new 970 mobo info ( macbidouille )

[maradong]

Macbidouille posts the following :


Bon, prenez tout ça avec les plus extrèmes précautions comme à chaque fois.

Les cartes mères bipro devraient avoir en tout 8 emplacements mémoire. Il y en aurait 4 par CPU. Ces emplacements sont dors et déjà compatibles avec les barrettes de 4 Go. On aura donc une capacité mémoire maximale de 32 Go !!
Nous ignorons totalement comment le Mac va se débrouiller avec sa RAM.
Nous ne savons pas si chaque processeur aura sa mémoire allouée avec un éventuel Twin Bank ou si les slots de 4 seront en fait appariés pour faire du Twin Bank entre les deux.

Give me 10 minutes and i ll make you a little translation :D

[maradong]

First of all, here comes the google translation:


Well, take all that with more extrèmes precautions as each time. The mother charts bipro should have in all 8 sites report. There would be of them 4 per CPU These sites are sleep and already compatible with the bars of 4 Go. One will thus have a maximum memory size of 32 Go!! We are unaware of completely how Mac will manage with its RAM. We do not know if each processor will have its memory allocated with possible Twin Bank or if the slots of 4 in fact are paired to make of Twin Bank between the two.

[maradong]

Well, take all that with more precautions as each time. The motherboads bipro should have in all 8 ram slots sites report. There would be of them 4 per CPU These slots already compatible with the bars of 4 Go. One will thus have a maximum memory size of 32 Go!! We are completely unaware of how Mac OS X will manage that amount of RAM.
We do not know if each processor will have its memory allocated with possible Twin Bank or if the slots of 4 in fact are paired to make of Twin Bank between the two.
the translation, a bit changed.

[mathiasr]

4 slots per CPU would mean a cc-NUMA architecture like the Opteron.
I think that Apple will stick with a true SMP system with the 970, on the other hand the PowerPC 980 is rumored to have an integrated memory controller, thus high-end Macs will naturally turn to NUMA, Apple could try to train developers to this new architecture early to get full benefit on the next generation.

Before you ask:
http://lse.sourceforge.net/numa/faq/

NUMA stands for Non-Uniform Memory Access, and cc for Cache Coherent.

This could pave the way to some very high-end HPC systems like SGI does:
http://www.usenix.org/publications/library/proceedings/usenix-nt98/invited_talks/baskett_slides/ppframe.htm

Xserves could be turned into modules like in the Origin 350 for instance (32 CPUs Macs anyone?):
http://www.sgi.com/origin/350/

[ZeeOwl]

Just as an FYI, Here is a human translation from French by your's truly (French is my first language)...

"OK, take this with a major dose of salt, as usual.

The dual-processor motherboards should have 8 memory slots in all. There seems to be 4 per processor. These slots are currently compatible with 4 GiB memory modules. Therefore we should have a maximum memory capacity of 32 GiB !!

We have no clue how the Mac will manage it's RAM. We don't know if each processor will have it's own dedicated Twin Bank memory allocation, or if the two groups of 4 slots will be combined to form one large Twin Bank.

The question now remains: How much will the ultimate PowerMac 970 configuration with 32 GiB of RAM cost?"

[maradong]

thx

[Mr. Anderson]

Originally posted by ZeeOwl
The dual-processor motherboards should have 8 memory slots in all. There seems to be 4 per processor. These slots are currently compatible with 4 GiB memory modules. Therefore we should have a maximum memory capacity of 32 GiB !!

32 Gig? That's huge, and most likely for a server. There are very few applications that would use this much memory.

With the fiber channel and now this, I'm thinking that the board they have is not going to be on every new 970 - but might be for an XServe or XStation (desktop server like they had before the XServe).

D

[ZeeOwl]

Originally posted by Mr. Anderson
32 Gig? That's huge, and most likely for a server. There are very few applications that would use this much memory.

There's a caveat to that. They say "32 GiB maximum memory capacity". That's using 4 GiB modules. Do those even exist yet? As far as I know, the biggest ones currently on the market are 512 MiB. That would give the PowerMac 970 a total of 4 GiB of RAM with today's technology. I do 3D animation in 720 x 480 (DVD) resolution. On average my scenes need about 3 GiB to render finished frames. That would only leave 1 GiB for the OS and other apps. Not huge. My current machines only support 1 GiB of memory, so rendering uses virtual memory heavily at the moment. Having 4 GiB would probably at least double rendering speed. And that's not even taking into account that a PPC 970 is much faster than a G3/G4. :) For people doing HDTV or film rendering, about 18 GiB would be needed, by extrapolating. So 32 GiB don't look so huge now, does it?... ;) This is the market the PowerMac is aimed at: high-end video/audio and scientific. And Apple needs to keep the future in mind. This thing needs to be upgradeable.

[Vlade]

Yes there are 8GB modules, but they cost a whopping 3700 bucks! There are 4GB ones for 191 dollars, and 2 GB ones for 303 dollars (Odd)

[iJon]

isnt there a hardware conflict somewhere that wouldnt allow each processor to have their own bank by themselves. i was talking to a friend about it and he said something that i cant remember, i guess ill have to ask him again.

iJon

[P-Worm]

As Dukestreet pointed out, each rumor we get from this site seems to be more and more outrageous. either these aren't true, or they're not what we think they are.

P-Worm

[tjwett]

Originally posted by Vlade
Yes there are 8GB modules, but they cost a whopping 3700 bucks! There are 4GB ones for 191 dollars, and 2 GB ones for 303 dollars (Odd)

Link Please? The only evidence I've been able to find of a 4gb module in existence is a recently tested Samsung which doesn't look like it will be mass produced or affordable for a long long time. And I find it very hard to believe that a 4gb module would cost $191, I've spent more than that on a 512 SDRAM chip. Please prove me wrong, that would be exciting stuff.

[tjwett]

Originally posted by P-Worm
As Dukestreet pointed out, each rumor we get from this site seems to be more and more outrageous. either these aren't true, or they're not what we think they are.

