970 Motherboard and Chip Delivery?

[maxvamp]

Aiden

Aiden, I would like to note two things....

1.) you jumped all over my posts because they mentioned that there would be performance increases in the 970 and assumed that it was solely because the 970 proc was 64-bit. If you go back in this thread, you will see that my original post had the following, which I saud before the ars article was announced :

All, I understand that most of performance increase to be gained from the 970 is from basic architectural design changes, and most apps will not take advantage from the increased memory addressing available, but people keep forgetting that computer security has a bright future, and we will ( with the war on terrorism ) see encryption show up in more and more components of the average computer system.

I did, however try to explain that there will be apps that will be able to make use of this processor. I even tried to give a few examples of some I thought might. Your basic stance so far has been that 64-bits will not, nor ever make a difference. and quite frankly, I do not think that is right. Even on the disk example I give, I believe one application that will take advantage of all of this is a good backup program. They have databases to deal with and do a ton of I/O calculations, along with ( often ) software compression. There is a lot of math there... I don't know how this processor will not benifit in this area.

Long and short of this dialog is... We must agree to disagree. I make my stance that this proc will be faster, as I stated before. If you choose to read into it that I am saying this is because it is 64 bits, so be it ( but I am not ).

Thanks again,

Max

 

[AidenShaw]

Re: Aiden

Originally posted by maxvamp
1.) you jumped all over my posts because they mentioned that there would be performance increases in the 970 and assumed that it was solely because the 970 proc was 64-bit.

Actually, I thought that I "jumped all over" your inconsistency in comparing dual PPC systems with single P4 systems, and then criticizing XP Home for not supporting duals.

Most of the 64-bit comments were other discussions....

 

[jettredmont]

Originally posted by IJ Reilly
For a guy who claims he can't say nothin', this powerjack says quite a bit -- which all by itself seems kind of suspect. In the end all he really is saying is that Foxconn is not the contract manufacturer, somebody else is. As if that really matters.

Yes, I agree that MacWhispers says a bit too much if he's really being watched closely. If any of the information he is publishing is part of his NDA, even if he's also gotten it from other sources, he's liable to be brought into court for an NDA violation (and, at the very least, designated personna non gratis at Apple which will kill the likelihood of him working on Apple-related products pretty quickly).

On the other hand, if he is right, the implication there is that since FoxConn is not using the 970 chips then the 20k/40k/40k "arriving at Foxconn" is a complete fabrication. There's no way you could get information about how many units are arriving at an asembly plant and not get the plant itself right.

We'll see soon enough.

 

[Mr. Anderson]

After going through the original fiasco with Apple on the introduction of the first G4s over three years ago - I'm thinking I'm going to wait until the second version of the 970 comes out before I make a new purchase - even though I'm still using that original G4....

The potential here is fantastic, its going to be very hard not to buy one immediately.

D

 

[dongmin]

Re: Went to Foxconn's site and found this

Originally posted by leo
Went to Foxconn's site and found this:

http://www.foxconn.com/products/alphabet.asp?fmletter=M

What's an "Apple Monster Bus" (Customer Special)?

*feeds the fire*

leo


p.s.: At least Apple has been working with Foxconn in the past to some extent. Some of the connectors in my b/w G3 are labelled "Foxconn". I know, it's just connectors...

Great find! Could this be the mysterious 'Apple Processor Interconnect Bus' that was rumored a while back? Could be huge.

If you do a google search for hon hai and apple, you'll see that they've done quite a lot of work for Apple. Hon Hai (Foxconn's their trademarked named) produced the B&W cases for Apple.

 

[cc bcc]

Foxconn also has this one:

Cable Assembly, Modubus____Harness, Apple Monster Bus____1.27mm [0.050in] Cable Flat

link

Can anyone tell us what this means?

 

[maxvamp]

XP Home

Actually, I thought that I "jumped all over" your inconsistency in comparing dual PPC systems with single P4 systems, and then criticizing XP Home for not supporting duals.

Ok. Quick question then.... Will XP Home support a 3GHz P4 with hyperthreading in HT mode ( i.e. the system looks like it has multiple procs)? This is the bets way to wring extra performance out of those chips.

Max.

 

[ktlx]

Re: XP Home

Originally posted by maxvamp
Ok. Quick question then.... Will XP Home support a 3GHz P4 with hyperthreading in HT mode ( i.e. the system looks like it has multiple procs)?

Yes. Windows XP Home will support two logical processors but not two physical processors. So as long as the motherboard supports it and the feature is turned on, Windows XP Home will recognize the HT-enabled Pentium 4s as being two logical processors.

 

[silvergunuk]

dual 970s?

