset dual Athlon motherboards were released/announced by multiple companies such as Asus, MSI, Tyan, and Abit. They all use standard ATX power supplies and will range in price from $199-$299.

[ 01-08-2002: Message edited by: Eskimo ]</p>

So, to sum up the thread there are exactly two sources for the near future relaease of the G5:

1) The Tinkerbell effect - apprently, if we all wish hard enough, it will be true (aka, the JYD/Ryan Meader effect).
2) A two year old link on ZDNET.

The evidence for the G4 Apollo (SOI or die shrink or both) continues to carry the OVERWHELMING evidence.

I wish rumors died faster.

The G5 is not coming in the next year. To wish otherwise is to set yourself up for disappointment.

Don't count out this Apollo chip, kids. SOI and die shrinks are serious enhancements to speed scaling. Combined with dual processor configs, the Powermac G4 will hold its own soon enough.

SdC

[ 01-08-2002: Message edited by: suckfuldotcom ]</p>

[quote]Originally posted by orb24:
<strong>
They'd be underestimating the power needs of FCP Adobe AE and even OSX. If that's their plan they'll regret not making headway into a market that they've wanted for years now simply because the users are ready but not willing to wait.</strong><hr></blockquote>

This sounds as though Apple is doing this all on purpose. I don't believe that's the case at all. They're not underestimating anything; I think they're doing everything they can, processor-wise. (Mobos are a different story.) I think they've got themselves into a tight spot with Motorola and there's nothing left to do at this point other than ride it through to the light at the end of the tunnel (the much-anticipated G5).

i believe the matrox realtime editing is out for mac i think its $1000 just their website

[quote]Originally posted by suckfuldotcom:
<strong>
The evidence for the G4 Apollo (SOI or die shrink or both) continues to carry the OVERWHELMING evidence.
</strong><hr></blockquote>

You seem to believe that multiple chip designs cannot be in progress at the same time. No chip design companies anywhere in the industry work this way. They all overlap their design cycles by anything from 6 months to two years. The current generation is refined and moved to better processes while the next generation is getting rolling.

Your refusal to admit that the G5 might exist makes it clear that you don't want to set yourself up for disappointment, but this doesn't have anything to do with whether or not the chip exists -- by exactly the same logic that you dismiss the optimists. The truth is most likely to be somewhere in between.

[quote]Originally posted by suckfuldotcom:
<strong>So, to sum up the thread there are exactly two sources for the near future relaease of the G5:

1) The Tinkerbell effect - apprently, if we all wish hard enough, it will be true (aka, the JYD/Ryan Meader effect).
2) A two year old link on ZDNET.

The evidence for the G4 Apollo (SOI or die shrink or both) continues to carry the OVERWHELMING evidence.

I wish rumors died faster.

The G5 is not coming in the next year. To wish otherwise is to set yourself up for disappointment.

Don't count out this Apollo chip, kids. SOI and die shrinks are serious enhancements to speed scaling. Combined with dual processor configs, the Powermac G4 will hold its own soon enough.

SdC

[ 01-08-2002: Message edited by: suckfuldotcom ]</strong><hr></blockquote>

Sorry, I'm new to this thread. Go to
<a href="http://www.theregister.co.uk/" target="_blank">The Register</a> and do a search for "G5". You'll find they've been posting articles about the supposed progress of the G5 pretty regularly for over two years (their first mention of the Moto G5 is October 6, 1999). That's the source of most of my expectations. While The Reg is occasionally off the wall, usually they are pretty reliable. I tend to believe that there's something to it.

[quote]Originally posted by orb24:
<strong>
How can you compare embedded market to desktop market? and if you can then I can confirm that G5's are being used in embedded devices and should be seen in towers soon. It's not a question as to wether or not they exist it's just wether or not apple is using them now or later.

I too have seen the quad chip linux rendering briq's etc. so thanks for the reminder but I don't think that they represent the same thing at all. Although I just use it, don't build it.
</strong><hr></blockquote>


What the heck did that mean? The message you are replying to simply stated that the G4 works perfectly fine in quad processor configurations (and probably higher too). Apple just hasn't shipped such a machine, but the existing chip design would support it.

The main technical issue would be that current G4 machines are already memory bandwidth limited, so doubling the number of processors would certainly not double the performance. Improving the memory subsystem would make a quad processor setup more effective.

There have been a variety of other rumours floating around for a long time from testers of prototype Apple machines. No doubt you will just dismiss these as invented, but you're not going to get much better than that from people who have signed the NDA agreements that Apple requires.

The performance increases tossed about (I mean the reasonable ones, not the silly 10x ones) are unlikely without a substantially better memory subsystem, and the G4's memory system is pretty much maxed out. Even the Apollo has no indication that it has gone beyond 133 MHz x 64-bits. So reports of big speed jumps are either of versions of the Apollo with an alternate bus, or of something else. An Apollo with an alternate bus is pretty much a G5 though.

Of course there are always wild prototypes floating about, so it doesn't tell us anything about product intro dates... but it doesn't justify the wildly pessimistic feelings post-MacWorld keynote. The keynote wasn't a setback, it just wasn't a step forward for the Pro line. The iMac buyers are certainly a happy lot (by and large).

[quote]Originally posted by orb24:
<strong>
How can you compare embedded market to desktop market?
</strong><hr></blockquote>

Those G4s on the embedded boards I'm talking about are the very same chips used by Apple, so if they can be made to work in a quad setup in the embedded market, the same is true in the desktop market.

