But what's the alternative, the G4? I'll take the 970 any day.</strong><hr></blockquote>
Absolutly! And besides, even if its not faster than these chips, at least it's in the ballpark. I just hope there are no "gotchas" that delay the release of it.
Can someone explain to me how the regular P4 can beat the P4 Xeon in both SPECint and SPECfp? And it's by no small margin either.
I've heard that the compiler Intel use for doing SPEC-marks is highly optimized just for providing impressive benchmarking score. To what extent have the done so, if all?
And.. I fore one have no problems with coming fifth in fp-performance behind likes of Sparcs, Power4s, Itaniums and Alphas. Tell me if I'm wrong but single one of those processors cost about as much or even more than a complete Power Mac.
I could not care less about SPEC. Back in 1999 did any PC buyers worry that the Pentium was beaten by th G4 in RC-5 scores? Any mac users out there switching to Intel to get better spec marks <img src="graemlins/lol.gif" border="0" alt="[Laughing]" />
If a Mac with IBM970 gets higher DOOM III frame rates than any Intel PC and compress mp3 and video files faster who cares about spec, RC-5 or even BLAST
If 3 GHz P4 boxes render videos way faster than dual 1.25 GHz G4s this hurt Apple far more than any spec marks going this or that way.
The real test of any system is widely used power demanding applications. If Apple will use the IBM970 web redering, java applets, Office and various compressing, decoding AV stuff as well as games will be more important than syntetic benchmarks or even scores for various projects in distributed computing!
With new higher quality video standards coming soon bus speed and CPU horse powers will be in sersioud demand...
I'm going to be uncharacteristically optimistic and say that we'll see a PPC970 based mac at MWSF, even if they can't ship anything with it for a month or more, we'll see it, and be able to order it, and that the rest is a smokescreen designed to keep pathetic G4 offerings saleable through the X-mas season.
If the 970 arrives, it will bench in the ballpark of a modern CPU, but have the FSB to set altivec and MP configs free, that should make for good performance.
Also, the whole point of the GPuL is that the chips are relatively small, cool and cheap, for the performance that they provide. PPC970 based macs should not cost a penny more than G4 based machines. In fact, as laptops have currently made a slight move down the price ladder, desktops should do the same, probably with the death of the CRt iMac and the subsequent downward adjustment of each line-up's price.
I don't know about the 970, but I am also strangely optimistic about MWSF bringing something interesting. Perhaps it's because there has been nothing genuinely interesting in a long time. Or perhaps it's the two glasses of wine with lunch. <img src="graemlins/embarrassed.gif" border="0" alt="[Embarrassed]" />
<strong>I'm going to be uncharacteristically optimistic and say that we'll see a PPC970 based mac at MWSF, even if they can't ship anything with it for a month or more, we'll see it, and be able to order it, and that the rest is a smokescreen designed to keep pathetic G4 offerings saleable through the X-mas season.
</strong><hr></blockquote>
Why set yourself up for disappointment? They have said 2H '03 and there is no reason to disbelieve them.
[quote]<strong>
If the 970 arrives, it will bench in the ballpark of a modern CPU, but have the FSB to set altivec and MP configs free, that should make for good performance.
Also, the whole point of the GPuL is that the chips are relatively small, cool and cheap, for the performance that they provide. PPC970 based macs should not cost a penny more than G4 based machines. In fact, as laptops have currently made a slight move down the price ladder, desktops should do the same, probably with the death of the CRt iMac and the subsequent downward adjustment of each line-up's price.</strong><hr></blockquote>
This I agree with for the most part. They may eat into Apple's margins for a while because the high speed logic boards and chipsets are going to be more expensive, even if the 970 itself isn't.
It's me again, just pointing out that Hannibal has given us some information on when we'll see the 2nd half of his write-up on the IBM 970. Follow the link to know more
I don't know about the 970, but I am also strangely optimistic about MWSF bringing something interesting. </strong><hr></blockquote>
I have to add to that some rumors from macbidouille: they say that, according to an anonymous but very reliable source, Apple is going to move to 64bit CPU's in the next months and that the supplier would be....(drums please)...AMD and not IBM, since the former has right now a superior production capacity. So now the question is: switch to x86 or PPC from AMD?