P-Worm

Yo, seriously. Imagine what everyone will say when WWDC comes and Apple releases a (drumroll).........
Dual 1.8Ghz G4 PowerMac:(

[Mr. Anderson]

Originally posted by ZeeOwl
I do 3D animation in 720 x 480 (DVD) resolution. On average my scenes need about 3 GiB to render finished frames. That would only leave 1 GiB for the OS and other apps. Not huge. For people doing HDTV or film rendering, about 18 GiB would be needed, by extrapolating. So 32 GiB don't look so huge now, does it?... ;) This is the market the PowerMac is aimed at: high-end video/audio and scientific. And Apple needs to keep the future in mind. This thing needs to be upgradeable.

Thats somewhat true - and a real pro machine would be nice. I'm also doing 3D animation on an older machine and I'd be perfectly happy with more RAM. Although I never thought that 32 Gig would be required. And it might be for planning on the future here - which would be nice for a change.

I'm personally looking forward to getting one of the 970s - hoping for an order of magnitude faster rendering...:D

D

[EponymousCow]

.Originally posted by ZeeOwl
There's a caveat to that. They say "32 GiB maximum memory capacity". That's using 4 GiB modules. Do those even exist yet? As far as I know, the biggest ones currently on the market are 512 MiB.

I believe that 1GB modules have just become available. Apple has done this before, making their motherboards compatible with a memory size not yet available. In fact, the MDD powermacs have 4 slots that can use 1GB DIMMs, but the motherboard can only handle 2GB total. Must be limited to 31-bits of addressing. :rolleyes:

[joelc]

I thought my 512 megs was pretty good...*looks sheepishly at Indigo iMac*
This is crazy stuff. How common is what ZeeOwl described? Do people actually need that kind of capacity?

[freundt]

Originally posted by tjwett
Yo, seriously. Imagine what everyone will say when WWDC comes and Apple releases a (drumroll).........
Dual 1.8Ghz G4 PowerMac:(

not too hard to imagine at all...

*loud boo-ing, some scattered yells of "this is *********, and the sounds of lots of feet heading for the door*

oh, wait... I forget, these are MAC fans we are talking about..We will all clap in a nice mannerly way and then come back here and specualte on the pushed back relase date for the 970's and what could have happened. ;)

_f

[Chobit]

I'm just glad that someone's starting to use the correct GiB instead of GB. Its 32 Gibibytes because it is 2^30 bytes, not 10^9. I don't know.. It really doesn't matter too much, but sometimes its annoying not knowing if things (especially harddrives) are being measured in Giga or Gibi bytes. People use Giga (which is supposed to be pronounced "jiga" even though I still say it with a hard g out of habit and people not understanding me) when they mean gibi and it all makes my head hurt!

[Steradian]

I really hope that people aren't dissapointed if all that is announced at WWDC is panther and a few upgrades on normal iApps... We DO need a power mac of this caliber however I doubt that it existists

[Coca-Cola]

http://www.xbitlabs.com/news/memory/display/20030319175250.html

It seems that Samsung and Micron have 4gb memory in the works.

1gb's are available at crucial.

[MacsRgr8]

Ofcourse it's possible and, yeah, take it with a ton of salt (possibly a rumor? :D )
Let's get a Pentium blowing 970 desktop Mac first....

[Nutzoids]

ummmm! lets see. For a 970 with 32GB of RAM...2 words, First Born... Thats what I would spend.

[MacFan25]

I bet apple is wondering where macbidouille is getting all this info.

originally posted by tjwett
Yo, seriously. Imagine what everyone will say when WWDC comes and Apple releases a (drumroll).......
Dual 1.8Ghz G4 PowerMac:(
could the g4 go that fast? :confused:

[t^3]

Originally posted by tjwett
Imagine what everyone will say when WWDC comes and Apple releases a (drumroll).........
Dual 1.8Ghz G4 PowerMac:(
Imagine what kind of cooling would be needed to keep it from burning up! The 7457 isn't scheduled to ship until the end of this year, so any new G4 chip will continue to be the 7455. All 7455's running past 1GHz are running really hot, especially if you look at the size of the heatsinks on the dual 1.4GHz PM's.

[alset]

I'm just sick of MB. They get more unrealistic every day. I bet if everyone stopped paying attention, they would stop making such outrageous claims. Good luck getting everyone to ignore them, though.

MB is quickly becoming my least favorite site on the net.

Dan

[Lazy]

Originally posted by tjwett
Link Please? The only evidence I've been able to find of a 4gb module in existence is a recently tested Samsung which doesn't look like it will be mass produced or affordable for a long long time. And I find it very hard to believe that a 4gb module would cost $191, I've spent more than that on a 512 SDRAM chip. Please prove me wrong, that would be exciting stuff.

Vlade got those prices from pricewatch.com, but not carefully enough. The 2GB for $303 is a kit, meaning multiple modules. In this case, it means 8 256MB modules! That's pretty useless. For months now the smallest size module I'll buy is 512MB. The cheapest actual 2GB module listed is $780.

The 4GB for $191 is not even memory; it's some widget of some sort that's needed for added memory to some kind of server. It works with up to 4GB memory though. :)

You can get 1GB PC2100 modules for $125 though. I think I could possibly see my way clear to scraping by with 4GB for $500 vs. buying 2GB modules for $780.

[cdburrows]

i was pretty optimistic about what might come this summer, after hearing reports from MacB. Now, with a report like this, it is as though MacB. is making fun - yet again leaving the future in the hands of Apple!

They can still pull it off... and at this rate it will be a complete surprise to us all! (considering Apple has really turned out the lights)

Make it a happy summer...please?

[Niknar]

Looks like Apple's planning for the future with this one.

It dosn't sound like an xStation more like a new xServer.

So maybe a new 970 xServer but maybe no Powermac yet.

We should trty not to get our hopes up to much :-)

[Longey Nowze]

Originally posted by Chobit
I'm just glad that someone's starting to use the correct GiB instead of GB. Its 32 Gibibytes because it is 2^30 bytes, not 10^9. I don't know.. It really doesn't matter too much, but sometimes its annoying not knowing if things (especially harddrives) are being measured in Giga or Gibi bytes. People use Giga (which is supposed to be pronounced "jiga" even though I still say it with a hard g out of habit and people not understanding me) when they mean gibi and it all makes my head hurt!

huh? Gibibyte?! so all ram uses GiB? what? when? how?! where? and I'm having a hard time trying to speak the word Gibibyte!

