PDA

View Full Version : G5 = IBM Powe4


Mr Jobs
Jul 13, 2002, 05:42 PM
MORS are claming that the G5 next year will be based on the IBM Power4 chip architecture, i really hope so...bring on the 64bit of AMD and Intel.:D :D :D

rice_web
Jul 13, 2002, 05:56 PM
Impossible, the Power4 sells for far too much. But, it would be cool.

Mr Jobs
Jul 13, 2002, 06:01 PM
Originally posted by rice_web
Impossible, the Power4 sells for far too much. But, it would be cool.

-it does say "based on the Power4 architecture" so it might be a cut down version

- by next year the prise may come down low enough

- by apple useing Power4 in powermacs IBM will be selling way more Power4 chips then they ever did so they can afford to sell for less

Megaquad
Jul 13, 2002, 06:02 PM
they cant call it G5,thats motorolas trademark if im not mistaken
and whats with all those programers who were working really hard to optimize apps for altivec? especially big players like adobe etc.

Mr Jobs
Jul 13, 2002, 06:08 PM
no the Gx is an Apple trade mark, G3 was originally a internal code name but everyone know about it and it was so popular Apple kept it as a release name.

Funkatation
Jul 13, 2002, 06:20 PM
http://www.theinquirer.net/Default.aspx?article=4266

read near the bottom...

"On the IBM side, expect a very different approach in POWER5 for the 2004 - besides the usual clock, cache, memory, CPU execution parallelisation and interconnect improvements, there will be "Fast Path" hardwiring (on-chip hardware acceleration) of some common tasks like TCP/IP processing - later maybe even stuff like high-level database or bioinformatics routines. The 0.13 micron copper POWER5 will have 2-way simultaneous multithreading, just like Xeon, and is expected to run well above 2 GHz. It should be much cheaper, cooler and less power-guzzling than POWER4. Don't be surprised to see Apple Macs or thin blade servers based on POWER5."

rice_web
Jul 13, 2002, 06:32 PM
I like that last possibility. I sounds plausible enough.

And yes, I do realize that the next generation processor used in the PowerMac could be a "dumbed down" version of the Power4, how much would they have to cut to make it affordable?

jadam
Jul 13, 2002, 07:02 PM
well... the really expensive Power4's have like 128mb DDR L3 cache, and 32mb L2 cache... so... i dont think they would need to be Dumbed down that much, just lower the cache!

dongmin
Jul 13, 2002, 07:03 PM
Can anyone provide any links to descriptions/analyses of the Power4? How does it compare to Moto's 8500?

It's an exciting rumor; the prospect of ditching Motorola is tantalizing. But it seems like a bit of a shot in the dark. And we'll have to wait at least a year...

rice_web
Jul 13, 2002, 07:36 PM
Well, the Power4 cannot be compared to the 8500 series, because the G5 from Motorola isn't ready in any form but an embedded one.

But, I found this link. It's fun to read:
http://commerce.www.ibm.com/content/home/shop_ShopIBM/en_US/eServer/pSeries/high_end/690.html

vonboom
Jul 13, 2002, 08:15 PM
[FONT=courier new][SIZE=3]its gotta be better than a motochip eh!!
all my money is belong to apple:D
"dont fall in love with your next idea"

Catfish_Man
Jul 13, 2002, 08:42 PM
...but it kicks ass because it's huge and expensive. 125 watts of power, costs a fortune, and has 10GB/sec of memory bandwidth (which it gets by having a multichannel bus that would be rather impractical on a non-server machine, afaik). The Motorola 8540 (there is no 8500, officially) is very similar to a stripped down embedded version of AMD's Hammer. It is designed for modularity, which should allow it to be easily customized for desktop use (add floating point and Altivec, up the voltage a bit, take out the embedded only stuff. It'd be hard on a normal chip, but much easier, afaik, on the 8540). Also, the Power4 would eliminate the benefit of the investment in Altivec. If Apple were going to the Power4 we would see a decrease in Altivec optimization, not the increase we're seeing (10.2 looooves Altivec). I think this rumor is BS spread by the people who can't believe that Motorola is capable of making a decent chip. Go read up on the 8540, it's quite a bit better than just "decent".

3G4N
Jul 13, 2002, 10:15 PM
Just got some juicy and semi-timely details from a Senior
Motorola Engineer over a beer last night, here in the Heart of
Motorola, DQ, and Lone Star Country. I am fairly confident of
this info. I didn't push him too hard, as he seemed as if he didn't
want to talk about it in too much detail. He didn't seem to know
much about macs, other than as a curiosity. At least he doesn't
have the same insane fever most of us reading this have.

The scoop:

-Just "tapped out" what he refered to as Apollo7 this last week.

-Said it wouldn't be 2ghz, but "more like one and a half".
Odd, I thought, not "one point four," or "one point five,"
but "one and a half"... Commented that current powermacs
use Apollo6 chips.

-Said that they had been focusing on network and embedded chips
manufacture (8540s?), thinking that Apple was a gonner.
Only now that the networking market has slowed dramatically, and they
have seen the strength (and future hollywood following - my words)
of OSX have they renewed focus on Apple, and are making a more concerted
effort to move to high-power, big-guns, processors. He commented
that the next round or two of G4s *will* be coming from IBM, but
after that Motorola will be back and swinging with some more whopper
innovations like Altivec. Apparently this major move has only come
within the last month or so.

-Now for the doozy. According to internal roadmaps, he said they
expect to hit 5 to 6 ghz (!!) by late 2003, early 2004. He said their
next chip will be called a G7!! (He was as astonished and confused about
it as I was.) I don't think I've ever seen a roadmap play out as predicted,
but the numbers along with the renewed Apple/PPC/G4 movement is
heartening news after Motorola's early lackuster performance with the G4.

How long is it, typically, between the time when a chip has
been "tapped out" and when PowerMacs ship? A month? ;;>>

Sayer
Jul 13, 2002, 10:20 PM
Just take whatever little Ryan says after getting home to his parent's basement from the movie theatre and go the absolute OPPOSITE direction to get the truth.

Notice how he is slowly swinging around to match the infamous No new G4s from ThinkSecret/CNet/PMSNBC. You get the idea.

DarkNovaMatter
Jul 13, 2002, 10:54 PM
I do admit this doesn't seem possible because of the work required to make the chip (get it to do ppc- unless it is the ppc coding). What would be fun though is the ppc arch. is based off the power architechture, what could we do if we could "upgrade" the ppc from its parent architechture? Obviously its a risc chip but can it do ppc code good? Another maybe problem is the 125 watts, thus maybe and upgrade of ppc. One thing that would nice to see would be Motorola developing new technologies and doing a limited chip production and IBM using or liscencing those techs (or jointly making them with Moto) and making most of the stack of chips.

drastik
Jul 13, 2002, 11:03 PM
well, Moto and IBM are ina na alliance, but I have another thought. Apple has been on a bit of a buying spree, maybe they are going to buy the Altivec tch form Moto, and liscence it to IBM. That would be pretty sweet, especialy if Moto has been leting Apple dangle for the last while. Htat would probably piss off apple enough to do something

j763
Jul 14, 2002, 01:21 AM
motorola suck... they can't even make phones properly...

plz steve! let's see some IBM power!

alex_ant
Jul 14, 2002, 02:06 AM
If any of this POWER-derived Apple CPU stuff turns out to be true, it was ME who predicted it first!

In this post (http://www.macrumors.com/forums/showthread.php3?s=&postid=53274#post53274)

And this post (http://www.macrumors.com/forums/showthread.php3?s=&postid=54176#post54176)

Bwahaha. C'mon Ryan Meader, you yuppie bastard, don't fail me now.

Wes
Jul 14, 2002, 04:12 AM
Does this ibm chip officially exist and is it ready to ship? If apple decided to make it from this chip how long would it take to design the hardware?

iGav
Jul 14, 2002, 05:18 AM
Originally posted by W-_-W
Does this ibm chip officially exist and is it ready to ship? If apple decided to make it from this chip how long would it take to design the hardware?

The Power4 does....... but the G5 that is rumoured to be based on it is just that at the moment.... a rumour....... :(

Something needs to be done with the PPC..... although IMHO the current G4 isn't a bad chip... so I'm not going to jump on the super-hyper-fashionable knock motorola to the ground bandwagon........ I'd much prefer that Apple, IBM and Motorola all work together to try and come up with a plan for the future of the PPC and it's uses on the next generation Apple hardware........

Blackcat
Jul 14, 2002, 06:51 AM
The Power4 Lite is due in August, so it might be doable next year.