Personally, I don't wait dramatic announcements in MWSF, only optimism of statistical origin could be justified at this moment, as Belle has pointed out.
[quote] I have to add to that some rumors from macbidouille: they say that, according to an anonymous but very reliable source, Apple is going to move to 64bit CPU's in the next months and that the supplier would be....(drums please)...AMD and not IBM, since the former has right now a superior production capacity. So now the question is: switch to x86 or PPC from AMD? <hr></blockquote>
If the AMD is radically different than the PPC, the cost of migration both in the migration itself and the loss of old applications is very substantial.
This would only make sense if A. The AMD is very very god. B. The IBM 970 is a POS that never will compete well.
Apple would also end up as AMD: in head to head competetion with a much larger company. The AMD CPUs are directly comparable with the Intel ones. If Apple design a box using IBM HD, nVIDIA GPU and a AMD CPU and it runns OS X applications at a certain speed and the same hardware in Dell box and WinXP does it better, there is not AltiVec driven PS filter to hide behind, nothing about comapring Apples and Oranges <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
To stay in the race Apple as everybode else has to work with back up plans. When the 68040 got beaten by the 486 and the 68050 and 68060 they migrated to PPC. Now withe the G3 at half the clock speed of the celeron and the G4 at less than half of the P4, they are looking for ways out, nota bene, plurality. They will work on different options out.
Are you feeling OK this morning, Matsu?</strong><hr></blockquote>Matsu's motivation is to give himself another year of complaining: If he says "Apple is going to release the 970 Mac in January," he can bitch and moan when they don't. He's just investing in the future.
Comments
<strong>It comes to something when we are desperately cheering a processor that appears to come second. </strong><hr></blockquote>
If the 970 came out today it would be 2nd in SPEC INT and 5th in SPEC FP.
<a href="http://www.aceshardware.com/SPECmine/top.jsp" target="_blank">http://www.aceshardware.com/SPECmine/top.jsp</a>
But what's the alternative, the G4? I'll take the 970 any day.
<strong>
If the 970 came out today it would be 2nd in SPEC INT and 5th in SPEC FP.
<a href="http://www.aceshardware.com/SPECmine/top.jsp" target="_blank">http://www.aceshardware.com/SPECmine/top.jsp</a>
But what's the alternative, the G4? I'll take the 970 any day.</strong><hr></blockquote>
Absolutly! And besides, even if its not faster than these chips, at least it's in the ballpark. I just hope there are no "gotchas" that delay the release of it.
<strong><a href="http://www.aceshardware.com/SPECmine/top.jsp" target="_blank">http://www.aceshardware.com/SPECmine/top.jsp</a></strong><hr></blockquote>
Can someone explain to me how the regular P4 can beat the P4 Xeon in both SPECint and SPECfp? And it's by no small margin either.
I've heard that the compiler Intel use for doing SPEC-marks is highly optimized just for providing impressive benchmarking score. To what extent have the done so, if all?
And.. I fore one have no problems with coming fifth in fp-performance behind likes of Sparcs, Power4s, Itaniums and Alphas. Tell me if I'm wrong but single one of those processors cost about as much or even more than a complete Power Mac.
If a Mac with IBM970 gets higher DOOM III frame rates than any Intel PC and compress mp3 and video files faster who cares about spec, RC-5 or even BLAST
If 3 GHz P4 boxes render videos way faster than dual 1.25 GHz G4s this hurt Apple far more than any spec marks going this or that way.
The real test of any system is widely used power demanding applications. If Apple will use the IBM970 web redering, java applets, Office and various compressing, decoding AV stuff as well as games will be more important than syntetic benchmarks or even scores for various projects in distributed computing!
With new higher quality video standards coming soon bus speed and CPU horse powers will be in sersioud demand...
If the 970 arrives, it will bench in the ballpark of a modern CPU, but have the FSB to set altivec and MP configs free, that should make for good performance.
Also, the whole point of the GPuL is that the chips are relatively small, cool and cheap, for the performance that they provide. PPC970 based macs should not cost a penny more than G4 based machines. In fact, as laptops have currently made a slight move down the price ladder, desktops should do the same, probably with the death of the CRt iMac and the subsequent downward adjustment of each line-up's price.
Are you feeling OK this morning, Matsu?
