By the way, "Algol", once upon a time I used a programming language called that. I think that it was assimilated into one of the Berkeley 'nix variants. Is that where your handle came from?
Regards
Gaack!!! I still have a copy of "Fanget An: An AlgolW Primer" on my bookshelf next to me, from my Freshman year at college... 25 years ago...
according to AMD, the opteron @ 2GHz specs out at:
specint: 1,202
specfp: 1,170
the 970 @ 1.8 GHz:
SPECint - 937
SPECfp - 1051
it has not been published where prescot will land, but I find it really hard to believe that the 970 will be able to top the alternatives in performance, cost, and arrival date. and even if it does, what will apple charge for it with their 28% profit margins... in addition, for lightwave, most of the stuff I need to do is real time... I'm not too concerned with final renders, mostly modeling/animating performance. anyway... just thought I'd add the numbers that have been in my head to the conversation. I really don't care about a 10% disadvantage from a mac workstation in price, or performance. but more than that is stretching it, and arrival date is HUGE.
I don't think that I would necessarily put much faith in AMD right now. Their theoretical SPEC numbers look good compared with the 970, but with the amount of cash AMD is hemorrhaging, they will struggle to keep up with IBM and Intel in scaling their chips. Intel is stuck with 32 bit mode in the consumer arena for the forseeable future and HPaq will do everything in their power to see the Itanic succeed in the pro and server market as the idea behind EPIC came from their own engineers. Gateway is a skeleton of a company and Dell is the only real hope that AMD and its Opteron has. Intel will ensure that Dell stays in the Pentium/Itanium fold by ensuring that HPaq otherwise gets preferred pricing on all Pentium, Itanium and Centrino processors. I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy. IBM itself feels that AMD's days are limited. So the market really comes down to Intel and IBM. Let's not fool ourselves on this one. AMD is in very real trouble here and the Opteron processor will need some serious help in gaining any type of marketshare. Intel will simply price their 32 bit chips so low that AMD won't be able to compete. The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes. Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it. Though the Itanium emulates x86 at 486 speeds, HPaq might just as well put together a box with an Itanium chip along with a Pentium or Centrino processor to run 32 bit x86 applications for those who absolutely need this type of solution. Intel will be more than happy to provide both processors for less than the cost of a single AMD Opteron. Intel's size is the ultimate trump card that AMD just cannot compete against. I am glad Apple chose to partner up with IBM. They are the only company with the size and resources to take on Intel and offer a better solution. It is over for AMD. They should have gotten into the graphics chip market when they had the chance. They might at one time had a chance to compete against ATi. With NVidia in the game and using IBM fabs. AMD can't go here either. It's too bad because they are an innovative company. But business being business, the best solution doesn't always win. So realistically it is the 9x0 series vs. the Itanic series. The Opteron is dead for all practical purposes. The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps. IBM won't be affected by price pressure from Intel as Apple is committed to the PowerPC. Lets be realistic here, Apple and IBM are in a superior position with the PowerPC and OS X. Intel, AMD and Motorola missed the boat. Good companies always find a way to survive. Sony messed up with betamax, but the beta vs. VHS argument is now meaningless. Apple and IBM messed up also. Apple with the Mac OS, and IBM with the PC. Microsoft and Intel double crossed them both. It is only fitting that Apple and IBM finally return the favor. But, I don't want anyone to mistake my comments to mean that MS and Intel will go under. They are too big and have too many resources to see that happen. Even IBM reinvented itself. But it had the size and resources to do so as do Intel and MS.
Still, I will wait for Apple's 970 machines. I'll get to have the best OS on a viable platform that will scale quickly with IBM's commitment to take on Intel with the PowerPC line of chips.
[snip] I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy. IBM itself feels that AMD's days are limited. So the market really comes down to Intel and IBM. Let's not fool ourselves on this one. AMD is in very real trouble here and the Opteron processor will need some serious help in gaining any type of marketshare. Intel will simply price their 32 bit chips so low that AMD won't be able to compete. The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes. Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it. Though the Itanium emulates x86 at 486 speeds, HPaq might just as well put together a box with an So realistically it is the 9x0 series vs. the Itanic series. The Opteron is dead for all practical purposes. The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps
[snip]
Even IBM reinvented itself. But it had the size and resources to do so as do Intel and MS.
