[GraphiteBoi]

Just wondering what this forum thinks about the eventual coming of the G5? How will it perform? How fast will it get? Improvements? What will it look like?

[I'mDaMac]

If IBM develops the G5, which I believe they are, then there's hope. OTOH, if Moto's behind the G5 then there's something to worry about. It's too early to tell though.

[11011001]

Like this I guess.

[This message has been edited by 11011001 (edited 06-01-2001).]

[OverclockedHomoSapien]

http://e-www.motorola.com/webapp/sps...=M943030621280

Click on the PPC road map. The G5 is going to be done by Motorola, probably not for several years. It will be fast by today's standards, but very slow by the standards of the day when it is released and shipping.

[Nimisys]

Originally posted by 11011001:


Like this I guess.

damn thats an ugly case...

[Cipher13]

I like it

What a tank!

The G5 will be here soon, and its gonna scream...


------------------

AIM: Cipher1387
ICQ: 48111606
mail: [email protected]

[PeteL999]

Originally posted by OverclockedHomoSapien:
[URL=http://e-www.motorola.com/webapp/sps/technology/tech_tutorial.jsp?catId=M943030621280]http://e-www.motorola.com/webapp/sps/technology/tech_tutorial.jsp?catId=M943030621280[/UR L]

Click on the PPC road map. The G5 is going to be done by Motorola, probably not for several years. It will be fast by today's standards, but very slow by the standards of the day when it is released and shipping.

It will be much sooner than several years from now I'm thinkin' 2002 2003 somewhere in there. It WILL be fast

[macvillage.net]

I don't know, I kind of like the case.

------------------
Robert Accettura
Owner/Webmaster of

[Scotttheking]

Yuck!

[tyoda]

Well you know, these speculated G5 cases look a lot like the descriptions and rumors floating around about the next gen G4 towers, possibly due out this summer.

It follows the line that it has some type of clear outer frame and silver box inside (like the Cube).

The renderings look a bit chunky and squashed for some reason, but the idea looks valid....I guess.

[Fyre4ce]

I'm fairly certain that the G5 is going to be done primarily by IBM.

Please to not post such oversized graphics! Most of that image is empty black space anyway!

EDIT: One more thing: PLEASE DON'T PUT SUCH A BIG IMAGE IN YOUR QUOTE!

------------------
Fyre4ce

"I need a vacation." - Terminator robot

[This message has been edited by Fyre4ce (edited 05-23-2001).]

G5 is in testing by Apple with theid party hardware testers. These testers are primarily corperations which Apple has had good business with.

No word yet on case, model was seen in a G4 case.

Hope the G5 will be here in no time. Imagine a Quad 2 GHz G5.
I am dreaming on... !!

[capone]

Stop making me drool I have a new shirt

------------------
-capone

G4/400 448 30,10 Rage128, Radeon
Performa 6116CD/60/700
Quadra AV
LC 575
LCIII
Apple IIgs

[austeros]

rumors of the g5 might just convince me to get a ibook, and save some cash now... rather then a tibook. if the rumors are true, apple could make one killer of a laptop... it would really be nice to start seeing some of the old pentium snail stuff again.

------------------
-rok

[11011001]

Motorolla is currently experimenting with the G5, they got it up to 1.3 Ghz. Ya, I know that sounds bad, but the potential is way higher. At that speed it is going way to fast for its own good, and it crashes.

By the time it is released we should see speeds of at least 2 Ghz.

The processor will be 64 bit, which I think means that the max amout of ram you can have is many millions of terabytes! (2^64) compared to the maximum 4 gigs (2^32) on the current 32 bit chips like the G4 and the G3. Of course, the chip will be compatable with the current 32 bit applications.

When I say 2^32, I am taking about binary, and how the computer points to places in memory. If you have a variable that is 32 bits long, that means that the largest number it can store is 4 billion something. Since the computer uses things called pointers to point to a location in RAM, such as the 1st location, that means that a 32 bit computer can point to no more than 4 billion points in RAM, thus you can only have a max of 4 gigs. No on a 64 bit comp, you can have a number that is beyond a trillion, I don't know how to name it. So, the G5, if the rumors are correct will be able to support millions of terabytes of memory.

Oh, and Motorola is also looking at the G6. Expect this about 2 years after the G4.

The G4 may come out as early as the end of this year!

oh, ya, feel free to correct me if I made a mistake!

[I'mDaMac]

Originally posted by 11011001:
Oh, and Motorola is also looking at the G6. Expect this about 2 years after the G4.

The G4 may come out as early as the end of this year!

oh, ya, feel free to correct me if I made a mistake!