THANK YOU
MaT

[shadowfax]

surely they have at least 2 GB dimms in decent production. i have seen 1GB so-DIMMS coming around soon, and if you can pack that into an so-DIMM, you can definitely fit 2 GB or 4 into a full DIMM.

[Flowbee]

Originally posted by Longey Nowze
huh? Gibibyte?! so all ram uses GiB? what? when? how?! where? and I'm having a hard time trying to speak the word Gibibyte!

THANK YOU
MaT

I'm with you. Gibibyte? I personally don't see that word coming out of my mouth any time in the near future (unless I stumble into a CompUSA after drinking a six-pack of beer). However, I did find some info on the giga vs. gibi:

http://www.romulus2.com/articles/guides/misc/bitsbytes.shtml

[Chobit]

Harddrives and RAM and other data storage devices should be measured in Kibi/Mebi/Gibi/Tebi etc. bytes as they are always in some grouping of a power of 2, not a power of 10. Mhz Ghz etc. will probably never move over to a binary standard as there is no reason to measure clockcycles per second as a power of 2. Data transfer is normally measured in Mega/Giga bytes per second, but I think it will change (or should change) eventually, because if you have a 80 GiB harddrive and can transfer things at 100 MB per second you'll have to convert to binary from base ten in order to find out how long it would take you to transfer something, so it only makes sense to measure those in binary even though it isn't the current correct standard. Of course, since no one really follows the proper standards anyway, none of this really matters.

[Jeff Harrell]

Originally posted by Chobit
Harddrives and RAM and other data storage devices should be measured in Kibi/Mebi/Gibi/Tebi etc. bytes as they are always in some grouping of a power of 2, not a power of 10."And hillbillies prefer to be called 'sons of the soil.' But it's not gonna happen."

(Obligatory Simpsons quote. :D)

[Vlade]

Originally posted by Lazy
Vlade got those prices from pricewatch.com, but not carefully enough. The 2GB for $303 is a kit, meaning multiple modules. In this case, it means 8 256MB modules! That's pretty useless. For months now the smallest size module I'll buy is 512MB. The cheapest actual 2GB module listed is $780.

The 4GB for $191 is not even memory; it's some widget of some sort that's needed for added memory to some kind of server. It works with up to 4GB memory though. :)

You can get 1GB PC2100 modules for $125 though. I think I could possibly see my way clear to scraping by with 4GB for $500 vs. buying 2GB modules for $780.

OOPS! Sorry, Thanks for correcting me. I hate when pricewatch does stuff like that

[shadowfax]

i still don't see why you can't measure hard drives etc. in GB and MB. in the same way that netweorks transfer in bytes or bits per second, a hard drive holds both bytes and bits. it's just how you look at it. and 8 bits isn't really any different than one byte. or whatever the equivalent is.

[job]

Originally posted by alset
I'm just sick of MB. They get more unrealistic every day. I bet if everyone stopped paying attention, they would stop making such outrageous claims. Good luck getting everyone to ignore them, though.

What happens if they're right? ;) :D

Personally I don't think the 970 will be first in Power Macs. Like people have been mentioning, I see them as a prime candidate for Apple servers, rackmounted or otherwise.

[JoeRadar]

Originally posted by Mr. Anderson
32 Gig? That's huge, and most likely for a server. There are very few applications that would use this much memory.

I admit, I cannot see most users needing that much space, but then again, I used to say the same thing about similar disk sizes.

My speculation: With the new machines Apple will come gunning for researchers, including bioinformatics. Imagine the entire human genome in searchable data structures in main memory.

Swap space? We don't need no stinking swap space.

[macmunch]

No, there should be no problem !

Its logical I think, first I thought Apple will have 4 RAM slots paired with another 4 to get TWIN BANKING ....

That meant that they will have 6,4 GB/s which the PPC970 needs but if you have a dual CONFIG the two chips must divede the memory, so each chip would have only a connection with 3.2 GB/s, which is not very good !

So now I understand Apple will make 4 RAM slots for each Chip, but the Salt in this is that you will only have 2 RAM Slots per Chip because this two must be TWINED to reach the 6.4 GB/s, same on the other Chip, so at the end you can support both Chips with its needed full Bandwith !

Now you can also understand why in the rumors is said that on 4 SLOTs is a mark "dont use". This person which saw that board saw only a version of the board with one Chip !!! When you put two in it than you must take the other 4 !

Apple is a member in the HyperTransport Forum, now I know why ! :D

The Ram Banks will be connected to each Chip with Hypertransport to deliver the big Bandwith. And from there they will connect it with a Controller Chip.

Maybe now you all have the Key from the AMD-APPLE rumors Apple asked AMD for technical help because of there Hypertransport technology and connect options for Multiprossecor boards. AMD does this self or need it for there Opteron Systems ( I saw boards and technical datasheets of some opteron boards) Its a very good solution so I think we´ll see a nearly identical solution from Apple !

Therefore I say this rumors from Macbidouille are true ! But the information Person they have dont know much about technics ! For me so far the used technics are very very logical !

But what I heard I would say that this version of the Baord is (because of its Fiber connector) a Xserve Board. Without this port it could be a Powermac board.

[Chobit]

You can always write a harddrive size in MB or GB, but you have to realize there is a conversion factor. Somtimes spaces are written without considering the conversion and are therefore bigger or smaller than you may think they are. Advertized Harddrive spaces ARE normally written in real Gigabytes (1,000,000,000 bytes per Gigabyte) however, when your computer reads it, I believe (I am not positive) It calculates the size in Gibibytes however writes GB. Before the prefixes were changed (1998) there could be either 1024 bytes in a kilobyte, or 1,000 and you rarely would know which one. Now you only know for sure what someone's talking about if they use the new base two prefixes as so many use the base 10 prefixes either way.

Anyway, this is all really off topic and we should probably get back to the actual thread.