[ 11-21-2002: Message edited by: Amorph ]</p>
I don't know about the 970, but I am also strangely optimistic about MWSF bringing something interesting. Perhaps it's because there has been nothing genuinely interesting in a long time. Or perhaps it's the two glasses of wine with lunch. <img src="graemlins/embarrassed.gif" border="0" alt="[Embarrassed]" />
<strong>I'm going to be uncharacteristically optimistic and say that we'll see a PPC970 based mac at MWSF, even if they can't ship anything with it for a month or more, we'll see it, and be able to order it, and that the rest is a smokescreen designed to keep pathetic G4 offerings saleable through the X-mas season.
</strong><hr></blockquote>
Why set yourself up for disappointment? They have said 2H '03 and there is no reason to disbelieve them.
[quote]<strong>
If the 970 arrives, it will bench in the ballpark of a modern CPU, but have the FSB to set altivec and MP configs free, that should make for good performance.
Also, the whole point of the GPuL is that the chips are relatively small, cool and cheap, for the performance that they provide. PPC970 based macs should not cost a penny more than G4 based machines. In fact, as laptops have currently made a slight move down the price ladder, desktops should do the same, probably with the death of the CRt iMac and the subsequent downward adjustment of each line-up's price.</strong><hr></blockquote>
This I agree with for the most part. They may eat into Apple's margins for a while because the high speed logic boards and chipsets are going to be more expensive, even if the 970 itself isn't.
<strong>high speed </strong><hr></blockquote>
<homer voice>mhmmm... high speeed....</homer voice>
<a href="http://arstechnica.com/archive/news/1037928794.html" target="_blank">here</a>
He says he'll finish it no later than 17th December.
<strong>
Can someone explain to me how the regular P4 can beat the P4 Xeon in both SPECint and SPECfp? And it's by no small margin either.
</strong><hr></blockquote>
I'm guessing the difference is caused by the chipset. Apparently the Xeon chipsets are more complex and thus slower.
<strong>
I don't know about the 970, but I am also strangely optimistic about MWSF bringing something interesting. </strong><hr></blockquote>
I have to add to that some rumors from macbidouille: they say that, according to an anonymous but very reliable source, Apple is going to move to 64bit CPU's in the next months and that the supplier would be....(drums please)...AMD and not IBM, since the former has right now a superior production capacity. So now the question is: switch to x86 or PPC from AMD?
Personally, I don't wait dramatic announcements in MWSF, only optimism of statistical origin could be justified at this moment, as Belle has pointed out.
If the AMD is radically different than the PPC, the cost of migration both in the migration itself and the loss of old applications is very substantial.
This would only make sense if A. The AMD is very very god. B. The IBM 970 is a POS that never will compete well.
Apple would also end up as AMD: in head to head competetion with a much larger company. The AMD CPUs are directly comparable with the Intel ones. If Apple design a box using IBM HD, nVIDIA GPU and a AMD CPU and it runns OS X applications at a certain speed and the same hardware in Dell box and WinXP does it better, there is not AltiVec driven PS filter to hide behind, nothing about comapring Apples and Oranges <img src="graemlins/hmmm.gif" border="0" alt="[Hmmm]" />
To stay in the race Apple as everybode else has to work with back up plans. When the 68040 got beaten by the 486 and the 68050 and 68060 they migrated to PPC. Now withe the G3 at half the clock speed of the celeron and the G4 at less than half of the P4, they are looking for ways out, nota bene, plurality. They will work on different options out.
<strong> <img src="graemlins/surprised.gif" border="0" alt="[Surprised]" />
Are you feeling OK this morning, Matsu?</strong><hr></blockquote>Matsu's motivation is to give himself another year of complaining: If he says "Apple is going to release the 970 Mac in January," he can bitch and moan when they don't. He's just investing in the future.
<strong>I'm guessing the difference is caused by the chipset. Apparently the Xeon chipsets are more complex and thus slower.</strong><hr></blockquote>
P4 : PC1066 RDRAM, FSB533
Xeon: PC800 ECC RDRAM, FSB400
RAM and FSB is much slower for the Xeon. Take a look at html files.
BTW a P4 2800MHz is not faster than an Athlon 2250MHz if both are runnig on PC2700 SDRAM.
<strong> ... He's just investing in the future.</strong><hr></blockquote>
I like to plan ahead.