I totally agree with what you are saying... but next time, may I suggest shorter paragraphs. It makes it easier to read
My take on it [as written in the Thin Client thread] is that as new processes are needed to compete, fewer and fewer companies will be able to fab or do proper R&D, and therefore survive. This effectivly leaves IBM and Intel.
according to AMD, the opteron @ 2GHz specs out at:
specint: 1,202
specfp: 1,170
the 970 @ 1.8 GHz:
SPECint - 937
SPECfp - 1051
it has not been published where prescot will land, but I find it really hard to believe that the 970 will be able to top the alternatives in performance, cost, and arrival date. and even if it does, what will apple charge for it with their 28% profit margins... in addition, for lightwave, most of the stuff I need to do is real time... I'm not too concerned with final renders, mostly modeling/animating performance. anyway... just thought I'd add the numbers that have been in my head to the conversation. I really don't care about a 10% disadvantage from a mac workstation in price, or performance. but more than that is stretching it, and arrival date is HUGE.
At equal speed, the 970 has similar performance than the opteron, at a cost of a smaller chip with a smaller cache (512 K versus 1024 K) and less heat issue. Add a WMX engine, much more efficient than the various SIMD units (MMX, 3DNow, SSE, perhaps SSE2) of the Opteron, and you have a winner. Consider also the presence of a top end front side bus, and you bet that the chip of IBM may be better.
Consider also that the management of the 32 bit mode of the Opteron is less efficient than the one of the PPC 970, and you will be able to guess that the new IBM chip appear to have a brightest future.
At equal speed, the 970 has similar performance than the opteron, at a cost of a smaller chip with a smaller cache (512 K versus 1024 K) and less heat issue. Add a WMX engine, much more efficient than the various SIMD units (MMX, 3DNow, SSE, perhaps SSE2) of the Opteron, and you have a winner.
That's more what I was trying to say. (That's _VMX_ == altivec == velocity engine) It extends to duals as well, where dual hammers or dual Xeons or whatever start costing serious money -> price parity.
In addition, I can see that 1051 is less than 1170ish. But it's not 1/5th of 1170, nor is it 1/3 of 1170... it's within the 'benchmarking error' area. 10-15%ish. I'm not a lightwave guy, but the historical reason reported for poor Mac performance on lightwave is way sub-par FPU performance. YMMV of course.
If you have to buy, you have to buy. I'm holding my machine together with duct tape, baling wire, and sweat until the 970's available is all I was implying.
I wonder how eager developers will be to port their applications to 64-bit X-86. With HP pushing the Itanium processor, developers are torn between the two or faced with porting to both. Developers may decide to wait a while, to see which one takes off, unless the job of porting is trivial.
The situation is much more clean and clear with Apple and the IBM 970. The 64 bit PPC design makes the 64 bit transition as simple as possible.
I guess paragraphs would make things easier. I do tend to go into great detail as my PC using colleagues are very difficult to reason with at times.
I'll try to organize things better and into paragraphs. Right now I am dealing with PC geeks who keep telling me that the Opteron is the future. They love to point out that it has better numbers than the 970 and how innovative AMD is. They won't listen to any logic.
In my emotional state of mind, I got carried away and did not take the time to organize my thoughts better.
In any case, HPaq's dedication to the Itanium spells certain death for the Opteron. Dell would be crazy to touch the Opteron as it will likely upset the folks at Intel who would immediately ensure that HPaq gets very preferred pricing on Intel chips. Without a compelling reason to move to 64 bit chips, Dell would lose out as a price leader on 32 bit chip computing where the market will remain for the near future. There is no way that AMD can price the Opteron low enough to compete with Intel's chips either. At least, and make a profit.
There were some comments about 970 and 0.90 nm process. The IBM plans that have been on the web do not show any such combination. IBM plans to go to the 980 as soon as possible next year. Initially the 980 will be in the 0.130 nm process and then go to a 0.90 nm process.
The 970 chip is, apparently, viewed as just a transitional chip. The 980 is supposed to be superior in performance in comparison the the 970 and is supposed to consume less power and generate less heat and (being on a smaller die) cost less to produce. It is only a guess, but presumably Apple will also move to the 980 as it becomes available for all of the above reasons.
One can only hope that Apple will have done enough testing to get things right the first time on this one. They don't need any more "oops!" machines.
One thing that I have not seen discussed about the new machines is how well they might work with Real PCs soon (everyone hopes) to be released unix based Windows emulator. You would hope that Apple would have been working very closely with them on this. Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination.