The cheapest Power4 *server* is around $43k, but includes massive disks, gigs of very fast ram and a multiple cpus. These things sell in thousands, not hundreds of thousands like Macs, so you can expect just a cpu bought in quantities of 500k might be a lot cheaper.

However, $43,000 per unit to $700 per unit is a big drop.

mmmdreg
Jul 14, 2002, 06:53 AM
I don't care what the G5 is..i just want it and I want it NOW...

Pants
Jul 14, 2002, 07:47 AM
hmm...apple did a way good job of selling Altivec to you guys eh? ;)

I hate to be a heretic, but Im not sure why everyone is so hung up on it - yes, at the time it was a better implementation of what was around, but other manufacturers have caught up. Not only that, but its hardly been used by any software since it was introduced. I wouldnt be exactly unhappy to see it disappear and be replaced by a similar set of extensions, and a better FPU.

moJoMR
Jul 14, 2002, 07:56 AM
That's a sad comment.. what if they take a G3 kill all the cacha (ala Celeron) and name it G5? I hope you enjoy your new G5 not even capable of OSX :)

((Don't say you want something just because the name for it has been hyped..))

synergy
Jul 14, 2002, 09:13 AM
Originally posted by Pants
hmm...apple did a way good job of selling Altivec to you guys eh? ;)

I hate to be a heretic, but Im not sure why everyone is so hung up on it - yes, at the time it was a better implementation of what was around, but other manufacturers have caught up. Not only that, but its hardly been used by any software since it was introduced. I wouldnt be exactly unhappy to see it disappear and be replaced by a similar set of extensions, and a better FPU.

Very true. Apple will drop it like a hot potato if they can get their hands on something much more faster and much more powerful. I would commend them for doing it. Having something on chip to handle TCP/IP processes and more? That would be major cool

Morky
Jul 14, 2002, 09:16 AM
This is a quote from macosrumors.com posted this morning:

"Also, Apple sources have recently told Rumors that when Apple ships computers branded as PowerMac G5s next year, they will almost certainly not include Motorola PowerPC 8500 processors, which are the G5 as we know it today -- they will be IBM-built chips based on its Power4 architecture, and may even include multiple cores on a single chip. More on this after Macworld...."

Very interesting...

iGav
Jul 14, 2002, 10:05 AM
Originally posted by Pants
hmm...apple did a way good job of selling Altivec to you guys eh? ;)

I hate to be a heretic, but Im not sure why everyone is so hung up on it - yes, at the time it was a better implementation of what was around, but other manufacturers have caught up. Not only that, but its hardly been used by any software since it was introduced. I wouldnt be exactly unhappy to see it disappear and be replaced by a similar set of extensions, and a better FPU.

Well OSX makes full use of Altivec......... pretty much all the software I use is Altivec aware....... (most of the software that requires intensive processing requires it) the only people that don't really require are the ones that think demanding tasks involves emailing, word processing and having a moochy around the net......... I'd always take a new G4 over a new G3 anyday of the week.......

nickgold
Jul 14, 2002, 10:38 AM
It would certainly be a pisser if future versions of Gx chips drop Altivec support. As the above poster said, just about all intensive software makes use of the technology these days, as does Mac OS X itself. I can't see them just dropping it, after so many companies took a while to implement support into their products. I can see Apple perhaps _adding_ functionality beyond Altivec, but that doesn't mean they have to drop Altivec altogether. At least keep it around during the interim, while software companies incorporate its (possible) follow-up.

ZoodTube
Jul 14, 2002, 10:39 AM
Originally posted by 3G4N

How long is it, typically, between the time when a chip has
been "tapped out" and when PowerMacs ship? A month?

3G4N: Tape Out is a term (dating back to when mainframe computers/servers still used large tape drives) in the semiconductor industry that indicates that a final layout design has been verified and approved and has been copied to "tape" so that it can be delivered to the mask writing shop. The next step in the process is creating the mask set used for production of the wafers. Usually, it takes several months from the time of tape out to first silicon. So, I wouldn't expect working silicon to be available from this particular design for at least several months.

Marianco
Jul 14, 2002, 10:49 AM
Getting rid of Altivec will cripple the PowerPC CPU compared to the Athlon and Pentium4. Without Altivec, the PowerPC cannot compete against the Athlon's and the Pentium4's version of Altivec. Most of the Pentium4's multimedia power comes from it's own implementation of Altivec!!! The Pentium4 is known to have a crippled floating point unit.

Remeber: Having Altivec is the similar to having multiple parallel CPUs when you are processing multimedia data. A single great floating point unit cannot match up to multiple parallel floating point units - which is what Altivec is, when you are crunching video or sound or other data which can be chunked into several pieces. When you encrypt data, for example, Altivec can be 16 times faster than a floating point unit. Quicktime greatly benefits from Altivec. The software DVD player greatly benefits from Altivec.

The Power4 processor is a great server processor, probably the best for now. But it cannot do multimedia as well as a single Athlon or Pentium4. The Power4 processor is similar to having 2 PowerPC G3 CPUs with a very fast bus. With the Athlons and Pentiums evolving to include very fast data busses, for multimedia data, the Power4 is going to be outstripped.

It is better for Apple to have a G5 processor which includes Altivec. If anything, Altivec should be improved so that it includes double-precission floating point. With double-precission floating point, Altivec can then be used to greatly speed up more scientific and engineering programs - expanding the market and blowing away the Athlon and Pentium4.

GPTurismo
Jul 14, 2002, 11:11 AM
I have to agree. Dropping Altivec would be idiotic.. It's 128 processing and MOSX really needs it to run at max power :\

Also the power series chips are all corporate/enterprise level chips. To run corrctly it needs all that cache and power. To power it down is like having a ferarri with a 50 horsepower motor.

ALso, 64 bit processing is still very expensive. Itanium boxes are still in the 5 digit range, and the Itaniums 2 are gonna cost even more. When you start talking enterpriese level equipment, you are talking 15 grand at least, and unfortunately, 64 bit is enterprise level. If you want to see powermacs sore in price again, keep praying for them to drop a 64 bit chip within the next year. As we can all see, price point is the most important aspect for the GENERAL public to buy a machine. And a 5000$ bare bones mac would only be the final nail in the coffin at this time.

Brent Turbo
Jul 14, 2002, 12:05 PM
Originally posted by GPTurismo
When you start talking enterpriese level equipment, you are talking 15 grand at least, and unfortunately, 64 bit is enterprise level.

Does this mean I can sell my Nintendo 64 for $15k? Might be hard to get rid of without the rack ears....

Wry Cooter
Jul 14, 2002, 12:07 PM
Don't think of it as dropping altivec, as much as dropping Motorola, who can't seem to hit yeilds or deadlines for beans.

Also remember that IBM has its very own altivec core clone which could be completely compatible and extend the code more.

A third option- Apple could buy Altivec from Moto and have it applied to other chip designs by other manufacturers.

Altivec is used more commonly that a lot of people realize. Watched any Quicktime lately? Ripped any MP3s in iTunes? Played with iMovie? These are all much slower without altivec helping with the codecs. Altivec is used for more than simply a photoshop filter or two, and it will probably have to be included in future chips that Apple may use, in one form or another.

If it were not for altivec, Motorola may already have been out of the picture for a couple of years now. It bought them time to become a more conscientious vendor for Apple; who is to blame if they haven't been able to put that time to better use towards this end?

gbojim
Jul 14, 2002, 12:15 PM
Also remember that IBM has its very own altivec core clone which could be completely compatible and extend the code more.



Do you have a link to any info on this?

Pants
Jul 14, 2002, 12:32 PM
Originally posted by iGAV


Well OSX makes full use of Altivec......... pretty much all the software I use is Altivec aware....... (most of the software that requires intensive processing requires it) the only people that don't really require are the ones that think demanding tasks involves emailing, word processing and having a moochy around the net......... I'd always take a new G4 over a new G3 anyday of the week.......

well it does, but thats more a reflection of the (sluggish) state of the g4 than a ringing endorsement for Altivec. Photoshop uses it yes, but coding for it is a pain - few compilers are even aware it exists, let alone be able to optimise code for it. an example - anyone remember iD's 'fun' with codewarrior and altivec? Yes, you and i would take a g4 over a g3, but again, is that a ringing recommendation or like saying "Id rather a cold than measles"? ;)

If anyone here thinks that Altivec is the only SIMD implementation around they are very much mistaken - 3dnow? MAJC? even the playstation2 has one! . As for the g4's '128 bit' processing....ahem....hardly - if it were, why bother with a g5? ;)

Silver Dragon
Jul 14, 2002, 12:40 PM
Couple of posters have hit this right on.... But let's clarify a few points.

