Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ibm vs amd

i unfortunately think that apple might have put their money on the wrong horse by going with IBM. They probably should have picked AMD last year for the new Power Mac's and XServe's.

The problem with IBM is that this truly isn't their core business. Not trying to be critical.
 
IBM is a safe bet

river_jetties said:
i unfortunately think that apple might have put their money on the wrong horse by going with IBM. They probably should have picked AMD last year for the new Power Mac's and XServe's.

The problem with IBM is that this truly isn't their core business. Not trying to be critical.

Why should they have done this? By switching to AMD, Apple whould abandon the PowerPC platform and throw compatibility out of the window plus requiring a port of Mac OS X to x86. Moreover they would need a rewrite of ALL the current applications for the new hardware.
I believe that switching to IBM was the wisest thing Apple has done in the last few years.They kept their PowerPC platform, progressed to a chip with great performance(G5's floating point performance is bettern than the Opteron's, see flop/s ratings and the virginia tech interview) and placed their bet on IBM who uses the POWER processors for most of her systems. So it will definetely support and progress the architecture and it can always uses this progress to future PowerPC chips.
I am not worried about this, I am only worried if we will actually see the PowerPC 975(if it is named like this) this year and a PowerBook G5 till the first quarter of 2005.

Maverick
 
It is interesting that I see much of what you describe as far as corporate mentality goes with respect to Apple. I work for a large corproation, but outside of the main IT departments. It is an attitude that is largely unjustified in my opinion.

Frist in Apples defense they keep the same basic designs around longer than anyone in the industry. They have also kept their OS stable while doing a pretty amazing job of upgrading it.

From the corporate perspective I've seen a lot of strange things there. Standardizing on a new vendor every year or so is one of the moves common in IT departments which makes me wonder what value there is in any NDA. I could go into more detail but I think it is safe to say that Apples approach is used as an excuse at times but when it comes right down to it other things are a factor in IT attitudes.

I look at it this way if you standardize on Apple as a platoform you pretty much are at their mercy. It is not so much that they won't share the informaiton it is more a matter of the platform not offering alternatives.

Thanks
dave



Trekkie said:
IBM does not keep it's customers in the dark. I should know, I'm one of the 'keepers of the light' when it comes to the BladeCenter product lines. I'd happily tell a customer who has signed an NDA what we are doing well into 2006 and even 'thoughts' about 2007 at this time.

Apple is the enigma in the computer industry as far as keeping their mouths shut. That is probably what hurts them to some extent in the enterprise space. They don't like surprises and want to have a 24 month planning roadmap even for desktops. Having a new box show up out of the blue really ticks them off - also when a box stops before they were told it would isn't a good thing either..
 
I could bet you that when Steve made those 3ghz promises Apple has already had some working 3ghz prototype boxes in their labs. But god knows what Power derivative those bxes were using. I still think that 970 was rushed off to market (tucked-in AltiVec, Power4 core,no ddr controler on die) as stop-gap measure. Apple-IBM have certainly been working together on more than just 970. Question however is ,how long Apple will be milking 970's in their pro desktops before they decide to move upIn the last 12 months there certainly wasn't much pressure on them from Intel or AMD while everyone pretty much stagnated. They could more or less afford to sit on 2 Ghz for a year now. Intel is about to release 3.7 and 4.0 ghz P4's with upto 2 MB L2 cache + other enhancements, where 2.4 ghz 970 just won't cut anymore. Apple knows this. ;)
 
Translation

makkystyle said:
Here is an interesting link. It leads to a PDF in french that mentions IBM using "PowerPC 970 at 3Ghz" in their Blade servers. If someone out there reads french and can please give us a rundown on what this says that would be great. I can't vouch for the validity of this article, because I have no idea where it came from (the whole not being able to speak french thing). Also it doesn't lend itself to Sherlock translation very well because of the format. Hope someone can shed some light:

http://www.irisa.fr/orap/Publications/Bi-orap/biorap-35.pdf

I knew that C in college French would pay off...