One thing that I have not seen discussed about the new machines is how well they might work with Real PCs soon (everyone hopes) to be released unix based Windows emulator. You would hope that Apple would have been working very closely with them on this. Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination
A source close to Apple recently relayed information to oscast suggesting that Apple had intended to use emulation software to show the power and performance of its upcoming hardware. This individual stated that Windows software running in emulation performed certain instructions significantly faster than what a hard PC was able to do.
Another (separate) source told oscast that a key reason for Microsoft's acquisition of Connectix's PC emulator, was to "discontinue development of a version of [Virtual PC] that would take advantage of two unreleased hardware technologies that would significantly boost the software'sperformance."
Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination.
I don't think Apple will ever incorporate windows compatibility into the os. Too dangerous. Many companies could stop developing for osx if there was a way too run windows apps without problems on the mac.
Having PC-emulation software on the computer is very useful and I think Apple would help RealPC in every way they can.
[I]A source close to Apple recently relayed information to oscast suggesting that Apple had intended to use emulation software to show the power and performance of its upcoming hardware. This individual stated that Windows software running in emulation performed certain instructions significantly faster than what a hard PC was able to do.
I don't think that I would necessarily put much faith in AMD right now. Their theoretical SPEC numbers look good compared with the 970, but with the amount of cash AMD is hemorrhaging, they will struggle to keep up with IBM and Intel in scaling their chips.
Actually, though AMD is making a loss they are in apparently good financial position:
Especially the last paragraph; "AMD could be considered to be cash rich..."
How is it so? Well how is it that Apple has a reported $4 billion in cash after making hugh losses in the '90s and has only been making modest profits since then?
AMD also just signed up as a partner with IBM to work on process tech together. I presume AMD will also get IBM to fab some of their chips as part of this deal since IBM is looking for work to fill its fabs.
Quote:
Intel is stuck with 32 bit [...] HPaq [...] to see the Itanic succeed [...] Gateway is a skeleton [...] Dell is the only real hope that AMD and its Opteron has. Intel will ensure that Dell stays in the Pentium/Itanium fold by ensuring that HPaq otherwise gets preferred pricing on all Pentium, Itanium and Centrino processors.
IBM is not small player in x86 and are reportedly going in with both hands and feet using Opteron chips. Dell is also a likely customer, contrary to what you think.
Quote:
I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy.
A short sighted statement that might only have strength if AMD had started development last year as their market share began to fall again. However, AMD started work on the K8 years ago, when their market share was going up so it is not a last gasp effort as you say. They company does depend on its success but all reports say it will be very competetive with the P4 and Xeon chips.
Quote:
IBM itself feels that AMD's days are limited.
No they don't of they wouldn't be planning to use their chips or have signed a partnership agreement.
Quote:
Intel will simply price their 32 bit chips so low that AMD won't be able to compete.
No so. AMD's Opteron and Athlon 64 dies are small. On the same process the A64 is small than a P4 and can hence be cheaper to produce. The Opteron is smaller than an Itanic by far so it can be much cheaper - price performance wise it will hit Itanic for six.
Quote:
The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes.
No they can't. AMD has much room to spare with their Barton core and it competes well with the P4. If A64 has some startup woes they will introduce faster AthlonXPs to stay in touch with Intel. The big point is, and the same goes for the 970, it is not so much 'backwards compatible' with 32 bit but natively 32 bit with native 64 bit capability too.
Quote:
Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it.
And to Opteron. IBM's DB2, Oracle, etc. No small fry in 64-bit land.
Quote:
Intel will be more than happy to provide both [a P4 and Itanic] for less than the cost of a single AMD Opteron.
Absolutely not true at all. See above. AMD has the upper hand with respect to pricing. Itanics cost thousands of dollars! See this article:
"A 3GHz PENTIUM 4 for a Canterwood will cost $415..."
Quote:
I am glad Apple chose to partner up with IBM.
So am I. I think the 970 will be excellent for Apple and I think that IBM will follow on with derivatives of the 970 and then the Power5 and really give them the hardware to match OS X.
Quote:
But business being business, the best solution doesn't always win.
This is an argument against the 970 as much as the Opteron. Neither currently have market share. Both perform well (or are expected to) against competitors.
Quote:
So realistically it is the 9x0 series vs. the Itanic series.
No, it's the Power series vs the Itanic. Those two families are aimed at different markets than the 970 and it's successors.
Quote:
The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps.
The Prescott is starting on 90 nm and will come in a mobile version but the Pentium M is Intel's real mobile chip anyway.