1 - a 64 bit processor does *not* mean that it's faster. It means that the processor can address more memory. In most cases a 64 bit processor will slow down 32 bit applications; however, it looks like the 8500 series will actually be able to run 32 bit apps at full speed. THAT's what is so cool about the 8500. Just because 64 is bigger than 32 does not mean that it's bigger and better.

2 - Altivec is actually pretty cool, but it also ties us to Motorola. Motorola can come up with some very cool concepts and designs, they just can't seem to produce these designs in bulk. Remember how we were stuck at 500MHz forever and a day? I blame that on Motorola. Now let's talk about IBM. IBM has some *very* cool production facilities and has taken the PowerPC architecture to the extreme :::echo, echo, echo:::. It would be a beautiful day if IBM took over the PowerPC fulfillment for Apple instead of Motorola... What would be even nicer is if IBM and Motorola had rights to produce all chips for Apple. Now I'm just dreaming. Point is, Altivec is cool, Motorola can't produce the chips (in the quantities that we demand) with Altivec on it. IBM can produce the chips, but then we loose Altivec. Perfect world, Apple buys Altivec and allows both Motorola and IBM to produce chips. This would re-create the alliance that the PowerPC was supposed to bring. Sorry, I rambled here.

3 - We will not see the Power4 in a consumer Macintosh. That chip is not designed for consumer computers. It's a server chip plain and simple. We *might* see Power4 technology in a Macintosh, but not the chip itself.

4 - G5 is just a name. What would everyone do if Apple decided to skip Generation 5 and just go to Generation 6? Would we say "WOW, this is so much better than the G5!!!" or would everyone realize that it's Marketing. Just because it's a G5 does ***not*** mean it's better.

5 - I believe that old Steve has a trick up his sleeve. See, the X86 architecture is (in my opinion) about to hit a brick wall. There's only so many times you can build upon a bad foundation before the whole house falls. That is when Apple will step in with the Uber processor. His Steveness (as per AtAT) has eluded to having a processor plan many times in financial calls. He has said not to worry, Apple has it under control. Personally, I trust that he knows more about upcoming products than I do, so I trust him.

Anyhow, that's my $3.50. Some of that info may be right, other info wrong. It happens. Hope it made sense and I didn't ramble on too long.

-SD

type_r503
Jul 14, 2002, 01:11 PM
What may interest some of you is that Motorola was inactive for several days during the fourth of july week (trying to save money). The G4 is (let's not kid ourselfs) lagging behind in perceived speed to the AMD/Intel camp. How could apple expect someone to "Switch" from a +3Ghz to a 1.4 or 1.5Ghz. Intel has caught up to the G4 regardless of Altivec, Apple is tired of dealling with Motorola and so would I. Remember 500Mhz?


Type R503

gbojim
Jul 14, 2002, 01:16 PM
Originally posted by Pants


well it does, but thats more a reflection of the (sluggish) state of the g4 than a ringing endorsement for Altivec. ... If anyone here thinks that Altivec is the only SIMD implementation around they are very much mistaken - 3dnow? MAJC? even the playstation2 has one! . As for the g4's '128 bit' processing....ahem....hardly - if it were, why bother with a g5? ;)

The use of Altivec is absolutely NOT a reflection of the sluggish state of the G4. When AIM decided on the PPC architecture, they could have added a bunch of pipeline stages like Intel and jacked the clock rate, or they could opt for a different method to achieve performance which is the path they took - in this case adding vector processing, cache, etc. Apple was a big part of those decisions and is using Altivec because that was the direction they decided to go. With the differential in clock rate, a PPC CPU all by itself will never match Intel or AMD in integer performance. You may not agree with the decisions they made, so be it.

Sure there are lots of SIMD implementations. However, Altivec is the only generic one around that is being used in a desktop computer. All of the others are really designed to enhance multimedia. It is very difficult to do anything else with them.

Catfish_Man
Jul 14, 2002, 02:04 PM
...does the same things as SSE, MMX, 3dNOW, SSE2, etc... but it does them much better, and does some things that they don't. The Pentium4 has 1 pipeline for SIMD and it's shared with the FPU. The G4+ has 4 dedicated pipelines. Altivec is the way SIMD should be. Now if only the rest of the G4 were that advanced... (specifically, I want DDR and a .13 micron manufacturing process).

To the people who constantly whine about how nothing uses Altivec: How do you know?

alex_ant
Jul 14, 2002, 02:04 PM
Originally posted by Silver Dragon
5 - I believe that old Steve has a trick up his sleeve. See, the X86 architecture is (in my opinion) about to hit a brick wall. There's only so many times you can build upon a bad foundation before the whole house falls.
Can you elaborate on this? I would love for x86 to hit a brick wall too, but haven't x86 chips evolved so substantially from their earlier days as to be fundamentally different architecturally today?
That is when Apple will step in with the Uber processor. His Steveness (as per AtAT) has eluded to having a processor plan many times in financial calls.
Links, links I want links. :)

Alex

Silver Dragon
Jul 14, 2002, 02:07 PM
Originally posted by Pants
As for the g4's '128 bit' processing....ahem....hardly - if it were, why bother with a g5? ;)

Actually, the G4 does have 128 bit processing. I believe (and someone please correct me if I'm wrong) it's as follows:

CPU: 32bit
FPU: 64 bit
Altivec: 128bit

I *think* that's right, but I can't remember. Someone smarter than I feel free to correct the above statement.

As I have stated time, and time, and time again... 64 bit does not mean it's faster. In most cases if you have a 64 bit CPU, your 32 bit applications will run slower. Let me say that again. If you have a 64 bit CPU, in most cases your 32 bit applications will run slower. The nice thing about Motorola's 8500 series... It's a 64 bit processor that allows 32 bit applications to run without a speed hit. The advantage? The processor can address much more memory without slowing down your current apps. Twice as fast? Nope. Will you notice a speed difference? Probably not (based on equal processors, but one being 64 bit instead of 32 bit.) Let's get off this 32 vs 64 thing and focus on what would really increase performance!!!

Just my $1.75.

-SD

Silver Dragon
Jul 14, 2002, 02:18 PM
Originally posted by alex_ant

Can you elaborate on this? I would love for x86 to hit a brick wall too, but haven't x86 chips evolved so substantially from their earlier days as to be fundamentally different architecturally today?

Links, links I want links. :)

Alex [/B]

Simple...
x86, has had a long life. It is now to the point where they are pulling any rabbit out af the hat they can to make those chips seem faster. Are they faster? Yeah, a wee bit, but not much. More and more it's a bigger number with no performance gain. So I look at it like the chips have hit the end of their life. Notice how AMD and Intel are not churning out as many speed bumps as before? There are a couple of reasons for this, but one of them is the inability to do what they want the chips to do (I must applaud the engineers though, I never thought they would get as far as they did... ) The chips have changed... a lot. The problem is that the end result has to be a x86 based instruction. You can translate it as many times as you want, Windows only understands X86 (well, kinda. Theer was that PPC version of Windows, and then there is SSE and such... But we will ignore that for now.) Again, this is opinion and I may be wrong.

As for Steve having a trick up his sleeve... Listen to the old financial calls from Apple. Go to the end where there is a Q&A session. This happens on a few of them. Someone will ask about processor speed, and his Steveness will always come back with "Don't worry, we have a plan in place. It's under control. Just wait for what we have to offer down the road". Not a direct quote of course, but you get the idea. Not sure where to find the financial results anymore (QuickTime stream) but search Apple's site. Go back about year and start listning. I *highly* suggest everyone listen to these calls live when they happen. You get more insite into Apple in that hour than with all the rumor sites combined... At least I do :)

Remember, this is all opinion. I don't claim to have a crystal ball nor am I saying that any of this is fact. I'm just listning to what Apple has to say and using commom sense.

-SD

Marianco
Jul 14, 2002, 03:32 PM
In comparing the G4, Speculated G5, Power4, Athlon, and Pentium 4, processors, there are three factors which determine speed of the CPU: 1) SIMD/Altivec, 2) Pipeline length, and 3) Bus speed. The G4, Athlon, and Pentium 4 all have implementations of SIMD (which is Altivec on the G4). The Pentium 4 has the weakest floating point unit, the Athlon the strongest floating point unit. The Pentium 4 has the longest pipelines, the G4 the shortest. Pipelines allow you to work very fast serially with data, while SIMD units allow you to work very fast with data in parallel fashion. Pipelines allow you to speed up the clockrate of the CPU. The Power4 is the only CPU with a super speedy bus speed. It is capable of transferring 25 gigabytes of data a second. The other CPUs are limited to single gigs of data a second.