The section that mentions a 970 at 3ghz translates approximately as follows:

"The secret weapon of IBM will be, without any doubt, its server "Blade", a super-dense architecture: the same frame of 7U will be able to contain 14 "blades" Bi processors equivalent to either 28 Pentium 4s DP Xeon 2.4 Ghz with network gigabit Ethernet and network of administration integrated for 32 bit calculation, or a POWERPC 970 with 3Ghz (a denser version of the POWER4) at 3 Ghz with Myrinet network integrated for 64 bit calculation! The whole functioning under... Of course Linux. There is no doubt that the advertisement of this product will make the effect of a bomb..."

So basically, IBM's "secret weapon" is a processor which has the speed of a 3.0 GHz 970... Another false alarm. The article did get one thing right though. It surely did have the effect of a bomb... a bomb of disappointment =P

Linux? What is with this talk of Linux?

-Clive at Five
 
970 3Ghz

1-
So basically, IBM's "secret weapon" is a processor which has the speed of a 3.0 GHz 970...

the article from the irisa is date April 2003 :)

2-
Linux? What is with this talk of Linux?

IBM says "when you put 1$ in a server on m$ Windows, you have to put 1$ (or more) in software to work with... so buy an IBM server at 1.5$ and put linux on it... it's free... you would have a better and more powerfull hardware for less" :D
So... it's normal to talk about Linux... :cool:
And it's easier for them to offer drivers... and it's a p-series IBM server, not a X serve Apple... :D
What OS do you want they sell with ? :confused:
 
A full 24 hours have passed and I am still incredulous that someone wrote "Thaiwanese". I just can't get over it.

Anyway, everybody here keeps bashing the 970 processor and stating that it needs to be replaced with something else. Clock-rate aside, why does anyone want to replace the 970/970FX? Are you all climbing on the wagon and shooting in the air just because everybody else is?

So far as I can see, they're nice chips.

And just for the record, I still can't get over the fact he wrote "Thaiwanese".
 
This is ridiculous

Twonk said:
I think it's real. I want one. Would be kinda funny if it was all hot air though :D

This is getting ridiculous. Now the rumor sites are reporting about rumours about rumours.

There's nothing to see here. Back to your lives, citizens.
 
and the CDs ...

Telomar said:
Keep in mind that is largely a support contract price.

If you believe that, then please post the URLs where I can download Enterprise Linux 3.0 for free...

Or SuSe Professional for workstations, that lists at $89 or so.
______________

Most of the people buying IBM POWER servers aren't the kind of people who enjoy downloading random bits of their operating systems from various places, recompiling the kernels, and then trying to get it to boot and stay up.

The TCO of a $1295 contract is much better than "roll your own".
 
Wendy_Rebecca said:
Now the rumor sites are reporting about rumours about rumours.

Don't forget the "proof" which is really a quote from an IBM web page in Chinese from the fantasy island of Thaiwan which also mentions the rumour !!

Wow !!
 
More Bad News

I apologize if this has already been posted but here is story on CNet that says IBM own blade servers are being pushed back due to low chip yields. It also references difficulty with the Xserves:
However, shipments of Apple Computer's Xserve, which also uses the PowerPC 970, have been constrained by a shortage of chips from IBM. "Obviously we were not happy with the delivery we got," Timothy Cook, Apple's executive vice president of worldwide sales and operations, said in April, adding that the company hopes to have caught up with demand by the end of June.
Doesn't sound like a company getting ready to ship an all new, 3.0 GHz version of the 970 (or 975 or 980...).
I wonder if the best we can hope for is the 1.6 GHz PowerMac becoming a dual and a lower cost single processor line of the same speed chips.
:( :( :( :( :(
 
qubex said:
A full 24 hours have passed and I am still incredulous that someone wrote "Thaiwanese". I just can't get over it.

And just for the record, I still can't get over the fact he wrote "Thaiwanese".

Since most of us are not from Taiwan or Thailand it really doesn't seem like more than a typo. That extra letter "h" must really bug you...