Quote:
Still, I will wait for Apple's 970 machines.
Yes, so will I. I'm sick of what MS is trying to force me to do via the Window's platform and will become a switcher as soon as I decide to upgrade my Athlon 1.2 GHz machine I got in 2000 and 2001. I'm planning a PowerBook too - a 970 PB would be nice.
Quote:
I'll get to have the best OS on a viable platform that will scale quickly with IBM's commitment to take on Intel with the PowerPC line of chips.
Yes but the competition will be:
970 vs Pentium4 family and
Power vs Itanium
I still think you underestimate AMD and the 'hammer' line.
I still think you underestimate AMD and the 'hammer' line.
MM
I guess we shall see.
As I said, I think AMD has done an excellent job with this chip. It is not that I am underestimating them. I think most people are underestimating Intel. I just don't see AMD being able to compete with a company the size of Intel for any length of time.
Intel can use its size and economies of scale to produce chips more inexpensively than AMD. Just because they sell the Itanium for an obscene price, doesn't mean that they couldn't change gears and lower prices, slashing margins and putting intense heat on AMD.
Even with their huge investment in the Itanium, Intel is still profitable. Selling the Itanium chips at low prices means that it takes longer to recover their investment, but it is a luxury that Intel has.
Not only that, AMD is going against IBM's Power series with the Opteron, and the 9x0 series with the Athlon64. Again, I am not trying to underestimate AMD, but I just don't see them staying with IBM in the long run.
The upcoming Power5 will simply be too much for AMD. Intel, however, should be able to able to engineer and fabricate a version of the Itanium that could compete.
Besides, AMD will be dependent on Microsoft in order to succeed. Opteron really has no advantage in the Linux server market over Itanium, other than price. The chip does have an advantage in the x86 Windows markets. However, I am not sure that this is enough to overcome HP's full dedication, and Dell's strong bias, for Intel. DB2 and Oracle on Linux are just not enough. As I recall, DB2 and Oracle are also fully committed to the Itanium platform also.
I am guessing that the Opteron will face stiff competition with the Itanium in the server market. There will be relatively little software on both platforms, however. The desktop version of the Opteron processor is also not slated for release for perhaps another 6 months.
In the 64 bit desktop market, this chip has little chance against OS X on the 9x0 series. Apple will develop a slew of 64 bit applications for the desktop and also laptop. Most Windows developers will remain firmly in the 32 bit arena for that is where most of the legacy hardware is. I'm not sure that x86 developers will be able to compete with Apple in these types of apps, either.
At 80 to 90 watts of heat dissipation, we can forget about putting one of these AMD beasts into a laptop. They won't be able to compete with the Banias chip, much less the 970.
I just don't see Dell giving up its preferred pricing with Intel to take a gamble on the Opteron. They will certainly lose in the laptop market with HP getting preferred pricing on Banias chips.
No matter how I analyze the situation, I just don't see much hope for AMD. If Apple were taken out of the picture, AMD might have a chance. With Apple in the picture, AMD gets squeezed in a big way. 64 bit Windows and x86 apps. won't effectively compete against OS X for the desktop. That is AMD's only real chance as IBM and Intel are going to really turn up the heat for the high dollar server market.
I'm sorry, I see AMD running out of cash in the not too distant future and only two players remaining, Intel and IBM. To think AMD can compete in the crossfire of these two giants is a little wishful thinking, don't you think?
[ Rumour ] Marketing policy around the PPC 970 - Lionel - 09:04:48
The architecture of the mother charts PPC 970 64 bits should have like trade description Mach 64. The official logo will represent an apple around which a shuttle spaciale will orbit.
The commercial slogan will be "With Mach 64Technology"
Oldest will remember that Mach 64 was the code name of chip ATI Rage Pro.
To return with the PPC 970, the frequency of the top-of-the-range machines is not yet known. It will be to it on May 15. The forecasts however remain optimistic with at least 2 GHz.
Comments
Originally posted by RBR
By the way, "Algol", once upon a time I used a programming language called that. I think that it was assimilated into one of the Berkeley 'nix variants. Is that where your handle came from?