Edit: Replaced incorrect Megabyte with Kilobyte

[robotrenegade]

That much ram would be sick!!!!!!!:D

[beatle888]

Originally posted by Mr. Anderson
32 Gig? That's huge, and most likely for a server. There are very few applications that would use this much memory.


but for a market that demands it. it would be perfect. and apple has been trying to be "the one" in the industries that WOULD benefit from this.

[pilotgi]

We DO need a power mac of this caliber however I doubt that it existists
I'm 99% sure we will see at least one machine with PPC 970 announced at WWDC. Probably the xServe and the PowerMac. And other announcements about PPC processors in other machines. WWDC is only three weeks away and there still has been no 15" PB update.

I feels it in me bones. I think Apple is allowing some leaks to MacBidouille so we'll know something BIG is coming.

Why they picked MB, I have no idea. MacRumors/Arn would have been fine.

[Furious Tiger]

would it hurt apple so much if they tricked out the mobos like intel an other mobo manufacturers do to get the edge. 800MHz FSB has just been released

[macmunch]

No, the P4 has no true 800Mhz Bus it handles 4 packages per cicle as far as I know the PPC970 does true 800 or 900 Mhz

[Furious Tiger]

Originally posted by macmunch
No, the P4 has no true 800Mhz Bus it handles 4 packages per cicle as far as I know the PPC970 does true 800 or 900 Mhz

that is why i said "TRICKED OUT". I know that they do not. they do a trick called QUAD PUMPING while apple does not

[dethl]

At first, I thought MB was doing a good job, but now I'm seeing more and more outrageous rumors from them.

I'm taking this whole 32GiB (weird to say..gibibyte :P) with a major grain of salt. I'll be happy to see a PPC970 that goes to 4GB(or GiB if you wanna get technical) this summer.

[macmunch]

k sorry !

But the Apple solution is much better when you read my article 5 posts above

[AidenShaw]

Originally posted by macmunch
No, the P4 has no true 800Mhz Bus it handles 4 packages per cicle as far as I know the PPC970 does true 800 or 900 Mhz

What's the difference between doing 4 transfers per cycle on a 200Mhz bus, and 1 transfer per cycle on an 800 MHz bus. Effectively none, right?

And check your PPC970 facts - it has 2 unidirectional double-pumped (DDR) 32-bit busses, each at 1/4 of the CPU speed. You get 900MHz by double-pumping (DDR) the 450MHz bus on a 1.8GHz part. This is 3.2GB/sec read and 3.2GB/sec write.

The P4 is 64-bit bidirectional at 800MHz (200MHz quad-pumped), so the full bandwidth can be either read or write.

Guess who's "tricked out"???????????????????????????

[MetallicPenguin]

I wonder what will happen come WWDC.

[makkystyle]

Arn, are all these stories from MacB going on the front page becuase you have heard any kind of corroborating info or, just because they are fairly big/interesting news?

I would love it if all of these rumors are on the money, but if they were so close to the truth and coming out this early, wouldn't Apple legal be sending some fairly stern words to MacB? I mean they sent lawsuit threats to Spymac over iSync compatible phone lists being leaked. Seems to me that Apple would have tried to nip this in the bud several weeks ago if MacB were really putting out info that hit close to home.

Anybody else reading these "leaks" with a whole hand-full of salt?

[iJon]

Originally posted by makkystyle
Arn, are all these stories from MacB going on the front page becuase you have heard any kind of corroborating info or, just because they are fairly big/interesting news?

I would love it if all of these rumors are on the money, but if they were so close to the truth and coming out this early, wouldn't Apple legal be sending some fairly stern words to MacB? I mean they sent lawsuit threats to Spymac over iSync compatible phone lists being leaked. Seems to me that Apple would have tried to nip this in the bud several weeks ago if MacB were really putting out info that hit close to home.

Anybody else reading these "leaks" with a whole hand-full of salt?
well you cant be for certain, also the stie was thinksecret, not spymac. anyways, thinksecret had ipod pictures 2 weeks before the ipods were released. the picture were EXACTLY how the ipods turned out to be, apple never touched them.

iJon

[Longey Nowze]

Originally posted by makkystyle
wouldn't Apple legal be sending some fairly stern words to MacB?

maybe french law is different... maybe there this kinda thing is considered to be freedom of speech or something? i don't really know much about anything that has to do with law :o

edit: why are all the romours coming from MacB?! isn't there any other sites that is posting this much stuff?

thank you
MaT

[Freg3000]

These stories make me seriously doubt the possibility of a PPC 970 @ WWDC. MacB keeps coming out with this far out stuff, which, in my eyes, kills their credibility.

:(

[bousozoku]

Originally posted by Chobit
You can always write a harddrive size in MB or GB, but you have to realize there is a conversion factor. Somtimes spaces are written without considering the conversion and are therefore bigger or smaller than you may think they are. Advertized Harddrive spaces ARE normally written in real Gigabytes (1,000,000,000 bytes per Gigabyte) however, when your computer reads it, I believe (I am not positive) It calculates the size in Gibibytes however writes GB. Before the prefixes were changed (1998) there could be either 1024 bytes in a megabyte, or 1,000 and you rarely would know which one. Now you only know for sure what someone's talking about if they use the new base two prefixes as so many use the base 10 prefixes either way.

Anyway, this is all really off topic and we should probably get back to the actual thread.

There are 1024 x 1024 bytes in a Megabyte, 1024 in a Kilobyte.

32GB (or even GiB) is only 35 bits, wasting 29 bits of the addressing unless it's constrained to fewer than 64 bits. Perhaps, it is constrained to 35 bits of memory plus one bit to signal whether it's real or virtual memory--an IBM big machine habit.

[wizard]

At first, I thought MB was doing a good job, but now I'm seeing more and more outrageous rumors from them.

I'm taking this whole 32GiB (weird to say..gibibyte :P) with a major grain of salt. I'll be happy to see a PPC970 that goes to 4GB(or GiB if you wanna get technical) this summer.