Eh? I thought the G4 was already out

ya, I meant the G5.

The G5 may be out by the end of the year.

And 2 years after that we should see the G6.

Interesting thing, is that even when the G5 is out the G3 will still be produced by Motorolla. So, perhaps the G3 will be around a while longer...

[Nimisys]

Originally posted by 11011001:

The processor will be 64 bit, which I think means that the max amout of ram you can have is many millions of terabytes! (2^64) compared to the maximum 4 gigs (2^32) on the current 32 bit chips like the G4 and the G3. Of course, the chip will be compatable with the current 32 bit applications.

When I say 2^32, I am taking about binary, and how the computer points to places in memory. If you have a variable that is 32 bits long, that means that the largest number it can store is 4 billion something. Since the computer uses things called pointers to point to a location in RAM, such as the 1st location, that means that a 32 bit computer can point to no more than 4 billion points in RAM, thus you can only have a max of 4 gigs. No on a 64 bit comp, you can have a number that is beyond a trillion, I don't know how to name it. So, the G5, if the rumors are correct will be able to support millions of terabytes of memory.

The G5 may come out as early as the end of this year!

oh, ya, feel free to correct me if I made a mistake!

actually for the 64bit CPU it will allow them to address files over 4gigs also, which is what makes them so important for database peoples... ahd thats what makes them such a big deal, is the larger file addressment.

as for it being 64bit, where'd you get that info on it? i am still studying the 64bit approaches being taken by the assorted mfgs, so i would like to see the resources on it. if it is also 32bit compatible, how much of the legacy RISC will be included, and what is the 32bit vs 64bit native performance estimites? will be like the AMD hammer line that will have superior 32bit, will still proving decent 64bit, or Intels Itanium style that has superior 64bit but dog slow 32bit? hows the 32 bit applications being done? in emulation? multi-core hardarwe (such as with the hammer line)? or is the 64bit side just a native extension of the PPC architure similar to AMD's x87-64 style or is it complete new like Intels IMAR (FORGOT ACTUAL NAME)?

[olePigeon]

Ugh, tell me about it. One of my HDs crashed and I tried to use a Copy Partition. It wouldn't let me copy the partition cuz it was over 2GBs.

[capone]

It will be done by moto. I still hope IBM gets back in the picture though. Combine the technologies for a better mac.

[11011001]

Here is some info from a maccentral document.

In fact, the G5 chip will scream at speeds of over to 2GHz if Motorola's predictions are on target. The microprocessor -- which should see the light of day by 2002 at the latest -- will come in 32-bit and 64-bit flavors, and so would sport the ability to "transparently" run present 32-bit applications. For the technically minded, the chip will have an extensible architecture, a new pipeline, new bus topology, and a 0.10 micron processor with SOI. The G5 is described as a "high performance microprocessor targeting computing and high-end embedded" markets.

As for the current speeds, from sources such as www.macosrumors.com (I can't remember if this is where I got it), they are at 1.3Ghz. They tested this with internet explorer, any higher than this the chip is going to fast for its own good, and ya start to get errors.

I think that using 32-bit applications under a 64-bit processor will not require emulation. It would require emulation though if the processor code were upgraded, to say a new version of RISK, but if this happened we would have that whole 68k, PPC deal all over again (which would not be so bad if the chips were a hell of a lot faster).

However, i doubt that there is a change to the chip language, because they were running the chip under Mac OS (probably X), and were running Internet Explorer and a few other programs. If the language were different these programs would have to be remade, and recompiled in order to work. So I guess we are stuck with the same old CPU language, huh.

As far as 32-bit, vs 64-bit. Hmmm, this might pose a problem. We are still using a 32-bit memory manager in our OS, so RAM is still limited (correct me if I am wrong). As for running a 32-bit application under a 64-bit processor, I don't think this should make much of a difference. I mean we should not see huge performance penalties for such a thing. Since this is apple and Motorola, I very much doubt that they would compromise 64-bit processing for the older 32-bit programs. These two companies always seem to want the cutting edge, might I even say revolutionary, technology. So the processor is porbably built for fast 64-bit.