The limitation of SIMD units is that your data must be able to be chunked up in pieces to be useable and worked on in parallel. Data which are great for SIMD units include video, sound, data to be encrypted, genetics data. For these, an SIMD unit is like having several CPUs working in parallel. If the data cannot be chunked, then you have to use the CPU's integer or floating point units. The Pentium 4 is slower than a Pentium 3 at the same clockspeed when it's SIMD unit is not used.

The strength of a long pipeline is that as you crank up the speed of the CPU, the speed of processing data through the integer and floating point units increases as well. The limitation of a pipeline is that if the program you are working with makes frequent decision changes, such as with random data, you have to reload the pipeline from memory, wasting a lot of time.

The weakness of using the G4 is that many programs rely on serial processing of data (i.e. work using the integer or floating point unit), not parallel processing. This includes database work, word processing, games, etc. These programs work faster with a higher speed CPU, such as the Pentium 4. In games, the Graphics Processing Unit does the heavy SIMD work, relying mainly on the integer unit on the CPU to get things done.

The speculated G5 (or G4++ CPU) has a longer pipeline. This allows it to have both SIMD and a higher clockspeed. The implementation of DDR and faster bus (such as Rapid-IO or Hypertransport) will further improve on the three factors in CPU design, giving the G5 a greater ability to surpass the Pentium 4.

If you are waiting for the Pentium 4 to hit a wall - you will be waiting a long time. People have made this prediction for years. What Intel has done is to improve on the *86 architecture so that it has RISC features, has SIMD, longer pipelines, and greater bus speed. As it rose past 2.5 GHz in speed, it outstripped the Athlon (which has a slow bus speed and shorter pipelines, though better SIMD and Floating point unit).

If more pipelines and a better bus can be added to the Athlon, it would be a monster chip. Already, it is so much faster in Photoshop than the G4 (because the Athlon has a much stronger floating point unit, and a nice SIMD unit), that Steve does not compare a dual G4 to a single Athlon. The single 1.7 GHz Athlon beats a dual 1 GHz G4 in Photoshop, and is cheap to boot.

Once more pipelines (to allow higher clockrates) and a better bus and a better floating point unit are added to the G4, it will then beat the Athlon.

This leaves one matter: multiple CPUs. With Mac OS X capable of supporting multiple CPUs, the two easiest ways to improve Mac speeds are to have Quad G4 Mac with faster busses.

beatle888
Jul 14, 2002, 03:41 PM
finally we have an intelligent soul on board.
One that doesnt speak in a condescending tone.

SEE ALL YOU BRATY TEENS OR ANGRY OLD PEOPLE!
TAKE A LESSON YOUR MOMMY DIDNT TEACH YOU,
HOW TO BE HUMAN.

I like how Silver Dragon uses the disclaimer at
the end. He's obviously aware of you little twits.

sorry, just tired of all the little bitches on the site
:p

pgwalsh
Jul 14, 2002, 03:46 PM
Originally posted by Silver Dragon

Theer was that PPC version of Windows, and then there is SSE and such... But we will ignore that for now.)

I have a version of NT that supports ppc. However, I'd never to that to the ppc.

Marianco
Jul 14, 2002, 03:46 PM
The Power4 is a great server processor. It has a very fast bus to push data through. It has two processor cores (essentially two G3s) to work on the data in parallel. It does the job of pushing data very well. Having two processor cores allows you to have redundancy if one fails.

A weakness of the Power4 is that if your data cannot be worked on in parallel, then one of the processor cores is not used (half of the CPU is wasted), reducing it to a 1.2 GHz G3 processor, essentially with a fat data bus.

Another weakness of the Power4 is that if you try to use it for data which can be processed in parallel - such as multimedia, video, movies, sound, data for encryption, genetics data, bit-mapped data (such as used by Photoshop), Altivec blasts the Power4 to bits at the same clockrate. If the data is chunked in 8 bit lengths, then Altivec is the same has having 16 parallel processing units, whereas the Power4 will have only two. For this type of data, Altivec will be 8 times faster than the Power4's two processing cores.

Note that the G4 is already instruction compatible with the Power4. Thus it is really no biggie to claim that the G5 is based on the Power4. In fact, you could say that the Power4 is two G3s with an extremely fast data/instruction bus. Give the G4 an extremely fast data/instruction bus and it will destroy the Power4 when using the Altivec unit.

Sun Baked
Jul 14, 2002, 04:45 PM
With IBM you can almost expect that a dual G5 will reside on a single chunk of silicon using a large combined cache.

Going from the 603/604 to the G3/G4 was supposed to make multiple CPUs on a single chip a reality, but they never made it - just multiple processing units (integer, fpu, altivec) for a single CPU.

The really interesting choice should be what type of bus Apple decides on for the next gen of CPU. I'd almost expect something safe like Hypertransport, but Apple could leapfrog to something much faster.

jadam
Jul 14, 2002, 05:32 PM
silver dragon... your very wrong. 64 bit code IS NOT SLOWER than 32bit code. 32bit Code BEING EMULATED on a 64bit IS SLOWER. Hence the Itanium which has to emulate 32bit x86 instructions on its 64bit IA-64 ISA. The Power4 runs PPC code Natively, hence it runs it VERY FAST. 64bit code IS slower if you have a slow bus.

The G3 used in the iBook 600/700 is .13u BTW. And IBM has a .1u plant in FishKilsl NY. The Power4's use 125watts because they are 2-xx cores on one chip.

Yes, and Altivec unit WOULD murder a two core Power4 processor in multimedia, but if... this dual Core Power4 had TWO 256bit AltiVec units, which a next gen chip SHOULD have, then ... well, you know the rest :).

Ohh yeah, Silver Dragon- the moto 8500's are not as good as the Power4's.

And Intel WILL NOT hit a brick wall, i mean seriously, the PIV's can have a 1.2ghz bus, probally even more, they are designed to scale up to 10ghz. probally more. And the PIV i think take up like 75 watts too. so, if you take out 30mb of L2 cache and 120mb of L3 cache from that 125watt Power4 processor then... Plus remember, they are designed to stay on for a long time, and they need more then enough juice just in case.

AND Sun sells $900 64bit Sparc stations OK

Silver Dragon
Jul 14, 2002, 05:47 PM
jadam, I don't think I was as clear as I should have been. Please let me clarify.

I did not say that 64 bit code was slower. I said that 32 bit code on a 64 bit processor is slower. However, the 8500 changes this as 32 bit code runs just as fast as 64 bit code. That's really, in my opinion, one of the only cool things about the 8500 series. I did not really elaborate and made a few assumptions about the reader. My fault.

64 bit still does not mean it's faster. It can just allocate larger chunks of memory. Please correct me if I'm wrong. 64bit code on a 64bit process will take no speed hit, agreed. I don't believe I ever said it would, but if I implied that I apologize.

As I said about the x86 line, that's just my opinion. I have nothing to back it up, nor do I claim to have insider info. I'm just looking at what Intel and AMD were putting out vs what they are putting out today. I'm also seeing more and more problems crop up and more and more people actually looking at performance. If that's the case, they may have a problem. Again, in my opinion.

And finally, I don't remember saying that the 8500 was better than the Power4. I did say that the Power4 does not belong in consumer computers as it stands right now, which I believe to be true. Take the cool parts of the Power4 that belong in consumerville add them to a new IBM processor, call it the G5 and I'm happy. If I gave the impression that I want the 8500 in my next Mac, I'm *very* sorry. I think it's cool, but I would much prefer go back to IBM for the production of PPC chips (please note my signature.)

jadam, I think we agree on most points, but I don't think I was clear on my intent. Accept my apologies, and try re-reading my posts with the idea that I want IBM, don't really like Motorola, and think that too many people are hyped over nothing over 64 bit when they all own 32 bit applications.

Just my $2.25. Hope that clears some things up.

-SD

dongmin
Jul 14, 2002, 05:56 PM
Originally posted by Silver Dragon

x86, has had a long life. It is now to the point where they are pulling any rabbit out af the hat they can to make those chips seem faster. Are they faster? Yeah, a wee bit, but not much. More and more it's a bigger number with no performance gain. So I look at it like the chips have hit the end of their life. Notice how AMD and Intel are not churning out as many speed bumps as before?