Im no tech geek so i was wondering if some one could help me out. As a Power 4 is a dual core and a 970 is a single core....couldn't a dual 970 set up act as a dual core chip? it could explain that dodgy bryce render.

 

[Rincewind42]

Re: dual 970s?

Originally posted by silvergunuk
Im no tech geek so i was wondering if some one could help me out. As a Power 4 is a dual core and a 970 is a single core....couldn't a dual 970 set up act as a dual core chip? it could explain that dodgy bryce render.

There is a small but significant difference between a Dual Core chip and a Dual CPU system. As far as the OS & Software is concerned they are identical, but from an engineering standpoint there are significant advantages to a Dual Core CPU design (and to a lesser degree advantages to a Dual CPU design). The Bryce rendering benchmarks post by MacBid could be either a beta of Bryce 6 that does take advantage of multiple CPUs/Cores or just made up, but either way Bryce 5 doesn't take advantage of multiple CPUS.

 

[silvergunuk]

Thanks for the explanation

 

[anonmac]

bryce 6 beta...

while there are people claiming that the bryce 6 beta does not exist, if you search google, you will find several people who claim to be in on the 'beta test'. if you're not one of them, who can say the truth is "this"?

 

[maxvamp]

RE: XP Home

INNNTERESTINK....

Max

 

[jettredmont]

Originally posted by Rocketman
I do not know if the 970 has powersaver features built-in but the OS does.

Powersaving works in one of two (preferably both) ways:

1) Shut down internal components (modem, HD, monitor) when not in use, and support hibernate (suspend to disk) and/or suspend (shut off everything but memory and monitor for wake).

2) Lower the voltage (and hence frequency and hence performance) of the CPU, turn off specific unused units on the processor itself.

(1) is implemented in the OS, and in large part does not need anything special from the CPU (although it does require that the firmware or bios support turning off subsystems, but that's been pretty universal since '98 or so). (2), however, requires CPU support.

I can't say for sure, but others here are convinced that the 970 does not support anything in group (2) above. Essentially, then, you will get the same design headaches putting the 970 into a laptop as you get putting a non-mobile Intel design into a laptop (relatively low battery life, heat generation a problem, fan noise).

On the other hand, there have been quite a few notebooks produced, even in the last couple of years, with non-mobile processors, and they have sold fairly well. I don't think this rules out a low-frequency (1.2GHz or less) 970 in laptops.

 

[jettredmont]

Re: About AltiVec...

Originally posted by P-Worm
Ok, so AltiVec is like a 64 bit emulator right? So why do we need it in a 64 bit machine. Will it double the emulation to 128? Or am I just getting my numbers screwed up?

Altivec does not emulate 64-bit integer operations (nor 128-bit!). It handles multiple 32-bit (or 16- or 8-bit) operations in a single instruction instead of in multiple (4-16) operations.

Note, also, that the Altivec registers are not any wider in the 970 than they are in the G4. They will of course benefit from the increased FSB bandwidth, but they'll still work on 128 bits at a time.

 

[AidenShaw]

XP Professional does 4 processors

Originally posted by ktlx
Windows XP Home will recognize the HT-enabled Pentium 4s as being two logical processors.

And Windows XP Pro recognizes a dual HT Xeon as 4 logical processors, even though XP Pro is restricted to 2 physical processors.

To wit:
OS Name Microsoft Windows XP Professional
Version 5.1.2600 Service Pack 1 Build 2600
OS Manufacturer Microsoft Corporation
System Name COCO
Manufacturer Dell Computer Corporation
System Model Precision WorkStation 650
System Type X86-based PC
Processor x86 Family 15 Model 2 Stepping 7 GenuineIntel ~3055 Mhz
Processor x86 Family 15 Model 2 Stepping 7 GenuineIntel ~3056 Mhz
Processor x86 Family 15 Model 2 Stepping 7 GenuineIntel ~3056 Mhz
Processor x86 Family 15 Model 2 Stepping 7 GenuineIntel ~3056 Mhz
BIOS Version/Date Dell Computer Corporation A00, 11/7/2002
SMBIOS Version 2.3
Windows Directory S:\WINDOWS
System Directory S:\WINDOWS\System32
Boot Device \Device\HarddiskVolume2
Locale United States
Hardware Abstraction Layer = "5.1.2600.1106 (xpsp1.020828-1920)"
User Name COCO\aiden
Time Zone Pacific Standard Time
Total Physical Memory 2,048.00 MB
Available Physical Memory 1.70 GB
Total Virtual Memory 5.85 GB
Available Virtual Memory 5.41 GB
Page File Space 3.85 GB
Page File S:\pagefile.sys

 

[BaghdadBob]

So can someone tell me how it helps multiple processors for Windows to treat one processor as two? I don't get it. Are you saying that Windows XP treating one processor as two enables it to treat two processors as two, or are they then four?