Can anyone tell me why running 32-bit, under a 64-bit processor, might slow things down? ( Is it because of memory, (moving blocks to accomadate for this) or perhaps is it how it processes 32 bit compared to 64 bit. Help me out, cause I am not certain about this. )

PS -- Just typing in G5 processor, on a search engine reveals quite a bit of cool info about this PC killer

[Nimisys]

Originally posted by 11011001:
[B]
Can anyone tell me why running 32-bit, under a 64-bit processor, might slow things down? ( Is it because of memory, (moving blocks to accomadate for this) or perhaps is it how it processes 32 bit compared to 64 bit. Help me out, cause I am not certain about this. )

B]

it is because of the internals of the chip the hardware language will change... as such so does the compiler. in order to run the older apps, the chip needs to still be famliar enough to what the older compiler was expecting in order for the programs to run. thats why your ppc chips can't run x87 stuff unless it has been morphed into ppc first. so a 64bit chip effectivly needs to appear as a 32bit chip to older apps. there are a few way so doing so.

theres real-time code-morphing like what transmeta does with there notebooks.

theres base code extension like amd is saying they will do with there hammer line, by just adding the 64bit instructions into the existing x87 base.

there is intels apporach of making a completley new base that has a few limited instructions of the previous base

and then theres two seperate cores, one for 32bit, one for 64bit.

AMD seems to be doing 1,2 and 4. as such the 32 bit will remain faster than the 32 bit, but will still provide decent 64bit, but not at the same level of a pure 64bit chip. intel is doing #3 and as such 32bit will be much slower because of the limited base and the emulation needed. amd is managing to speed things up by using the code-morphing form transmeta and the multi-core of IBM, while extending the x87 to 64bit.

from what moto said then it seems that the g5 is probably #4 only..which could be a good or bad thing. it seems that way if they are running a OSX and current apps shows that the base language hasn't change significantly which means the internals must be similar to the current g3&4 lines of cpu, which further suggest multi-core. the fact that it would come into flavors means they need some way of disabling the 64bit side without creating two seperate chuips, once again multi-core.

for more info check arstechnica as they do a good job with the 64bit theory

[Spiffington]

what?!!?
i thought that the g4 was a 128-bit processor and that was why they performed much faster than pentiums...
this is what i read at apple.com:


The secret of the PowerPC
G4�s phenomenal
performance is its aptly
named Velocity Engine. It�s
the heart of a
supercomputer miniaturised
onto a sliver of silicon. The
Velocity Engine can
process data in 128-bit
chunks, instead of the
smaller 32-bit or 64-bit
chunks used in traditional
processors

have i mis-interperated???

[origin3]

Nimisys,

When IBM and Motorola "broke up" in 98/99 before the G4 (74xx-series) was introduced, they also made the 'Book E' specification for future 64-bit PowerPC processors. A presentation of this specification is given at www.chips.ibm.com/products/powerpc/booke_pz.pdf

The G5/75xx-series is such a 'Book E' processor. There will be new instructions for arithmetic and adressing. Chip implementations can be 32-bit or 64-bit, but Motorola have publicly stated that 32-bit code will run native on the 64-bit variant, unlike IA-64. (I don't know whether this means lower performance for 64-bit code...the presentation mentions "Optimize cost and performance for diverse system requirements".)

Over at the Future Hardware forum at Appleinsider (a place with low signal-to-noise ratio) there is some information from 'motoman' (PPC designer). We can expect the G5 to have 2*32kB L1 cache, 512 kb L2 cache (on-die), 13+ stages in the pipeline, RapidIO Interconnect Architecture. Too early too speculate in MHz and process technology, but I believe Mot and AMD are cooperating in these areas. 'motoman' also said it wouldn't arrive until fall 2002 or later. In the mean time we will see Apollo (7460?), which is an SOI-implementation of the current G4 7450.

I found the following on http://www.mdronline.com/epf/conf_1.html (Embedded Processor Forum is in about two weeks.)

Motorola's PowerPC Book E Core
Motorola Semiconductor Products Sector�Joe Chang, PowerPC Development Manager

Motorola's new core implements the PowerPC Book E architecture with extensions to optimize conditionals, context-switching, and numerics.This presentation covers previously undisclosed technical details of the new core's microarchitecture and physical implementation information and indicates how the core may be deployed in initial products.

'Book E' also allows for Application Specific Processing Units. (AltiVec - and maybe others later on.) IBM is working on a next generation PowerPC on their own, but it's too early to tell how different this will be from Mot's. RapidIO is certain, and also a SIMD unit (AltiVec ?). Take a look at IBM's site if you are interested. It seems like they are taking technologies from the POWER series and adding it to the PowerPC.
http://www.chips.ibm.com/products/powerpc/rdmap/

[anthology123]

The Veolcity Engine part of the G4 is 128 bit only, the core of G4 is a 32-bit processor. Velocity engine is a separate subsection of the G4, like the MMX or SIMD in a Pentium.