*ahem*

To stay on topic I'd like to say that, whatever IBM is calling the 970 successor, it will be in the new PowerMacs this year... probbably WWDC. I wouldn't be suprised. I guess I'm not going to lose any sleep over it, but I believe a 3.0 GHz part is coming from IBM sometime this year.

The nice thing is that a 3.0 GHz 975/980 whatever will be faster per clock than the 970/970fx. Sweeet.
 
pjkelnhofer said:
Doesn't sound like a company getting ready to ship an all new, 3.0 GHz version of the 970 (or 975 or 980...).
I wonder if the best we can hope for is the 1.6 GHz PowerMac becoming a dual and a lower cost single processor line of the same speed chips.

One of 2 scenarios have been covered: either the 970 delays ARE or ARE NOT going to effect the successor (975/980) shipments. My belief is that they will not, but I have no inside information. The 970 and 975 are different beasts and assuming they use different production processes we should be okay.

My fingers are crossed. :)
 
Frobozz said:
Since most of us are not from Taiwan or Thailand it really doesn't seem like more than a typo. That extra letter "h" must really bug you...

It was more than a typo. It was repeated, there was the Thai in who speak Thaiwanese. It is very funny go back a couple pages and read it.

To stay on topic I'd like to say that, whatever IBM is calling the 970 successor, it will be in the new PowerMacs this year... probbably WWDC. I wouldn't be suprised. I guess I'm not going to lose any sleep over it, but I believe a 3.0 GHz part is coming from IBM sometime this year.

The nice thing is that a 3.0 GHz 975/980 whatever will be faster per clock than the 970/970fx. Sweeet.

It is nice to dream isn't it, but there is absolutely no indication that IBM, who cannot even mass produce the 970 or 970fx chips with a good yield, is secretly producing a chip that is 150% faster in pure clock speed and even faster than that in performance. Maybe we will be suprised, but I doubt that they could keep the existence of a chip that would be essentially twice as fast as what is out there right now, this much of a secret.
 
pjkelnhofer said:
IBM, who cannot even mass produce the 970 or 970fx chips with a good yield

says who? i remember seeing a single slide of some ibm presentation where they had put an estimated good-cpu rate and an actual produced good-cpu rate on the same chart, and the actual numbers weren't so far off the estimate. so if they haven't estimated a failure, they aren't so far off - given that the slide i saw was a genuine one.

i think too little about this topic is based on facts.
 
qubex said:
Anyway, everybody here keeps bashing the 970 processor and stating that it needs to be replaced with something else. Clock-rate aside, why does anyone want to replace the 970/970FX? Are you all climbing on the wagon and shooting in the air just because everybody else is?

I was among the first to start up the '970 looks hokey' strikes after the initial furor over their release, and I stand by my reasoning to this day. The SIMD implementation is less efficient than the G4s and was just tacked onto the Power4 core so that it could do vector math the way Apple would want to. There's no on-die memory controller, which is one of AMD's big strengths and part of the only reason that their chips beath the G5 at similar clockspeeds on some tasks. If IBM and AMD were working on the same projects, sharing the same fab space, and the former was licensing technology to the latter, you'd think that they might have similar design philosophies, no?

The Power5 core has all the things the 970 does not. It has fewer pipeline stages but faster clock, on-die DDR and DDR2 control, SMT implementation, a multi-core design that could possibly be staged at the consumer level, and the rumors say that they 975/980/what-have-you is designed from the ground up to be a new chip. It could easily have a fully functional, even more optimized dual-precision SIMD unit, especially since that's a published feature of FreeScale's next-generation chip - 128-bit dual precision vector math.

So far as I can see, they're nice chips.

Nice is not the same as cutting edge, and the technology world is always moving. The 970 is starting to fall behind, and it just doesn't feel like it was really intended to be the flagship.
 
JFreak said:
i think too little about this topic is based on facts.


Fact: Order an XServe G5 - 5-7 weeks quoted for delivery

Fact: Apple said IBM's chip delay only reason for XServe delay

Fact: Virginia Tech's cluster is shut down, and probably won't be in the next Top500 list at all - they sold the PowerMacs and haven't been able to get the XServe G5s....

http://www.tcf.vt.edu/
http://www.thinksecret.com/news/virginiatech3.html
__________________

Those facts don't point to a huge success....
 