I don't know how you can say this. AMD's processors outperform the G4 in every aspect. It used to be before, the G4 had advantages in certain areas like running Photoshop. But now... If anything, the performance gap seems to be growing, and AMD seems to have a more promising immediate future.

And I'm talking about real-world performance here, not the gigahertz gap which is a whole another animal...


As for Steve having a trick up his sleeve... Listen to the old financial calls from Apple. Go to the end where there is a Q&A session. This happens on a few of them. Someone will ask about processor speed, and his Steveness will always come back with "Don't worry, we have a plan in place. It's under control. Just wait for what we have to offer down the road".

Don't you think this is a bit of wishful thinking on your part? It's been 3+ years since we had the 500 mhz hurdle, and the PowerPC is STILL lagging in development. We can't take Steve's words at face value; he is the master of the Reality Distortion Field after all. I want to see some concrete proof that the PowerPC is going places (Moto's optimistic powerpc roadmap doesn't count).

I love my Mac. I don't think I'll ever go over to the dark side. I love the design, the OS, the community, and the ambition to be original and innovate. But I don't have any illusions about it being the most advanced or fastest hardware out there.

Marianco
Jul 14, 2002, 05:57 PM
Server CPUs like the Power4 and Pentium4 Xeon chips are power hungry and very expensive. The Power4 itself costs about $2500. Try buying a Dell Server. The single CPU server seems reasonable. But adding another Pentium Xeon CPU to create a dual CPU server will cost you another $2000. These CPUs clearly are not consumer models. The latest AMD Opteron server procesors are likewise going to be very expensive.

alex_ant
Jul 14, 2002, 06:48 PM
Originally posted by Marianco
Server CPUs like the Power4 and Pentium4 Xeon chips are power hungry and very expensive. The Power4 itself costs about $2500. Try buying a Dell Server. The single CPU server seems reasonable. But adding another Pentium Xeon CPU to create a dual CPU server will cost you another $2000. These CPUs clearly are not consumer models. The latest AMD Opteron server procesors are likewise going to be very expensive.
Server chips/systems have massive margins. For instance, in 1998, SGI sold MIPS R5000 CPU modules for its O2 at a price in excess of $2000, when the R5K itself cost $50 per unit in quantity. It's fun to look at the price lists of IBM/Sun/SGI/HP etc. "24X CD-ROM drive: $399. 1GB parity SDRAM: $999. 36.4GB Ultra160 hard drive w/ proprietary mounting sled: $1299."

The big Unix players need high margins to fund their R&D. Intel sets the Xeon prices high because they can get away with it - not because the Xeon costs anywhere near as much to produce as what it sells for. So, I'm sure that with a scaled-down and redesigned Power4-derived CPU put in place inside a consumer item like a Macintosh, the price would be able to come way, way down.

Alex

T'hain Esh Kelch
Jul 14, 2002, 06:51 PM
Originally posted by Sun Baked
The really interesting choice should be what type of bus Apple decides on for the next gen of CPU. I'd almost expect something safe like Hypertransport, but Apple could leapfrog to something much faster.
Back to the real world.. 166 mhz DDR...

Chryx
Jul 14, 2002, 07:29 PM
Originally posted by T'hain Esh Kelch

Back to the real world.. 166 mhz DDR...

FYI, AMD's Hammer processors use hypertransport for AGP/PCI <> processor links.. and guess what kind of memory they use?.. 166Mhz DDR Sdram....

type_r503
Jul 14, 2002, 07:57 PM
1. For the Idiot that wrote Altivec is the same as MMX. The are not even in the same boat. MMX which means Multimedia Instructions puts common MM tasks into silicon. Altivec is a 128-bit wide number cruncher. Intel has nothing like altivec and probably don't care to ever implement it. Their "speed" is from super high clock rates on the CPU.

2. Motorola is hurting bad. This rumor might indicate that IBM will be purchasing the PPC unit from MOT which should have happened long ago.

3. For altivec to even be used software companies must optimize their code to take advantage of the features. OSX is not optimized for the the altivec engine.

4. A 64bit bus is not just for memory addresses but also for data lines. This allows for 64bits of data to enter the CPU on a given clock cycle. The G4 only allows for 32bits. Once the data is in the CPU anything can happen. This greatly reduces the bottle neck caused by the memory bus.

type r503

Chryx
Jul 14, 2002, 08:06 PM
Originally posted by type_r503
1. For the Idiot that wrote Altivec is the same as MMX. The are not even in the same boat. MMX which means Multimedia Instructions puts common MM tasks into silicon. Altivec is a 128-bit wide number cruncher. Intel has nothing like altivec and probably don't care to ever implement it. Their "speed" is from super high clock rates on the CPU.

Guess what SSE2 is...
a double precision int/fp number cruncher.....

3. For altivec to even be used software companies must optimize their code to take advantage of the features. OSX is not optimized for the the altivec engine.

I was under the impression that a lot of OSX code IS Altivec?

gbojim
Jul 14, 2002, 08:38 PM
Originally posted by type_r503
1. For the Idiot that wrote Altivec is the same as MMX. The are not even in the same boat. MMX which means Multimedia Instructions puts common MM tasks into silicon. Altivec is a 128-bit wide number cruncher. Intel has nothing like altivec and probably don't care to ever implement it. Their "speed" is from super high clock rates on the CPU.

2. Motorola is hurting bad. This rumor might indicate that IBM will be purchasing the PPC unit from MOT which should have happened long ago.

3. For altivec to even be used software companies must optimize their code to take advantage of the features. OSX is not optimized for the the altivec engine.

4. A 64bit bus is not just for memory addresses but also for data lines. This allows for 64bits of data to enter the CPU on a given clock cycle. The G4 only allows for 32bits. Once the data is in the CPU anything can happen. This greatly reduces the bottle neck caused by the memory bus.

type r503

Just a couple of corrections...

The common belief is MMX stands for MultiMedia eXtensions. However, MMX officially does not stand for anything. Info courtesy of Intel docs. BTW, the newest incarnation of MMX known as SSE2 is closer in concept to Altivec but cannot hold a candle to it for performance.

OS X - especially with the upcoming release of 10.2 is heavily optimized for Altivec.

The G4 already reads 64 bits of external data per clock. Internally it processes 32 bit integer, 64 bit FP and 128 bit vector quantities per clock.

Chryx
Jul 14, 2002, 09:05 PM
g{The common belief is MMX stands for MultiMedia eXtensions. However, MMX officially does not stand for anything. Info courtesy of Intel docs. BTW, the newest incarnation of MMX known as SSE2 is closer in concept to Altivec but cannot hold a candle to it for performance.}g


SSE2 has nothing to do with MMX, totally seperate, MMX was integer only and couldn't be run in parallel with the floating point unit for starters... (modern x86 chips do still support it, although I believe the Athlon can still use the FP hardware whilst running MMX code?)

SSE2 OTOH, can do (IIRC) 2 double precision FP ops per cycle. (or 4 single precision)

then you have to factor in the bus/clockspeed advantage the P4 holds over the G4

Altivec is VERY nice, but Intel have something of a brute force advantage ATM.

gbojim
Jul 14, 2002, 09:21 PM
SSE2 has nothing to do with MMX, totally seperate...

That is not exactly true. SSE2 is quite different than MMX which I stated in my earlier post when I said the design is much closer in concept to Altivec. However, all Intel SIMD processors including SSE2 derived from MMX, are fully backward compatible to MMX and the design intent remains the same.

Chryx
Jul 14, 2002, 09:24 PM
Originally posted by gbojim


That is not exactly true. SSE2 is quite different than MMX which I stated in my earlier post when I said the design is much closer in concept to Altivec. However, all Intel SIMD processors including SSE2 derived from MMX, are fully backward compatible to MMX and the design intent remains the same.

I tend to look at SSE2 as a superset of MMX (and at the functions it shares it's much faster at than MMX was) I seem to have a vague recollection of the MMX and SSE2 execution hardware being seperate on the P4 though?.. might just be my frazzled brain though..


as for design intent?.. well, yeah.. making repetetive (instructionally) tasks faster was the design intent, so yeah :)

Marianco
Jul 14, 2002, 09:37 PM
The G4 Altivec parallel processing unit and the Pentium 4 Deep Pipelines are both brute force methods of processing data. The Altivec unit processes chunks of data in parallel, the Pentium 4 processes chunks of data serially but at high speed.

The Pentium 4 copied the G4 by adding an SIMD/Altivec-like unit. It one-upped the G4 by having a better bus to transfer data and instructions. It then upped the G4 by having deeper pipelines to process the data at higher speed.