Regards
Gaack!!! I still have a copy of "Fanget An: An AlgolW Primer" on my bookshelf next to me, from my Freshman year at college... 25 years ago...
according to AMD, the opteron @ 2GHz specs out at:
specint: 1,202
specfp: 1,170
the 970 @ 1.8 GHz:
SPECint - 937
SPECfp - 1051
it has not been published where prescot will land, but I find it really hard to believe that the 970 will be able to top the alternatives in performance, cost, and arrival date. and even if it does, what will apple charge for it with their 28% profit margins... in addition, for lightwave, most of the stuff I need to do is real time... I'm not too concerned with final renders, mostly modeling/animating performance. anyway... just thought I'd add the numbers that have been in my head to the conversation. I really don't care about a 10% disadvantage from a mac workstation in price, or performance. but more than that is stretching it, and arrival date is HUGE.
I don't think that I would necessarily put much faith in AMD right now. Their theoretical SPEC numbers look good compared with the 970, but with the amount of cash AMD is hemorrhaging, they will struggle to keep up with IBM and Intel in scaling their chips. Intel is stuck with 32 bit mode in the consumer arena for the forseeable future and HPaq will do everything in their power to see the Itanic succeed in the pro and server market as the idea behind EPIC came from their own engineers. Gateway is a skeleton of a company and Dell is the only real hope that AMD and its Opteron has. Intel will ensure that Dell stays in the Pentium/Itanium fold by ensuring that HPaq otherwise gets preferred pricing on all Pentium, Itanium and Centrino processors. I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy. IBM itself feels that AMD's days are limited. So the market really comes down to Intel and IBM. Let's not fool ourselves on this one. AMD is in very real trouble here and the Opteron processor will need some serious help in gaining any type of marketshare. Intel will simply price their 32 bit chips so low that AMD won't be able to compete. The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes. Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it. Though the Itanium emulates x86 at 486 speeds, HPaq might just as well put together a box with an Itanium chip along with a Pentium or Centrino processor to run 32 bit x86 applications for those who absolutely need this type of solution. Intel will be more than happy to provide both processors for less than the cost of a single AMD Opteron. Intel's size is the ultimate trump card that AMD just cannot compete against. I am glad Apple chose to partner up with IBM. They are the only company with the size and resources to take on Intel and offer a better solution. It is over for AMD. They should have gotten into the graphics chip market when they had the chance. They might at one time had a chance to compete against ATi. With NVidia in the game and using IBM fabs. AMD can't go here either. It's too bad because they are an innovative company. But business being business, the best solution doesn't always win. So realistically it is the 9x0 series vs. the Itanic series. The Opteron is dead for all practical purposes. The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps. IBM won't be affected by price pressure from Intel as Apple is committed to the PowerPC. Lets be realistic here, Apple and IBM are in a superior position with the PowerPC and OS X. Intel, AMD and Motorola missed the boat. Good companies always find a way to survive. Sony messed up with betamax, but the beta vs. VHS argument is now meaningless. Apple and IBM messed up also. Apple with the Mac OS, and IBM with the PC. Microsoft and Intel double crossed them both. It is only fitting that Apple and IBM finally return the favor. But, I don't want anyone to mistake my comments to mean that MS and Intel will go under. They are too big and have too many resources to see that happen. Even IBM reinvented itself. But it had the size and resources to do so as do Intel and MS.
Still, I will wait for Apple's 970 machines. I'll get to have the best OS on a viable platform that will scale quickly with IBM's commitment to take on Intel with the PowerPC line of chips.
Originally posted by herbivore
[snip] I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy. IBM itself feels that AMD's days are limited. So the market really comes down to Intel and IBM. Let's not fool ourselves on this one. AMD is in very real trouble here and the Opteron processor will need some serious help in gaining any type of marketshare. Intel will simply price their 32 bit chips so low that AMD won't be able to compete. The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes. Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it. Though the Itanium emulates x86 at 486 speeds, HPaq might just as well put together a box with an So realistically it is the 9x0 series vs. the Itanic series. The Opteron is dead for all practical purposes. The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps
[snip]
Even IBM reinvented itself. But it had the size and resources to do so as do Intel and MS.
I totally agree with what you are saying... but next time, may I suggest shorter paragraphs. It makes it easier to read
My take on it [as written in the Thin Client thread] is that as new processes are needed to compete, fewer and fewer companies will be able to fab or do proper R&D, and therefore survive. This effectivly leaves IBM and Intel.
but next time, may I suggest shorter paragraphs
I'm sorry. Didn't mean to be so verbose.