AidenShaw said:
Fact: Order an XServe G5 - 5-7 weeks quoted for delivery
Fact: Apple said IBM's chip delay only reason for XServe delay

this can also mean that business purchase decisions have been delayed in anticipation of the G5 going into xserves, and apple is facing more demand than they predicted in the light of G4 xserve sales. my interpretation is not a fact, but a fact is i can very well be right.

AidenShaw said:
Fact: Virginia Tech's cluster is shut down, and probably won't be in the next Top500 list at all - they sold the PowerMacs and haven't been able to get the XServe G5s....

someone has then made a stupid decision for selling the production hardware without knowing the replacement unit's delivery date. that is poor business practice on behalf of virginia tech, IF they need to have the supercomputer up and running at all times. that is not apple's fault, nor ibm's. they should have made such a deal that the existing hardware leaves the building a day before the replacement arrives, IF they need to have the supercomputer up and running at all times.

if they can afford downtime, they may have reasons other than staying on the list. we can only guess.
 
To thatwendigo:

Fair enough, I see your line of reasoning, even though I don't necessarily agree with it.

But having just recently introduced the POWER4-derived 970, doesn't it seem odd that IBM would run off and introduce the POWER5-derived "975" (or whatever) a mere twelve months later?

POWER4 and POWER5 are radically different designs, as you indicate (and I know first hand because I have an AS/400 Advanced Series with POWER4 chips inside). It seems to me that would instantly vanify all the hard work and money invested into the 970: in other words, retrospectively, investing all that hard work into the 970 would look like a very bad idea if they were planning to introduce the 975 already a year ago.

I can foresee a hypothetical 975 as being an "evolution" of the 970/970FX with a few additional features - for example, SMT or even revised AltiVec. But I think it is simply too early for the Next Generation to be unveiled.

That's just my own opinionated view. I'm not really informed so I guess it's just idle speculation.
 
JFreak said:
says who? i remember seeing a single slide of some ibm presentation where they had put an estimated good-cpu rate and an actual produced good-cpu rate on the same chart, and the actual numbers weren't so far off the estimate. so if they haven't estimated a failure, they aren't so far off - given that the slide i saw was a genuine one.

i think too little about this topic is based on facts.

Once again here is the quote from the CNet article:
However, shipments of Apple Computer's Xserve, which also uses the PowerPC 970, have been constrained by a shortage of chips from IBM. "Obviously we were not happy with the delivery we got," Timothy Cook, Apple's executive vice president of worldwide sales and operations, said in April, adding that the company hopes to have caught up with demand by the end of June.
It's been a nearly year and so far there has been absolutely no increase in the speed of 970's that had been made public.
That is a fact.
 
Wendy_Rebecca said:
This is getting ridiculous. Now the rumor sites are reporting about rumours about rumours.

There's nothing to see here. Back to your lives, citizens.
That's the 95% of what's out there and that's what the internet is good for. And you're forgetting that the original purpose of MacRumors was to provide a kind of a digest of floating rumors.

As for chip yields, the 90 nm process has been a disaster so far. All this talk of working prototypes of 980 this and 975 that is pointless if IBM can't produce these chips in MASS quantities, not just samples for testing. I'm of the camp that believes the 975/980 faces the same manufacturing hurdles as the 970fx. Now, IBM is claiming that they're getting a handle on the problem. We'll see. As someone else have noted, the Xserves still show 5-7 week shipping times; that hasn't changed since March I think.
 
VaTech was a given...

JFreak said:
apple is facing more demand than they predicted in the light of G4 xserve sales.

Well, the VaTech cluster systems didn't have to be predicted - and Apple wasn't able to fill that order in a timely manner.
 
AidenShaw said:
Well, the VaTech cluster systems didn't have to be predicted - and Apple wasn't able to fill that order in a timely manner.

And, they received preferential treatment when they got the original G5's, you would have thought they would get that for the Xserves as well.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.