The lower clockspeed G4 can keep up with the Pentium 4 because the Altivec unit is so much better than the Pentium 4's SIMD unit and because complex instructions to the CPU cause frequent dumps of the pipeline contents, slowing down the advantage of the long pipeline. However, the higher clockspeed of the Pentium 4, when coupled with less complex intructions (i.e. few branchpoints), allow the Pentium 4's Integer units and Floating point units to seriously beat the G4 in common tasks and games. Most common programs such as databases or word processors or games do not use the CPU's SIMD unit. Games use the graphic processors SIMD units instead.

Thus the G4 is at a disadvantage running at lower clockspeeds, for the many programs which need only integer processing. This alone makes me want to get a fast Pentium 4 computer to run Filemaker Pro 6.

The latest G4+ incarnations do add more pipeline depth, thus bringing more balance into the equation. Hopefully Apple unveils the 10 deep pipeline G4+, with a much faster bus (such as hypertransport, or whatever). This should allow the G4 to match or overrun the Pentium 4, maybe even the Athlon. However, the Athlon chip has deep pipelines, a great floating point unit, and and an SIMD unit which can match the G4. It's bottleneck is it's fairly slow bus.

Chryx
Jul 14, 2002, 09:40 PM
Originally posted by Marianco
However, the Athlon chip has deep pipelines, a great floating point unit, and and an SIMD unit which can match the G4. It's bottleneck is it's fairly slow bus.

The AthlonXP already has a much better bus than the G4 though (although it can't use L3 cache)

133Mhz DDR Is significantly less of a bottleneck than 133Mhz SDR :)

(Clawhammer should be frickin' impressive though)

PyroTurtle
Jul 14, 2002, 09:56 PM
i still find it funny as to how much mac users seem to know (or, at least think they know as they themselves say) about the PC world...
i can sum all these treads up...

Better performance now, and give us those chips in high quantity, and very cheap! and preferably not from moto...maybe keep altivec if it really does work...

did i miss anything?

Chryx
Jul 14, 2002, 09:58 PM
Originally posted by PyroTurtle
i still find it funny as to how much mac users seem to know (or, at least think they know as they themselves say) about the PC world...
i can sum all these treads up...

Better performance now, and give us those chips in high quantity, and very cheap! and preferably not from moto...maybe keep altivec if it really does work...

did i miss anything?

I think you're supposed to say something snide about using PC133 on the top of the line machines in the year 2002? :p

DavidRavenMoon
Jul 14, 2002, 11:58 PM
Originally posted by T'hain Esh Kelch

Back to the real world.. 166 mhz DDR...
That would be Xserve :) Actually it has 266MHz PC2100 DDR SDRAM

I'd imagine we will see DDR in the new G4s.. when ever they come out.

DavidRavenMoon
Jul 15, 2002, 12:09 AM
Originally posted by Chryx
Guess what SSE2 is...
a double precision int/fp number cruncher.....


Except as in the case of the P-IV, it can't do FP and vector ops [SSE] at the same time, since the one unit handles both.

The G4 has a double precision FP unit, and four AltiVec vector processors. Altivec doesn't do double precision FP however. The "G5" is supposed to fix that.

BTW, I have read several times that IBM did indeed licence AltiVec from Moto, but just doesn't seem to be interested in using it. But I guess they could if Apple wanted them too.

Also someone was wondering... yes the POWER4 is a PowerPC chip.

Chryx
Jul 15, 2002, 12:10 AM
Originally posted by DavidRavenMoon

That would be Xserve :) Actually it has 266MHz PC2100 DDR SDRAM


Actually, it has 133Mhz DDR PC2100 SDRAM, the 166Mhz DDR he was referring to would be PC2700

the "266Mhz" is very much a marchitecture term >:(

DavidRavenMoon
Jul 15, 2002, 12:27 AM
Originally posted by Chryx


Actually, it has 133Mhz DDR PC2100 SDRAM, the 166Mhz DDR he was referring to would be PC2700

the "266Mhz" is very much a marchitecture term >:(

I copy and pasted that off the Apple Xserve page:

Memory
*256MB or 512MB of 266MHz PC2100 DDR SDRAM with up to 2.1GB/s throughput
*Four DIMM slots supporting up to 2GB of DDR SDRAM using the following:
128MB or 256MB DIMMs (64-bit-wide, 128Mb technology)
512MB DIMMs (64-bit-wide, 256Mb technology)

http://www.apple.com/xserve/specs.html

If you are clocking the chip at 133Mhz, and it's reading on the rise and fall of the clock cycle, that would be 266DDR (Double Data Rate). 133 X 2.

PC2700 is ~166MHz (333DDR)

" PC2100 DDR is designed for use in systems and motherboards which require a 133MHz front-side bus. While the system or motherboard may operate a 133MHz front-side bus, its effective front-side bus speed is 266MHz in DDR operation. To get the effective front-side bus speed of a DDR system, double the operating frequency which in this case is 133 MHz to get the 266 MHz.

(PC2100 is also backward compatible for systems with a 100MHz front-side bus.)

PC2700 DDR is designed for use in systems and motherboards which require a 167MHz front-side bus, with an effective front-side bus speed of 333MHz. "

http://www.crucial.com/crucial/pvtcontent/topdef.asp?device=motherboard_extended

Chryx
Jul 15, 2002, 12:42 AM
Originally posted by DavidRavenMoon

PC2700 DDR is designed for use in systems and motherboards which require a 167MHz front-side bus, with an effective front-side bus speed of 333MHz. "


That's my point exactly, "effective 333Mhz" isn't the same thing as "actual 333Mhz"

and you can't buy "Actual 266Mhz" DDR ram (well, not for system memory, it's available for videocards)


Like I said, it's marchitecture, it isn't actually running at 266mhz, but they claim it is because it "sorta is kinda".. bah

TMay
Jul 15, 2002, 12:51 AM
One thing that I would bet on. Apple won't go with a different architecture than PPC.

two reasons:

1) PPC architecture is scalable from imbedded, low power consumer devices to PowerX server processors (or massively multiprocessor supercomputer architectures). There is no other architecture in the world that will do that. This portends OSX derived consumer devices, which will be a further boon to developers. Read this anyway that you want, but I wouldn't rule out a devices with PDA functionality.

http://www-3.ibm.com/chips/products/powerpc/newsletter/

2) It is cheaper for Apple to buy out Moto's desktop PPC line and develop the design in house and send it to a foundry (even AMD) than to switch. Either way, I believe that Moto's interest is considerably higher now that the economy sucks. Motorola also has some significant GaS manufacturing technology that will come on line in 2005 that may scale to 70Ghz. Moreover, the 8500 is substantially the same manufacturing technology as the 7500 (G5), so there really isn't a reason that we won't see a G5 from Moto. And of course, IBM would be interested, though they still aren't all that keen on Altivec.

My own opinion is that G5 development is advancing quite well, but the impetus for Apple to market such a device today (or even January) has been killed by the state of the economy. Moto will naturally be encouraged by Apple to keep these developments quiet.

Note that Apple has had to put in substantial resources to OSX so the hardware designs have had to take a back seat. With the immenent release of Jaguar, I believe that more resources will be available to hardware and consumer device development.

Marianco
Jul 15, 2002, 01:49 AM
I believe that there is a huge amount of pressure on Apple to come up with faster computers. Thus there is a lot of pressure to come up with the G5.

Apple, remember, has been gobbling up high end video firms and recently eMagic. THere was also that closed door meeting with the Hollywood big-wigs, where Steve asked the leaders what will it take to make the switch to Mac. He was told, it is rumored, that power was king. Dual and Quad CPU PowerMacs were necessary. The server is one result of that meeting. XServe can be used in render farms.

Marianco
Jul 15, 2002, 02:05 AM
Although IBM has been traditionally a server company and thus has not been interested in Altivec because it isn't needed when it comes to pushing data with server CPUs like the Power 4, it should definitely be eyeing the addition of an SIMD unit like Altivec on future CPUs.

What IBM missed is that SIMD units spectacularly speed up scientific programming tasks. Witness how the Dual 1 GHz G4 can do gene sequence searches 5 times faster using the program, Blast, then using Blast on a Pentium 4. Witness how much faster a 1 GHz G4 is in searching for encryption keys than the fastest Athlon or Pentium 4. Since IBM also makes massively parallel super computers for scientific research, it is shooting itself in the foot when it tries to use the Power 4 CPU in them compared to a super computer with multiple P4s or Athlons with their SIMD units. Again, the Power 4 is great at pushing data and doing server work (which is the bulk of what IBM makes profits on), but for scientific work (the other area where servers are also used), it acts like two G3s with a very fast bus. The AMD Hammers and Pentium 4 Xeons will nail the Power 4 on these tasks.