Originally posted by grad student
according to AMD, the opteron @ 2GHz specs out at:
specint: 1,202
specfp: 1,170
the 970 @ 1.8 GHz:
SPECint - 937
SPECfp - 1051
it has not been published where prescot will land, but I find it really hard to believe that the 970 will be able to top the alternatives in performance, cost, and arrival date. and even if it does, what will apple charge for it with their 28% profit margins... in addition, for lightwave, most of the stuff I need to do is real time... I'm not too concerned with final renders, mostly modeling/animating performance. anyway... just thought I'd add the numbers that have been in my head to the conversation. I really don't care about a 10% disadvantage from a mac workstation in price, or performance. but more than that is stretching it, and arrival date is HUGE.
At equal speed, the 970 has similar performance than the opteron, at a cost of a smaller chip with a smaller cache (512 K versus 1024 K) and less heat issue. Add a WMX engine, much more efficient than the various SIMD units (MMX, 3DNow, SSE, perhaps SSE2) of the Opteron, and you have a winner. Consider also the presence of a top end front side bus, and you bet that the chip of IBM may be better.
Consider also that the management of the 32 bit mode of the Opteron is less efficient than the one of the PPC 970, and you will be able to guess that the new IBM chip appear to have a brightest future.
Originally posted by Powerdoc
At equal speed, the 970 has similar performance than the opteron, at a cost of a smaller chip with a smaller cache (512 K versus 1024 K) and less heat issue. Add a WMX engine, much more efficient than the various SIMD units (MMX, 3DNow, SSE, perhaps SSE2) of the Opteron, and you have a winner.
That's more what I was trying to say. (That's _VMX_ == altivec == velocity engine) It extends to duals as well, where dual hammers or dual Xeons or whatever start costing serious money -> price parity.
In addition, I can see that 1051 is less than 1170ish. But it's not 1/5th of 1170, nor is it 1/3 of 1170... it's within the 'benchmarking error' area. 10-15%ish. I'm not a lightwave guy, but the historical reason reported for poor Mac performance on lightwave is way sub-par FPU performance. YMMV of course.
If you have to buy, you have to buy. I'm holding my machine together with duct tape, baling wire, and sweat until the 970's available is all I was implying.
The situation is much more clean and clear with Apple and the IBM 970. The 64 bit PPC design makes the 64 bit transition as simple as possible.
Originally posted by herbivore
I'm sorry. Didn't mean to be so verbose.
Verbose is OK, not using the return key is not.
Verbose is OK, not using the return key is not.
I guess paragraphs would make things easier. I do tend to go into great detail as my PC using colleagues are very difficult to reason with at times.
I'll try to organize things better and into paragraphs. Right now I am dealing with PC geeks who keep telling me that the Opteron is the future. They love to point out that it has better numbers than the 970 and how innovative AMD is. They won't listen to any logic.
In my emotional state of mind, I got carried away and did not take the time to organize my thoughts better.
In any case, HPaq's dedication to the Itanium spells certain death for the Opteron. Dell would be crazy to touch the Opteron as it will likely upset the folks at Intel who would immediately ensure that HPaq gets very preferred pricing on Intel chips. Without a compelling reason to move to 64 bit chips, Dell would lose out as a price leader on 32 bit chip computing where the market will remain for the near future. There is no way that AMD can price the Opteron low enough to compete with Intel's chips either. At least, and make a profit.
The game is over for AMD.
Originally posted by herbivore
. . . In my emotional state of mind, I got carried away and did not take the time to organize my thoughts better. . .
Hey, I believe your thoughts were organized fine, sequentially. It helps the reader when each new idea or issue has its own paragraph however.
The 970 chip is, apparently, viewed as just a transitional chip. The 980 is supposed to be superior in performance in comparison the the 970 and is supposed to consume less power and generate less heat and (being on a smaller die) cost less to produce. It is only a guess, but presumably Apple will also move to the 980 as it becomes available for all of the above reasons.
One can only hope that Apple will have done enough testing to get things right the first time on this one. They don't need any more "oops!" machines.
One thing that I have not seen discussed about the new machines is how well they might work with Real PCs soon (everyone hopes) to be released unix based Windows emulator. You would hope that Apple would have been working very closely with them on this. Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination.
One thing that I have not seen discussed about the new machines is how well they might work with Real PCs soon (everyone hopes) to be released unix based Windows emulator. You would hope that Apple would have been working very closely with them on this. Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination
Have you read this?
A source close to Apple recently relayed information to oscast suggesting that Apple had intended to use emulation software to show the power and performance of its upcoming hardware. This individual stated that Windows software running in emulation performed certain instructions significantly faster than what a hard PC was able to do.