Bye,
RazzFazz

[quote]Originally posted by Eskimo:
<strong>That was only true for the Tyan Thunder K7 motherboard. Tyan shortly thereafter released the TigerMP motherboard that cost $200 less and used a standard ATX power supply.</strong><hr></blockquote>

But (at least over here) it's still awfully hard to find a dealer that will actually sell one of those to you.


[quote]<strong>Recently a new wave of AMD MPX chipset dual Athlon motherboards were released/announced by multiple companies such as Asus, MSI, Tyan, and Abit. They all use standard ATX power supplies and will range in price from $199-$299.
</strong><hr></blockquote>

OK, didn't know that.
Besides, even once they are available, $299 seems pretty steep to me.

Bye,
RazzFazz

Suckful,
Admit it, you really don't know anything.

[quote]Originally posted by Programmer:
<strong>
The main technical issue would be that current G4 machines are already memory bandwidth limited, so doubling the number of processors would certainly not double the performance. Improving the memory subsystem would make a quad processor setup more effective.</strong><hr></blockquote>

I just read everything the Moto website had to offer in regards to the MPX bus, and as it seems, your last statement might not be true. From what the documentation says, it sounds like the G4's front side bus is *not* point to point, i.e. unlike the Athlon for example, every CPU in a G4 MP setup is *not* individually connected to the northbridge, but instead all G4s share the same FSB and consequentially share its bandwidth (roughly 1GB/s) too, so improving the memory subsystem beyond PC133 would not bring any benefits unless the G4s get an improved FSB first.

Bye,
RazzFazz

[quote]Well, I've owned the DP800 for the past 4-5 months and can tell you it was one of the most disappointing upgrades I ever made. For the money, the bang for the buck was minimal. I had upgraded from the DP533.<hr></blockquote>

Yeah, it does seem like a waste of money.

Not significant a upgrade to warrant a purchase IMO.

I would have waited to get a chip that had AT LEAST double the clock speed.

These people buying a new machine for 200-400 MHZ increase is just insane.

But if you got money to burn, what the heck.

orb24
"if you can then I can confirm that G5's are being used in embedded devices"

I was unaware that any G5's of any flavor had even been sampled yet. Please show me a link. I'll show you mine.

<a href="http://www.motorola.com/mediacenter/news/detail/0,1958,568_322_23,00.html" target="_blank">http://www.motorola.com/mediacenter/news/detail/0,1958,568_322_23,00.html</a>

Motorola Discloses First e500 Integrated Host Processor To Employ RapidIO Interconnect Technology

"Samples of the MPC8540 are expected to be available in the second half of 2002."

jimmac: I know what is reasonable and feasible. I know that all the hoping in the world doesn't change all the engineering in Moto's Semicon Div. I know that buying Ryan Meader's bullshit is the quickest way to delusion.

Let's face it kids: If the 8540 isn't even sampling yet, the full 85xx generation processor is, at best only being knocked in test quantities, and even that is being optimistic.

SdC

[quote]Originally posted by RazzFazz:
<strong>

I just read everything the Moto website had to offer in regards to the MPX bus, and as it seems, your last statement might not be true. From what the documentation says, it sounds like the G4's front side bus is *not* point to point, i.e. unlike the Athlon for example, every CPU in a G4 MP setup is *not* individually connected to the northbridge, but instead all G4s share the same FSB and consequentially share its bandwidth (roughly 1GB/s) too, so improving the memory subsystem beyond PC133 would not bring any benefits unless the G4s get an improved FSB first.

Bye,
RazzFazz</strong><hr></blockquote>

Agreed, although I had been thinking they could build a memory controller that had 2 MPX busses coming out of it. Less efficient to be sure (the bus snooping would be relying on the memory controller between the busses), but it could be made to work. I doubt they would do this if there was a new FSB on the horizon anywhere.

just recently i read an interview with some guy responsible for parts of the semiconductor division of moto. the interesting part about that interview was, that it wasn't related to apple that much, but rather general about the next couple of big products coming from moto.
the moto guy gave VERY detailed descriptions of the apollo chip, the state of development (which sounded like it would come out quite soon) and everything around that family of products. when they (repeatedly) got to the topic of the next desktop cpu line - while mentioning the 8540 - the moto official blocked completly. his only comment was, that moto has agreements with bigger client"s" enabling the clients to completely decide how much information about the development of a product of high interest for that client is made public. and he repeatedly said that with growing emphasis on "full control" which, i think, includes everything from interviews (as proven) to microprocessor forum presentations...

from now on it's only a short, very easy chain of assumptions which are quite reasonable.

client = apple (because he explicitly talked about the cisco contract, regarding the embedded e500 core based chips)
-&gt; sj controls flow information
-&gt; SILENCE (anybody will agree to that last one, right

whatever. that doesn't say ANYTHING about the release dates of any of the chips, BUT the clou is inside what it just leaves out: apple doesn't have an agreement with them on keeping quiet about the apollo... why not? they will be the number one customer for that cpu.... by far... and sj would keep anything top secret if he had the chance to do so.

i don't believe that the g5 will come out at any time. i will wait with that till i don't have to believe anymore, but can know instead. (knowing for sure is MUCH less frustrating as believing - something which might prove healthy for some of the poor guys still sitting in front of stacks of hats they gotta eat cause of no pro announcement in sf. - btw. what about that guy who wanted to eat a cockroach and promised to post a pic?!?)