In any case, there are only two more days left before Apple unveils their next computers. woo hooo!

All this speculation about the future means nothing since you can't use it today.

I'll buy the latest greatest when Apple unveils them. When the G5 comes out, I'll buy that one too. I'll also buy a Pentium 4 computer since they are cheap enough from Wal-Mart - without Windows installed, to use as a second computer or peripheral to the PowerMac. Accounting software, unfortunately, is more available on the Windoze side. But then again, I do cross platform development. The AMD Athlon is great, but not entirely compatible with Windows XP like the Intel chips. Also, the P4's clockspeed has risen high enough to outstrip the Athlon in most things. Don't underestimate Intel. The Pentium 4 will keep getting faster even if it has to copy features of the G4 or other CPUs.

Chryx
Jul 15, 2002, 02:16 AM
Originally posted by Marianco
The AMD Athlon is great, but not entirely compatible with Windows XP like the Intel chips.

Do you have any evidence to substantiate this or are you just spreading FUD? (or blaming AMD for VIA's inability to make chipsets that work?)

tjwett
Jul 15, 2002, 02:36 AM
Originally posted by Megaquad
whats with all those programers who were working really hard to optimize apps for altivec? especially big players like adobe etc.

by the looks of things most developers have not been putting too much effort into AltiVec. in the begining they were but now everyone just seems to be Carbonising. perhaps it's because they know the new processors won't need it.

Marianco
Jul 15, 2002, 07:15 AM
Athlon-based motherboards are not totally compatible with certain hardware and software packages like the Intel Pentium-based motherboards.

For example, I do a lot of videoconferencing. The videoconferencing hardware/software for a 3-ISDN line 380 kbps setup I have specifically excludes Athlon compatibility. I thus have to use a Pentium-based motherboard, despite how great the Athlon is.

Yes, perhaps it is the fault of the motherboard makers, but for this niche, the hardware surrounding the Athlon is different enough to make a big difference.

DavidRavenMoon
Jul 15, 2002, 09:01 AM
Originally posted by Chryx


That's my point exactly, "effective 333Mhz" isn't the same thing as "actual 333Mhz"

and you can't buy "Actual 266Mhz" DDR ram (well, not for system memory, it's available for videocards)


Like I said, it's marchitecture, it isn't actually running at 266mhz, but they claim it is because it "sorta is kinda".. bah

But actually it is running twice as fast as SD Ram, because it's running on both sides of the clock cycle. (Think of a square wave and you will see what that means.)

So for every tic of the clock it's advancing two tics. SD (Single Data) Ram only operates on either the rising or falling (I dont remember which), so PC133 is running at 133MHz, but DDR clocked at 133 is running at 266Mhz. The bus is 133Mhz, but it's reading twice as much data per clock cycle.

266MHz DDR would be 532DDR.

It's not just marketing. Bottom line is it is faster :)

GPTurismo
Jul 15, 2002, 09:06 AM
Well, processor speed = gimmick and marketting ploy. The main reason you need such powerful processors is because of the bloat. 95% of people can get by with a 500 megahertz p3 running win2k. Then people bring up games and yet again, the only area needing upgrading is graphic cards. The actual code is usually trash due to being built in Direct X.

As for the processor speed thing, it's an old argument and it seems idiots are still going to fall for it. Processor speed doesn't matter. Performance is what matters.

Also, the BIGGEST deciding factor got Joe Shopper is price, something even now Macs are still having a hard time beating.

I can get a 1.8 gig p4, 40 gig hdd, with a 17 inch crt, gf2, and 256 ram for 699. Sure it's a pos, but guess what, as long as it will surf the web, rip cds, has decent specs, and play games somewhat decently, people will always buy those over a mac.

Second is availability. Until we see macs in walmart (which last year 30% of all good were bought at walmart, from food to car parts) and kmart, the Mac will never dominate.

So neh :p

DavidRavenMoon
Jul 15, 2002, 09:23 AM
Originally posted by GPTurismo
Well, processor speed = gimmick and marketting ploy. The main reason you need such powerful processors is because of the bloat. 95% of people can get by with a 500 megahertz p3 running win2k. Then people bring up games and yet again, the only area needing upgrading is graphic cards. The actual code is usually trash due to being built in Direct X.

As for the processor speed thing, it's an old argument and it seems idiots are still going to fall for it. Processor speed doesn't matter. Performance is what matters.

Also, the BIGGEST deciding factor got Joe Shopper is price, something even now Macs are still having a hard time beating.

I can get a 1.8 gig p4, 40 gig hdd, with a 17 inch crt, gf2, and 256 ram for 699. Sure it's a pos, but guess what, as long as it will surf the web, rip cds, has decent specs, and play games somewhat decently, people will always buy those over a mac.

Second is availability. Until we see macs in walmart (which last year 30% of all good were bought at walmart, from food to car parts) and kmart, the Mac will never dominate.

So neh :p

In my field of work, the faster the better. Try editing high res images in Photoshop on a 132Mhz 604e or Pentium. It's painful. It's even painful on this 400Mhz G4. Same holds true for hard disk recording. Faster the CPU, the more effects you can run.

You will never see Macs in Walmark, just like you will never see Bose Wave radios in Walmark. Apple is going for the upscale market and that would cheapen their image.

Also it seems Dell has the market as far as PCs, and they don't sell them in Walmart either. Apple will not dominate anymore than Volvo will. But I'd rather drive a Volvo than a Ford.

GPTurismo
Jul 15, 2002, 10:44 AM
Well, as for speed, an older sgi running at 300 megahertz will blow anything you do out of the water. You don't understand what's going on under the hood if you think processor speed is the only factor contributing to performance.

Also, even with such a fast bus, PCs can't take full advantage of it because 90% of the components can't even take advantage of it, ranging from ethernet to the standard buses. To much legacy is in there.

As for Dell having the market, the only reason they can brag on that is BUSINESS MARKET, not consumer. Walmart sells so many make shift brands that it's not funny. and honestly, All the make shift brands and home brew comps out there is greater than the dells out there, especially in the consumer market.

As for what you do, I have done high end graphics from large screen video to billboards to photos to eveerything and I have done large scale databases (oracle, db2) to standard server farms. I know whats screwing up computers.

It's like this, SGI runs their workstations at 800 megahertz, with a single processor running 128 bit irix, and they blow the doors off of almost everything in the market, especially a 2.5 gigahertz p4 running a 533 FSB. The only reason SGI still doesn't dominate the graphics market is...

Adobe shafted them by not releasing Photoshop 4 after they said they would back around 96

and

cost. A good machine from the cost 10 - 15 grand.

So don't tell me it's processor speed.

GPTurismo
Jul 15, 2002, 10:47 AM
Originally posted by DavidRavenMoon
In my field of work, the faster the better. Try editing high res images in Photoshop on a 132Mhz 604e or Pentium. It's painful. It's even painful on this 400Mhz G4. Same holds true for hard disk recording. Faster the CPU, the more effects you can run.

Well, put a full speed system bus on there, or even eliminate that bus, with full speed ram, and everything else running a 133 Mhz, and guess what, that machine will FLY.

As for the volvo for thing, i find that ironic since Ford owns volvo now ;)

DavidRavenMoon
Jul 15, 2002, 10:58 AM
Originally posted by GPTurismo
Well, as for speed, an older sgi running at 300 megahertz will blow anything you do out of the water. You don't understand what's going on under the hood if you think processor speed is the only factor contributing to performance.

It depends on the model. I used to use a few SGI systems. At one shop I worked we had a high end Indigo and a hand full of Indys. We ran Photoshop on the Indys (Dual 250Mhz If I remember correctly) and our 9500s (132Mhz) were much faster!


The only reason SGI still doesn't dominate the graphics market is...

Adobe shafted them by not releasing Photoshop 4 after they said they would back around 96

Yeah, but Photoshop was sucky on Irix anyway.


and cost. A good machine from the cost 10 - 15 grand.

Yep, and the 6 grand Indys couldn't compete with the 3 grand 9500s


So don't tell me it's processor speed.

I never said it's all about processor *clock* speed. Memory throughput and hard disk access are just as important... probably more so, since they are much slower than the CPU.

But for some tasks it's all CPU, so every bit helps. But as your said, for your average Joe, most PCs are faster than they need.