Another (separate) source told oscast that a key reason for Microsoft's acquisition of Connectix's PC emulator, was to "discontinue development of a version of [Virtual PC] that would take advantage of two unreleased hardware technologies that would significantly boost the software'sperformance."
Now read this
Originally posted by RBR
Then again, Apple could simply buy them and incorporate it into the OS. A very high degree of compatibility and good speed would be a nice combination.
I don't think Apple will ever incorporate windows compatibility into the os. Too dangerous. Many companies could stop developing for osx if there was a way too run windows apps without problems on the mac.
Having PC-emulation software on the computer is very useful and I think Apple would help RealPC in every way they can.
Originally posted by COS
Have you read this?
[I]A source close to Apple recently relayed information to oscast suggesting that Apple had intended to use emulation software to show the power and performance of its upcoming hardware. This individual stated that Windows software running in emulation performed certain instructions significantly faster than what a hard PC was able to do.
Yeah, right...
Originally posted by herbivore
I don't think that I would necessarily put much faith in AMD right now. Their theoretical SPEC numbers look good compared with the 970, but with the amount of cash AMD is hemorrhaging, they will struggle to keep up with IBM and Intel in scaling their chips.
Actually, though AMD is making a loss they are in apparently good financial position:
http://www.theinquirer.net/?article=8982
Especially the last paragraph; "AMD could be considered to be cash rich..."
How is it so? Well how is it that Apple has a reported $4 billion in cash after making hugh losses in the '90s and has only been making modest profits since then?
AMD also just signed up as a partner with IBM to work on process tech together. I presume AMD will also get IBM to fab some of their chips as part of this deal since IBM is looking for work to fill its fabs.
Intel is stuck with 32 bit [...] HPaq [...] to see the Itanic succeed [...] Gateway is a skeleton [...] Dell is the only real hope that AMD and its Opteron has. Intel will ensure that Dell stays in the Pentium/Itanium fold by ensuring that HPaq otherwise gets preferred pricing on all Pentium, Itanium and Centrino processors.
IBM is not small player in x86 and are reportedly going in with both hands and feet using Opteron chips. Dell is also a likely customer, contrary to what you think.
I really see the Opteron as a dead duck and a last gasp effort by AMD to maintain market relevancy.
A short sighted statement that might only have strength if AMD had started development last year as their market share began to fall again. However, AMD started work on the K8 years ago, when their market share was going up so it is not a last gasp effort as you say. They company does depend on its success but all reports say it will be very competetive with the P4 and Xeon chips.
IBM itself feels that AMD's days are limited.
No they don't of they wouldn't be planning to use their chips or have signed a partnership agreement.
Intel will simply price their 32 bit chips so low that AMD won't be able to compete.
No so. AMD's Opteron and Athlon 64 dies are small. On the same process the A64 is small than a P4 and can hence be cheaper to produce. The Opteron is smaller than an Itanic by far so it can be much cheaper - price performance wise it will hit Itanic for six.
The Opteron might gain a small percentage of the market share with a few pro users who need 64 bit power with backwards 32 bit compatibility, but it is a small market at this point. Intel can mercilessly cut AMD to ribbons which is happening in front of our very eyes.
No they can't. AMD has much room to spare with their Barton core and it competes well with the P4. If A64 has some startup woes they will introduce faster AthlonXPs to stay in touch with Intel. The big point is, and the same goes for the 970, it is not so much 'backwards compatible' with 32 bit but natively 32 bit with native 64 bit capability too.
Pro applications that need a 64 bit processor will be ported to Itanium as Intel and HPaq will see to it.
And to Opteron. IBM's DB2, Oracle, etc. No small fry in 64-bit land.
Intel will be more than happy to provide both [a P4 and Itanic] for less than the cost of a single AMD Opteron.
Absolutely not true at all. See above. AMD has the upper hand with respect to pricing. Itanics cost thousands of dollars! See this article:
http://www.theinquirer.net/?article=8900
"The distributor price for the Model 240 is $275, for the 242 $670..."
Pentiums 4 don't even sell for that price.
http://www.theinquirer.net/?article=8907
"A 3GHz PENTIUM 4 for a Canterwood will cost $415..."
I am glad Apple chose to partner up with IBM.
So am I. I think the 970 will be excellent for Apple and I think that IBM will follow on with derivatives of the 970 and then the Power5 and really give them the hardware to match OS X.
But business being business, the best solution doesn't always win.