 

[adamfilip]

any chance of Serial ATA interface on the new 970 MB?

i would like to have a Western Digital 36 gig Raptor SATA drive

and then a larger slower drive for regular backups and general archiving like 250 gig or so

 

[AidenShaw]

extra power, nearly free

Originally posted by BaghdadBob
So can someone tell me how it helps multiple processors for Windows to treat one processor as two? I don't get it. Are you saying that Windows XP treating one processor as two enables it to treat two processors as two, or are they then four?

Yes, XP shows one processor as two, and two processors as four - when those processors have Hyperthreading.

Hyperthreading is a cross between a single CPU chip and a dual-core (2 CPU) chip. It has some of the circuitry of 2 CPUs, but not 2 complete CPUs. They are called "logical CPUs", so that you have 2 logical CPUs per chip.

If one logical CPU is using the floating point units, the other logical CPU can use the integer units at exactly the same time. This can improve performance.

It'a not double the performance (YMMV), but it's very cheap to add to a chip.

Look at http://www.reseller.co.nz/Reseller/...56C780072D42F?OpenDocument&More=Product+News, or the Intel pages at http://developer.intel.com/homepage/land/hyperthreading.htm.

 

[jettredmont]

Originally posted by BaghdadBob
So can someone tell me how it helps multiple processors for Windows to treat one processor as two? I don't get it. Are you saying that Windows XP treating one processor as two enables it to treat two processors as two, or are they then four?

If Windows were to treat a HT processor as a single processor instead of dual processors all benefits of HT would be erased. Windows treats a HT-enabled processor as two, thus keeping two threads active on the processor at once.

HT-CPUs are not the same as dual core CPUs nor of course as dual CPUs, because they "share" most of the CPU between what Windows would consider the two processors. This is why Intel's implementation gets about 25% better performance with HT instead of 100% better performance (note, however, that IBM's Power5 is said to get much higher benefit, close to 100%, from SMT aka HT ... must be that IBM Special Sauce ...).

 

[maxvamp]

SMP percentage

From what I remember from my OS/2 days, the most one could expect expect from a second processor was about 89% performance of the first. Even this took some IBM trickery. I would say that if someone is claiming 100%, that this would be more of a theory, than a constant statistic. Even the 89% was a peak rating.

BTW, thanks for the headsup on the XP HT info. I stopped keeping up with extreme details after Win2k. XP turned out to be such a piece of work, that I lost some interest in details. Win2k3, so far has been interesting though..

Max.

 

[MoldyJohn]

About those 8 slots...

Per the MacBidouille article of the new mobos ?8 RAM connectors - 4 of which were covered by a Do Not Use sticker? makes some sense. OSX10.3 (64bit) will not come preloaded on the 970 boxes. OSX10.2 (32bit) can only handle, what, 4MB memory? The extra slots are there waiting for 10.3 which will then open up a full 8MBs of memory. Voile!

 

[mim]

Originally posted by adamfilip
any chance of Serial ATA interface on the new 970 MB?

We'll, you'd think there'd be no reason >not< to include one, after all Apple is pretty hot on the video editing market and it's ripe for SATA drives (ie - speed without RAIDing)....
...but this is Apple. Who knows what they'll give us.

 

[trebblekicked]

Originally posted by mim
We'll, you'd think there'd be no reason >not< to include one, after all Apple is pretty hot on the video editing market and it's ripe for SATA drives (ie - speed without RAIDing)....
...but this is Apple. Who knows what they'll give us.

funny. HD is the one aspect of the 970 innerds that hasn't been speculated to death. SATA would be interesting, but i wonder how XRAID fits in with the afore mentioned Xstation rumors. A prosumer (power mac) level 970 would probably still have UATA, with expandability leaning towards FW800 drives/raids, IMO.

 

[mathiasr]

Re: SMP percentage

Originally posted by maxvamp
From what I remember from my OS/2 days, the most one could expect expect from a second processor was about 89% performance of the first. Even this took some IBM trickery. I would say that if someone is claiming 100%, that this would be more of a theory, than a constant statistic. Even the 89% was a peak rating.

The theory is called the Amdahl's Law. In practice you also have to consider how resources outside the CPU, especially memory bandwidth, will be shared.

100% is achieved by the distributed.net client, since it does scarcely access memory outside the L1 caches. The distributed.net client is not representative at all, it also uses AltiVec on a tiny data set for instance; but it does exist.