DavidRavenMoon
Jul 15, 2002, 11:00 AM
Originally posted by GPTurismo
As for the volvo for thing, i find that ironic since Ford owns volvo now ;)

Ha! I forgot that. I drive a BMW anyway ;)

alex_ant
Jul 15, 2002, 11:54 AM
Originally posted by GPTurismo
Well, as for speed, an older sgi running at 300 megahertz will blow anything you do out of the water. You don't understand what's going on under the hood if you think processor speed is the only factor contributing to performance.
Agreed, but I think that's a massive exaggeration. I have a 150MHz SGI (R4400SC) and it sure doesn't blow anything out of the water except maybe a 150MHz Pentium. The current MIPS R14000As have a good performance/MHz ratio, but since they only reach 600MHz, that doesn't count for much. A modern Athlon 2100 is faster in both integer and fp than a 600MHz R14000A.
Also, even with such a fast bus, PCs can't take full advantage of it because 90% of the components can't even take advantage of it, ranging from ethernet to the standard buses. To much legacy is in there.
What do you mean by this? Ethernet controllers have nothing to do with the system memory bus.
It's like this, SGI runs their workstations at 800 megahertz, with a single processor running 128 bit irix, and they blow the doors off of almost everything in the market, especially a 2.5 gigahertz p4 running a 533 FSB. The only reason SGI still doesn't dominate the graphics market is...
But the fastest CPU used in SGIs today is 600MHz. Were you overclocking? IRIX/MIPS is 64-bit, btw.
Adobe shafted them by not releasing Photoshop 4 after they said they would back around 96

and

cost. A good machine from the cost 10 - 15 grand.

So don't tell me it's processor speed.
Adobe discontinued Photoshop for IRIX because it wasn't selling. It wasn't selling because SGIs have an insanely poor price/performance ratio compared to Macs and PCs. SGI owns the high-end graphics market, where PCs and Macs can't scale to, but why run Photoshop on a $15k SGI when you can run it on a Mac or PC costing 1/3 as much and get better performance?

Alex

alex_ant
Jul 15, 2002, 11:56 AM
Originally posted by DavidRavenMoon

It depends on the model. I used to use a few SGI systems. At one shop I worked we had a high end Indigo and a hand full of Indys. We ran Photoshop on the Indys (Dual 250Mhz If I remember correctly) and our 9500s (132Mhz) were much faster!
Just a minor correction, the fastest Indy was 180MHz, and that one was rare. Most common models were 133 and 150MHz. None were dual-processor. Especially in their early years, they were weighed down heavily by IRIX (which was very bloated, and still is, but you can't tell as much because the hardware is so fast now).

Alex

Chryx
Jul 15, 2002, 01:36 PM
Originally posted by Marianco
Athlon-based motherboards are not totally compatible with certain hardware and software packages like the Intel Pentium-based motherboards.

For example, I do a lot of videoconferencing. The videoconferencing hardware/software for a 3-ISDN line 380 kbps setup I have specifically excludes Athlon compatibility. I thus have to use a Pentium-based motherboard, despite how great the Athlon is.

Yes, perhaps it is the fault of the motherboard makers, but for this niche, the hardware surrounding the Athlon is different enough to make a big difference.

In other words you ARE blaming AMD for VIA's incompetance.

Here's a hint, there are chipset makers OTHER than VIA, and they manage to make chipsets that actually work.

the Athlon "boards" have NOTHING to do with software compatability either.

Try AMD's 760/762 chipsets (the 762 is the SMP variant)
or SiS 735/745 chipsets (cheap 'n' very cheerful) (I'm using an SiS 735 here)
or Nvidia's Nforce.

I'd place money on ALL of them working with your video conferencing hardware. if it declares itself Athlon compatible or not.

(and for the record, Intel's chipsets aren't perfect, their current generation have a PCI bus bug that cramps the maximum PCI transfer rate to about 85MB/s )

Chryx
Jul 15, 2002, 01:44 PM
Originally posted by DavidRavenMoon


But actually it is running twice as fast as SD Ram, because it's running on both sides of the clock cycle. (Think of a square wave and you will see what that means.)

So for every tic of the clock it's advancing two tics. SD (Single Data) Ram only operates on either the rising or falling (I dont remember which), so PC133 is running at 133MHz, but DDR clocked at 133 is running at 266Mhz. The bus is 133Mhz, but it's reading twice as much data per clock cycle.

266MHz DDR would be 532DDR.

It's not just marketing. Bottom line is it is faster :)

It IS faster, but it's still not 266Mhz :)
(and there's evidence that it's not 100% faster either, a few companies released Geforce 2MX cards with DDR / 64bit memory and the SDR / 128bit cards ate them alive at the same clockrates..)


Also, IIRC, It's "double DATA rate" the command rate is still tied to the initial rising part of the clock pulse. (so it can fetch data faster, but it still takes the same amount of time to start fetching it.)

(oh yeah, poke around for posts by me, I posted ascii diagrams of SDR/DDR/QDR for someone else :))

stripes
Jul 15, 2002, 03:16 PM
Originally posted by DarkNovaMatter
I do admit this doesn't seem possible because of the work required to make the chip (get it to do ppc- unless it is the ppc coding).

The POWER4 does the POWER instruction set, the PowerAS instruction set, and also the PowerPC instruction set. It does not do AltiVec. IBM did finally licence AltiVec from Moto though, so maybe if they design a "cut down" POWER4 they will add AltiVec to it.

Of corse that takes time, so unless they started that about a year ago when they licenced AltiVec, we won't be seeing it this week!

Phil Of Mac
Dec 13, 2002, 08:30 PM
Ha! Now we're all talking about the 970!

THIS CAME TRUE! MOSR OWNZ ALL!!!!

jefhatfield
Dec 15, 2002, 11:56 AM
Originally posted by Mr Jobs
no the Gx is an Apple trade mark, G3 was originally a internal code name but everyone know about it and it was so popular Apple kept it as a release name.

i think it's about time the naming convention be changed to something more creative than "generation 3 chip", generation 4 chip, generation 5 chip, etc

i like the names athlon and duron...it beats pentium 1, 2, 3, 4, etc

cr2sh
Dec 15, 2002, 02:09 PM
Originally posted by DavidRavenMoon
266MHz DDR would be 532DDR.
It's not just marketing. Bottom line is it is faster :)

It is faster, but it's certainly not twice as fast. I felt a bigger difference going from a pc100 system to a pc133, than I did a pc133 to DDR. Why is that?

BenderBot1138
Dec 15, 2002, 02:19 PM
lotsa talk about cost, but what's the damage, especially in a larger production grid.

:cool:

Steradian
Dec 22, 2002, 02:12 AM
Cr2sh which system where you using when you didn't notice the differance between DDR and PC-133? was the computer a PM?
The PM's we have now use some sort of castrated motherboard that doesn't take full advantige of the DDR ram speed's.

PowerBook G4 1ghz

ELYXR
Dec 22, 2002, 02:24 AM
Do you guys know anything about QDR (Quad-Data-Rate) memory. Isn't that what Rambus uses in their RDRAM product?

Is that the next step for SDRAM?:confused:

cr2sh
Dec 22, 2002, 02:34 AM
Originally posted by Steradian
was the computer a PM?


No, both were Pc's.

alex_ant
Dec 22, 2002, 02:57 AM
Originally posted by Phil Of Mac
Ha! Now we're all talking about the 970!

THIS CAME TRUE! MOSR OWNZ ALL!!!!
MOSR doesn't make up its own rumors, it gleans them off of others. They likely copped this one from me, since I, being the prescient genius I am, predicted the 970 way back in April (almost 2 months before word broke of this chip). Nyah

Dr. Distortion
Dec 22, 2002, 08:04 AM
Whoa! Did we really have to dig up this old thread?

jefhatfield
Dec 22, 2002, 08:11 AM
Originally posted by Dr. Distortion
Whoa! Did we really have to dig up this old thread?

i think there are a lot of mac power users who are desperate for a new chip

...especially since the pentium 4s are flying around 3 ghz

nallu
Dec 22, 2002, 11:06 AM
There are rumors that g5 exists but yeilds are too low and hence Moto was forced to abonden it
dont you think it will be cool if IBM bought what is remaining of g5 and developed it. they have both the funds and tecnology.
It was said that the g5 gave 40-80 gflops with 5 built-in altivecs even at 1.6 ghz.
thats is almost 5-10 times the top current performance. imagine the same chip at 2.4 ghz!!!.

well atleast it may be worth a try.

and i seriously think its time to think big apart from diffferent.
a processor way above competition, software way above competition. then people will buy the machine for more than its design. "these are hard targets so is survivial"