__________________
Argh...I have to wait how long for my 30 gig iPod?! (June 10th to ship)

If the PC coming out around the time of the WWDC are 970 based why would they even bother to implement a smaller memory capacity. 4GB is the curent addressing range of the G4 (this is not implemented in any Apple design due to I/O decoding) so why would you go to a 64 bit chip and use an address range that is covered by older processors? Further the PC world has already moved to dual channel memory systems, it is the only way to get the required memory performance.

I do not believe that the rumors are outrageous at all considering trends in the industry. Just look at what AMD is doing or Intel for that manner. Wether those rumors are true or the board being discussed is for a PowerMac is another issue. The reality is that there is a big pent up demand for large memory systems, if apple can tap that demand it would be fantastic.

The biggest problem I have is the fear that Apple will do that tapping with a big screw. In other words make the systems so expensive that tears will well up in our eyes and bank accounts will be bleed dry. It is obviuos that if you want a large memory system today you can buy one, the issue is finding a system that runs software you want to use.

Next month will be very interesting indeed. My geuss is new server hardware, but who knows for sure.

Thanks
Dave

[Kurt]

Come WWDC Apple will introduce the next PowerMacs with the 970. It will be much faster than the current G4's and will be close to the top end P4s. BUT, everyone will be very disappointed because the new machines will not be as described by MacB. They may not have 8 RAM slots or fibre channel on the motherboard. They may not have built in 5.1 audio or 6 PCI slots.

I think that these rumors are getting more and more ridiculous.

Maybe MacRumors should rename itself as the English version of MacB:D They do seem to be quoting them very often.

[wizard]

There are 1024 x 1024 bytes in a Megabyte, 1024 in a Kilobyte.

32GB (or even GiB) is only 35 bits, wasting 29 bits of the addressing unless it's constrained to fewer than 64 bits. Perhaps, it is constrained to 35 bits of memory plus one bit to signal whether it's real or virtual memory--an IBM big machine habit.

__________________
LOOP: MOVE #$01, D0
CMP #0, D0
BNE LOOP
folding@home is good for you.

Or perhaps they sized the memory subsystme around other parameters. This could be the number of physical slots they where will to produce. The limits for current DDR module design, which by the way I have no idea what that limit is.

This is still an expansion of 16 times the current limit if I recall correctly. Its a rather large jump, but even then I'm sure many would find it limiting. Lets just hope that they designed things for a seemless upgrade in the future. Lets hope that apple has keep all those other bits reserved, though I imagine the top 2 to 4 bits are going to I/O usage.

Actually the sketch of a machine as described in the MacB reports could have been lifted form recent PC world motherboard specs. So I'm not sure at all why it would be unexpected for apple to be working on similar technology. The question comes down to when will we see it. I flip flop between next month and late fall.

Obviously Apple needs to deliver such a machine now. Having no PoaerMac sales to speak of has got to hurt.

Dave

[Rincewind42]

Originally posted by bousozoku
There are 1024 x 1024 bytes in a Megabyte, 1024 in a Kilobyte.

32GB (or even GiB) is only 35 bits, wasting 29 bits of the addressing unless it's constrained to fewer than 64 bits. Perhaps, it is constrained to 35 bits of memory plus one bit to signal whether it's real or virtual memory--an IBM big machine habit.

The PowerPC 970 can only address 42 bits or 4 TB of RAM. The memory controller wouldn't get sent signals indicating real or virtual memory either way, as it only knows about real memory. Most likely (if this rumor is true) the controller is not constrained specifically to 35-bits total, but the RAM technology that it can interface with implies this limitation.

[Sol]

With a theoretical maximum of 32 GB I wonder if we will see the return of RAM disks in the Mac operating system. For hi-res video capture the RAM disk could be used to capture tapes in 5-10 minute segments and then moved to a serial ATA drive for playback and editing purposes. Of course the price of RAM will have to be lowered to the point where it would not be cheaper to buy a SCSI RAID array instead.

[Rincewind42]

Originally posted by AidenShaw
What's the difference between doing 4 transfers per cycle on a 200Mhz bus, and 1 transfer per cycle on an 800 MHz bus. Effectively none, right?

And check your PPC970 facts - it has 2 unidirectional double-pumped (DDR) 32-bit busses, each at 1/4 of the CPU speed. You get 900MHz by double-pumping (DDR) the 450MHz bus on a 1.8GHz part. This is 3.2GB/sec read and 3.2GB/sec write.

The P4 is 64-bit bidirectional at 800MHz (200MHz quad-pumped), so the full bandwidth can be either read or write.


The practical difference between a quad pumped 200Mhz bus and a double pumped 450 Mhz bus is one of latency - request and response will return to the CPU faster on the 450Mhz bus than on the 200Mhz bus simply because the 450Mhz bus transfers data faster. So the 450Mhz double pumped bus will transfer a data chunk just over twice as fast as the 200Mhz quad pumped bus, although they may have the same theoretical peak bandwidth.

The Pentium will be at an advantage (though unlikely used one) of being able to completely focus it's bus on either reading or writing, but will have a lower actual throughput. The 970 will have the (likely rarely used) advantage of being able to read and write to memory at the same time.

[Rincewind42]

Originally posted by Chobit
I'm just glad that someone's starting to use the correct GiB instead of GB. Its 32 Gibibytes because it is 2^30 bytes, not 10^9. I don't know.. It really doesn't matter too much, but sometimes its annoying not knowing if things (especially harddrives) are being measured in Giga or Gibi bytes. People use Giga (which is supposed to be pronounced "jiga" even though I still say it with a hard g out of habit and people not understanding me) when they mean gibi and it all makes my head hurt!

I can't stand the new abbreviations personally - it just seems silly that we are creating new abbreviations because one segment of the computer industry has consistently failed to use them in the same way that the rest of the computer industry does. So instead of K=1024 for computers and K=1000 elsewhere, it becamse K=1024 for computers unless you are talking about storage devices, and some other value completely when talking about the high density floppy... blah =p

PS: Like your handle - I'm waiting for the 3rd DVD myself right now :D

[Rincewind42]

Originally posted by Sol
With a theoretical maximum of 32 GB I wonder if we will see the return of RAM disks in the Mac operating system. For hi-res video capture the RAM disk could be used to capture tapes in 5-10 minute segments and then moved to a serial ATA drive for playback and editing purposes. Of course the price of RAM will have to be lowered to the point where it would not be cheaper to buy a SCSI RAID array instead.