This is an argument against the 970 as much as the Opteron. Neither currently have market share. Both perform well (or are expected to) against competitors.
So realistically it is the 9x0 series vs. the Itanic series.
No, it's the Power series vs the Itanic. Those two families are aimed at different markets than the 970 and it's successors.
The Prescott has no realistic chance of competing either. It will be at a serious disadvantage in a laptop and won't be able to run 64 bit pro apps.
The Prescott is starting on 90 nm and will come in a mobile version but the Pentium M is Intel's real mobile chip anyway.
Still, I will wait for Apple's 970 machines.
Yes, so will I. I'm sick of what MS is trying to force me to do via the Window's platform and will become a switcher as soon as I decide to upgrade my Athlon 1.2 GHz machine I got in 2000 and 2001. I'm planning a PowerBook too - a 970 PB would be nice.
I'll get to have the best OS on a viable platform that will scale quickly with IBM's commitment to take on Intel with the PowerPC line of chips.
Yes but the competition will be:
970 vs Pentium4 family and
Power vs Itanium
I still think you underestimate AMD and the 'hammer' line.
MM
I still think you underestimate AMD and the 'hammer' line.
MM
I guess we shall see.
As I said, I think AMD has done an excellent job with this chip. It is not that I am underestimating them. I think most people are underestimating Intel. I just don't see AMD being able to compete with a company the size of Intel for any length of time.
Intel can use its size and economies of scale to produce chips more inexpensively than AMD. Just because they sell the Itanium for an obscene price, doesn't mean that they couldn't change gears and lower prices, slashing margins and putting intense heat on AMD.
Even with their huge investment in the Itanium, Intel is still profitable. Selling the Itanium chips at low prices means that it takes longer to recover their investment, but it is a luxury that Intel has.
Not only that, AMD is going against IBM's Power series with the Opteron, and the 9x0 series with the Athlon64. Again, I am not trying to underestimate AMD, but I just don't see them staying with IBM in the long run.
The upcoming Power5 will simply be too much for AMD. Intel, however, should be able to able to engineer and fabricate a version of the Itanium that could compete.
Besides, AMD will be dependent on Microsoft in order to succeed. Opteron really has no advantage in the Linux server market over Itanium, other than price. The chip does have an advantage in the x86 Windows markets. However, I am not sure that this is enough to overcome HP's full dedication, and Dell's strong bias, for Intel. DB2 and Oracle on Linux are just not enough. As I recall, DB2 and Oracle are also fully committed to the Itanium platform also.
I am guessing that the Opteron will face stiff competition with the Itanium in the server market. There will be relatively little software on both platforms, however. The desktop version of the Opteron processor is also not slated for release for perhaps another 6 months.
In the 64 bit desktop market, this chip has little chance against OS X on the 9x0 series. Apple will develop a slew of 64 bit applications for the desktop and also laptop. Most Windows developers will remain firmly in the 32 bit arena for that is where most of the legacy hardware is. I'm not sure that x86 developers will be able to compete with Apple in these types of apps, either.
At 80 to 90 watts of heat dissipation, we can forget about putting one of these AMD beasts into a laptop. They won't be able to compete with the Banias chip, much less the 970.
I just don't see Dell giving up its preferred pricing with Intel to take a gamble on the Opteron. They will certainly lose in the laptop market with HP getting preferred pricing on Banias chips.
No matter how I analyze the situation, I just don't see much hope for AMD. If Apple were taken out of the picture, AMD might have a chance. With Apple in the picture, AMD gets squeezed in a big way. 64 bit Windows and x86 apps. won't effectively compete against OS X for the desktop. That is AMD's only real chance as IBM and Intel are going to really turn up the heat for the high dollar server market.
I'm sorry, I see AMD running out of cash in the not too distant future and only two players remaining, Intel and IBM. To think AMD can compete in the crossfire of these two giants is a little wishful thinking, don't you think?
From Macbidouille (bottom)
[ Rumour ] Marketing policy around the PPC 970 - Lionel - 09:04:48
The architecture of the mother charts PPC 970 64 bits should have like trade description Mach 64. The official logo will represent an apple around which a shuttle spaciale will orbit.
The commercial slogan will be "With Mach 64Technology"
Oldest will remember that Mach 64 was the code name of chip ATI Rage Pro.
To return with the PPC 970, the frequency of the top-of-the-range machines is not yet known. It will be to it on May 15. The forecasts however remain optimistic with at least 2 GHz.