Even when G5 comes out, it will have to have MacOSX re-compiled for 64 bit, just like Win 2000 will be when Itanium hits the desktop. Don't tell me they won't be in desktops. Even though Itanium is billed as for servers only, don't count on it staying there for more than 3 months before appearing in standard consumer PCs.

[Nimisys]

the velocity engine or simd unit in the g4 can read upto 4 simotanous 3bit instructions or 2x64bit or 1x128. that is better than 3dnow, sse, and MMX, but then agin it came out after the others. thats what jobs menat by the 128bit part. technially you could call the k6-2 and beyond from amd and p3 and beyond from intel 128bit also as their simd units can read 2x64bit instructions thus giving 128 bit. but the interoir processing portion of the chip is still only 32 bit.

thats what we are referring to, moving the 32bit upto 64bit internally. the ability to run 32bit nativly within, seems to suggest the AMD approach or the multicore from IBM... i'll read the papers soon.

as for intanium in desktop machines...not blood likely..it cost to much and su><0rs in 32bit applications... i don't see any anouncments for 64bit photoshop, maya, softimage, lightwave, cad, ect. the only places that will matter will be serverside dtabsae that exceed 4gb in size.

the first real 64bit cpu that will get picked up by desktop enthusats will be AMD's K8 ClawHammer. Best yet it will incorparate the chipset into the cpu, giving each slot (pci,Pci-x,AGP, hdd, ect) its on bus. this will mean cheaper motherboard but possibly higher cpu costs. with amds approach no new recompile will be needed since the exsisting base is being expanded. thank you legacy ISA's. only if you want 64bit will you need to recompile the software to use it, but i am still wondering just howmuch that will be neccesary with transmetas code-morphing in it. i think moto/ibm will be taking a similar approach

[This message has been edited by Nimisys (edited 05-31-2001).]

[Nimisys]

Originally posted by Nimisys:
the velocity engine or simd unit in the g4 can read upto 4 simotanous 3bit instructions or 2x64bit or 1x128. that is better than 3dnow, sse, and MMX, but then agin it came out after the others. thats what jobs menat by the 128bit part. technially you could call the k6-2 and beyond from amd and p3 and beyond from intel 128bit also as their simd units can read 2x64bit instructions thus giving 128 bit. but the interoir processing portion of the chip is still only 32 bit.

thats what we are referring to, moving the 32bit upto 64bit internally. the ability to run 32bit nativly within, seems to suggest the AMD approach or the multicore from IBM... i'll read the papers soon.

as for intanium in desktop machines...not blood likely..it cost to much and su><0rs in 32bit applications... i don't see any anouncments for 64bit photoshop, maya, softimage, lightwave, cad, ect. the only places that will matter will be serverside dtabsae that exceed 4gb in size.

the first real 64bit cpu that will get picked up by desktop enthusats will be AMD's K8 ClawHammer. Best yet it will incorparate the chipset into the cpu, giving each slot (pci,Pci-x,AGP, hdd, ect) its on bus. this will mean cheaper motherboard but possibly higher cpu costs. with amds approach no new recompile will be needed since the exsisting base is being expanded. thank you legacy ISA's. only if you want 64bit will you need to recompile the software to use it, but i am still wondering just howmuch that will be neccesary with transmetas code-morphing in it. i think moto/ibm will be taking a similar approach

damn double posting unable to deleting ubb crap

[This message has been edited by Nimisys (edited 05-31-2001).]

I think there will be a G5 from Apple first but then WIndows will copy. Stupid Windows!

I love macintosh's

Kyle F

ICQ - 114294897

[sine -''-..-]

i have heard that the g5 will be a titanium'esque, sleek tower with some polycarb on the front. i expect it to be an obilisk (sorry spelling) of sorts, without handles but still sporting the side hatch.

[Nimisys]

Originally posted by <Macintosh_Kyle>:
<STRONG>I think there will be a G5 from Apple first but then WIndows will copy. Stupid Windows!

I love macintosh's

Kyle F

ICQ - 114294897</STRONG>

WTF are you talking about... Windows will copy?!... Windows isn't a ****ing CPU! nor is Apple... Intel already has the Itanium out the door, so technically speaking, they are the first. my guess is theat AMD will be next followed by IBM/Motorola closely. there is already the WinXP 64bit build in beta, so it also would be the first consumer if you are willing to call it that. true UNIX/LINUX (OSX not counting) has had 64bit for a while now, so they would be considered the first OS period. As for Consumer level OS, if you aren't willing to count WInXP64, then Windows would still be first as the two CPU mfg have already released their Suimulators for the chips, so software companies can start programming for the chips already...wheres Motorola/IBM's?

Damn Head-in-as$ zealots...