Well, the general reason why RAM disks went away with demand paged operating systems (Unix, Windows, MacOS X) is because the operating system should be caching as much as possible in RAM including your huge capture sessions. The actual disk writes happen in a DMA process that allows the capture device to continue to write to memory and that memory write to be sent to disk without a huge performance hit. I did once actually setup a RAM disk under Win 2k and not only was it a pain in the *** to figure out how to do, it didn't end up helping performance any.

[Phil Of Mac]

Originally posted by Chobit
Harddrives and RAM and other data storage devices should be measured in Kibi/Mebi/Gibi/Tebi etc. bytes as they are always in some grouping of a power of 2, not a power of 10.

Kilo=1000
Kilobyte=1024 bytes
1000 is approximately equal to 1024.
It is an approximation, not an error to refer to kilobytes as 1024 bytes, megabytes as 1024k, gigabytes as 1024 megs, etc.
It's a non-issue.

[Sol]

Originally posted by Rincewind42
The 970 will also have the (likely rarely used) advantage of being able to read and write to memory at the same time.

That sounds like a technology waiting for an application. Software written from the ground up to utilize this feature would be able to do things that ported software never could. Unfortunately most non-Apple software for the Mac market seem to be ports from Windows.

[york2600]

2 years ago I ordered a HP Netserver LH3000r. It has 2 1gig modules. 1 gigs have been out for a while now.


Originally posted by EponymousCow
.

I believe that 1GB modules have just become available. Apple has done this before, making their motherboards compatible with a memory size not yet available. In fact, the MDD powermacs have 4 slots that can use 1GB DIMMs, but the motherboard can only handle 2GB total. Must be limited to 31-bits of addressing. :rolleyes:

[Chobit]

Kilo=1000
Kilobyte=1024 bytes
1000 is approximately equal to 1024.
It is an approximation, not an error to refer to kilobytes as 1024 bytes, megabytes as 1024k, gigabytes as 1024 megs, etc.
It's a non-issue.

Actually, it was an SI standard to for Kilo to equal 1000 or 1024 depending on the context, not an apporoximation. Now the SI standard is Kibi as 1024 and Kilo as 1000, and when you get to higher prefixes, as we will soon, the differences become larger and larger. A Terabyte (1 000 000 000 000 bytes) shares only one significant digit with a Tebibyte (1 099 511 627 776 bytes).

I'm not saying you can't use Terabytes, as you most certainly can, I'm saying there's a distinction that should be made and things get confusing when people don't accept standards.

Anyway, I'm getting tired and don't care. 32 GB or 32GiB would be great to have in a computer. If only I had more money.

PS: Sorry if I misunderstood what you meant to say.

[arn]

Originally posted by makkystyle
Arn, are all these stories from MacB going on the front page becuase you have heard any kind of corroborating info or, just because they are fairly big/interesting news?

They are front page because they are of significant interest... and people are following it.

i dont have any collaborating info. I've heard varying things regarding the timeframe of these things... with slightly more relaible lean towards them not being available immediately.

arn

[fourthtunz]

Originally posted by Sol
That sounds like a technology waiting for an application. Software written from the ground up to utilize this feature would be able to do things that ported software never could. Unfortunately most non-Apple software for the Mac market seem to be ports from Windows.

I'm not sure about games but most of the best Audio,video and graphics software started life on the Mac.
Can I have a witness or an amen:D
daniel

[fourthtunz]

Originally posted by Rincewind42
Well, the general reason why RAM disks went away with demand paged operating systems (Unix, Windows, MacOS X) is because the operating system should be caching as much as possible in RAM including your huge capture sessions. The actual disk writes happen in a DMA process that allows the capture device to continue to write to memory and that memory write to be sent to disk without a huge performance hit. I did once actually setup a RAM disk under Win 2k and not only was it a pain in the *** to figure out how to do, it didn't end up helping performance any.

Ok, so for the slower ones in our viewing audience,(me) you're saying that the ram disk would no longer be needed because the os is doing the same thing behind the scenes?
Setting one up in OS 9 and previous was easy but 1 gig isn't enough at least for audio and video to matter.
So I guess my next questions would be how much ram can an app use under os X and can you set up a ram disk?
I would still love to have 12 gigs or so of ram disk to capture to, seems like it would have to speed up renders? peace
daniel

[wheezl]

Well the dual Opteron MB from Tyan already supports 12GB. One of the biggest wins for most users with a 64-bit chip is having lots of memory more easily addressable. So not only is 32GB pretty damn normal. (and not really a huge amount in the server world). it's damn handy for lots of video work. Try using Discreet Combustion and you would know why many of us would dig on 32GB of memory.

Still, the MB folks do seem pretty over the top with the rumors.

[AidenShaw]

Originally posted by Rincewind42
The practical difference between a quad pumped 200Mhz bus and a double pumped 450 Mhz bus is one of latency - request and response will return to the CPU faster on the 450Mhz bus than on the 200Mhz bus simply because the 450Mhz bus transfers data faster. So the 450Mhz double pumped bus will transfer a data chunk just over twice as fast as the 200Mhz quad pumped bus, although they may have the same theoretical peak bandwidth.

Huh?? Just over twice as fast?

First of all, the P4 bus is 64-bits wide, not 32-bits like the 970 bus. So, "one transfer" is 64-bits at 800 MHz, vs 32-bits at 900 MHz (assuming 1.8GHz CPU).

I don't think we know enough about the 970 bus and memory controller to compare latencies. Does the 970 have to use 2 transfers on the 32-bit bus to pass the 64-bit memory address? That will hurt latency.

Does the 970 (or the P4) immediately forward the first chunk of data up through the cache levels to the registers? Or does it wait to fill a cache line (32 to 64 bytes) before making the data available - big latency issue here.

[AidenShaw]

Originally posted by wizard
4GB is the curent addressing range of the G4 (this is not implemented in any Apple design due to I/O decoding)

Actually, the G4 (and the P4) both implement 36-bit physical addressing, supporting up to 64GiB.

You are correct, no Apple design goes above 2GiB.

Many P4 systems are available with up to 32GiB today (e.g. Dell 6650) or even 64GiB (e.g. IBM x440).

If the PPC970 board supports 32GiB, that's just an implementation limit for that board - most likely influenced by the maximum DIMMs available and the maximum number of DIMM slots that are permitted (apparently 8). Future systems could support more, but it's not unusual for a particular mobo/chipset combo to support less than the maximum supported by the chip. For example, Xeon workstations (e.g. IBM zPro) support up to 8 GiB, even though the Xeon chip could support 64GiB.

[mathiasr]

Originally posted by Sol
That sounds like a technology waiting for an application. Software written from the ground up to utilize this feature would be able to do things that ported software never could. Unfortunately most non-Apple software for the Mac market seem to be ports from Windows.
This is nothing particular, the way caches work already mimics this behavihour once they are full they have to write a cache-line back to memory to keep a new one.

In the code you'll find load and store instructions mixed up, the problem I would see is that there are often more loads than stores (how much more depends on the kind of computation of course) and when an instruction is read it generates a load (this should have minimal impact). Anyway it's not that easy to tell the effect outside the CPU since what basically enters and leaves the CPU are whole cache-lines (32 Bytes usually, could be 128 Bytes on the 970), a single cache-line could take many hits from the Load/Store Unit point of view but would only require a single memory transaction on the outside bus.

And even if the PowerPC 970 is able to read and write at the same time there is no memory technology that allows this kind of access, memory chips have a single interface, you cannot designate two memory cells at the same time at this level since there is only one place to put the address of the cell. This will lead to some sort of serialisation, the memory controller will handle this, may be with a few Apple tricks. How it will cope with 16 open datastreams (8 per PPC 970) would be even more insteresting to know.

Compared to the G4e the 970 has overkill capacities to handle more memory and way faster, the problem is what kind of memory could follow its pace ? AMD put the memory controller inside the Opteron to get the best out of DDR2700 and reduce latencies, IBM will probably take the same approach in the 980; what Apple managed to do with the 970 is an open question.


Actually there are a lot more software coming from the GNU/Linux and *BSD sphere than from the Windows area.

[Sol]

Originally posted by mathiasr
Actually there are a lot more software coming from the GNU/Linux and *BSD sphere than from the Windows area.

Granted that OS X has more in common with Unix, Linux, Irix, etc. than it does with Windows. The problem is that most of the companies that create software for both Mac & Windows seem to concentrate all their efforts on the Windows applications and treat the Mac versions as if they were less a priority. This would explain why software is usually released for Windows first, then Mac. As examples Corel comes to mind and so does Adobe (since the introduction of OS X). This is why Apple needs more developers like The Omni Group who make the effort to build software that takes advantage of OS X's features. Their port of Giant: Citizen Kabuto was the first ever Mac game to utilize dual CPUs. More of that please!

[jettredmont]

Originally posted by tjwett
And I find it very hard to believe that a 4gb module would cost $191, I've spent more than that on a 512 SDRAM chip. Please prove me wrong, that would be exciting stuff.

Recently? Seems like you got ripped off. Seems like 512MB RAM should be in the 100-150 range for high-quality parts ...

But, yes, 4GiB for just $50 more ... yeah, that's a bit off sounding ...

I do have a link for you, although it is specific to Sun servers (not likely the same as in our 970), and it seems the previous poster lost a digit on his estimates:

4GiB @ $1984 from Dataram, $3800 from Sun.

http://biz.yahoo.com/bw/030318/185022_1.html

Here's another for Compaq SDRAM at $2199.99 per stick:

http://www.chicagocomputersupply.com/com4gbsdramm.html

I doubt that 4GB of memory on a stick went down in price by a factor of 10 in two weeks.

[mathiasr]

Originally posted by jettredmont
I doubt that 4GB of memory on a stick went down in price by a factor of 10 in two weeks.
Especialy since the new motherboard is supposed to use DDR400, the Compaq memory you pointed to is older SDRAM, Sun memory could be 128 bits wide, you cannot compare those.

Corsair is reputed for its fast (CAS 2-2-2 @ 200 MHz) and reliable memory, these could be the chips that would fit in the new Macs:
http://www.corsairmemory.com/xms/xms_modules.html

Still no GiB on the horizon there.

[jettredmont]

Originally posted by AidenShaw
What's the difference between doing 4 transfers per cycle on a 200Mhz bus, and 1 transfer per cycle on an 800 MHz bus. Effectively none, right?


Effectively none on long bursts, but the latency on a quad-pumped 800MHz bus is 1/200M instead of 1/800M on a single-pumped 800MHz bus.

And, you are correct: the 970 is a 900MHz, dual-pumped bus. It transfers data at a rate equivalent to an 800MHz dual-pumped bus after overhead is taken into consideration.

[pretentious]

Well I'm just upset now... I was expecting to get atleast 40GiB of memory. This is just wrong, only 32...shaw.... come on,what are we going to do with that???

I want more!! darnit! and I'm going to leave Apple if that is what they are only going to give me.

lol

[solvs]

Originally posted by AidenShaw
Huh?? Just over twice as fast?

First of all, the P4 bus is 64-bits wide, not 32-bits like the 970 bus. So, "one transfer" is 64-bits at 800 MHz, vs 32-bits at 900 MHz (assuming 1.8GHz CPU).

I don't think we know enough about the 970 bus and memory controller to compare latencies. Does the 970 have to use 2 transfers on the 32-bit bus to pass the 64-bit memory address? That will hurt latency.


I always thought it was the other way around. P4 = 32 bit, 970 = 64 bit. But that's something else. I wonder how many stages the 970's pipeline has? I've heard all sorts of different answers to that one.

Suppose I could look it up, but I'm tired and don't really care right now.

-

BTW - GIGA/GIGI... doesn't really matter. Can't they just call a 185GB hard drive a 185GB hard drive like with FP monitor sizes now?

(sorry, I'm cranky today) :mad:

[SuzanneA]

Originally posted by fourthtunz
I'm not sure about games but most of the best Audio,video and graphics software started life on the Mac.
Can I have a witness or an amen:D
daniel

Well, at the risk of being pedantic, most of the best *Audio* software started life on the Atari ST, but MATURED on the Mac.

(Cubase, Logic (Notator), etc)

Still, the PC is still the 'last platform it was ported to' in these cases, and in most of them, its obvious when you try to use them (Logic and Cubase always feel like mac apps, even on PC)

[maradong]

Originally posted by SuzanneA
Well, at the risk of being pedantic, most of the best *Audio* software started life on the Atari ST, but MATURED on the Mac.

(Cubase, Logic (Notator), etc)

Still, the PC is still the 'last platform it was ported to' in these cases, and in most of them, its obvious when you try to use them (Logic and Cubase always feel like mac apps, even on PC)
true :D
hey a female humanoide on the boards :D nice...

[MetallicPenguin]

Here's some more of from MacB (not on the same thing, but interesting):

According to an internal source (Thank you Superced), Motorola started to manufacture chips in 0,09 Microns in the factory of Crolles 1. We do not know any more, but if Motorola manages to engrave of G4 to this smoothness, they could return in the race at the frequency. It will be difficult to catch up with the PPC 970 in installed capacity, but G4 with 0,09 given rhythm with 2GHz, would be an excellent candidate with the replacement of G3 in the entry of APPLE range, in particular in the portable. Remain to know if APPLE would manage to overcome its antagonisms to agree to have a common future. If one sticks to Roadmap Motorola (seldom followed by them same), G4 with 0,09 Microns would be 7457-rm (envisaged at the beginning into 0,10) which benefit from Rapid IO to maximize the band-width memory, large weak point of the current G4.

[Hattig]

Originally posted by AidenShaw
Huh?? Just over twice as fast?

First of all, the P4 bus is 64-bits wide, not 32-bits like the 970 bus. So, "one transfer" is 64-bits at 800 MHz, vs 32-bits at 900 MHz (assuming 1.8GHz CPU).

I don't think we know enough about the 970 bus and memory controller to compare latencies. Does the 970 have to use 2 transfers on the 32-bit bus to pass the 64-bit memory address? That will hurt latency.

Does the 970 (or the P4) immediately forward the first chunk of data up through the cache levels to the registers? Or does it wait to fill a cache line (32 to 64 bytes) before making the data available - big latency issue here.

Most likely the 970 interface is packetised, like HyperTransport is, whereas the P4 bus is the old fashioned bus type that isn't packetised. Packetisation adds to the latency of course, but means that you can have thinner busses, and variable width busses (e.g., HyperTransport can vary from 2-bits in each direction to 32-bits in each direction), and also run the bus at faster speeds.

Previous to this rumour, I thought that a dual 970 PowerMac would have 1 northbridge that would connect to the processors (one port per processor), and a dual-channel memory controller.

Now I believe that each processor has its own Northbridge, with dual-channel memory controller. This makes sense, as the same northbridge element can be used for 1 processor designs (PowerBook, iMac2004, etc) up to many-processor designs (e.g., 8-processor Xserve or something). Hence 8 slots on the motherboard. Hence the IBM 980 having this (or similar) northbridge logic integrated into the processor.

I expect that this northbridges connect to each other, and other system components, using HyperTransport - even if the IBM 970 bus is not HyperTransport itself (a standard 6.4GB/s HyperTransport configuration is 16/16 at 800MHz DDR, not 32/32 at 450MHz DDR).

However I don't really trust the rumours since a couple of MacB rumours ago ... so using them as a basis for the above reasoning is not the safest thing to do.

[AidenShaw]

Originally posted by solvs
Can't they just call a 185GB hard drive a 185GB hard drive like with FP monitor sizes now?


They do! A 185GB drive has 185,000,000,000 bytes. (Actually, most of them have a bit more, the advertised size is usually rounded down to a nice number - e.g. the 120GB IBM/Hitachi drives are really 123.5GB)

Unfortunately, many operating systems say that the drive is 172.3 GiB, and people think that they lost 12GB to "formatting".

[maradong]

Originally posted by AidenShaw
Unfortunately, many operating systems say that the drive is 172.3 GiB, and people think that they lost 12GB to "formatting".
lol ?

[ZeeOwl]

Originally posted by joelc
I thought my 512 megs was pretty good...*looks sheepishly at Indigo iMac*
This is crazy stuff. How common is what ZeeOwl described? Do people actually need that kind of capacity?

In the wide world of Wintel, it's pretty rare. But in the Mac community, this is common stuff. And I assume that Apple is targeting Mac users with this new design. ;) Most PowerMacs out there are being used for high-end video/audio/scientific work. And in those types of applications (especially now with HDTV and 5.1 surround) you actually need about 20 GiB of work space. Having this much RAM would eliminate the need to make heavy use of virtual memory. Just that in itself would boost processing speed substantially. Couple that with a processor that is about twice as fast, and you have a killer high-end machine. We're talking PowerMac/XServe here, not iMac.

[ZeeOwl]

Originally posted by SuzanneA
Well, at the risk of being pedantic, most of the best *Audio* software started life on the Atari ST, but MATURED on the Mac.

Ya, Ataris rock! lol I have two: a 1040ST and a Falcon 030. Then I 'switched' to a PowerMac B&W. Does that count? Can I be in a Switch ad? lol

[ZeeOwl]

Originally posted by AidenShaw
They do! A 185GB drive has 185,000,000,000 bytes. (Actually, most of them have a bit more, the advertised size is usually rounded down to a nice number - e.g. the 120GB IBM/Hitachi drives are really 123.5GB)

Unfortunately, many operating systems say that the drive is 172.3 GiB, and people think that they lost 12GB to "formatting".

Yup. And actually operating systems (including Mac OS X, you-hoo Apple!) incorrectly display it as 172.3 GB, instead of 172.3 GiB. They're misleading the user by not using the correct SI prefix. The number is right, it's the letters that are wrong. Sort of like when Americans come up here... If our speed limit signs said 100 mph, that would be misleading, because the limit is actually 100 km/h.

[AidenShaw]

Originally posted by ZeeOwl
In the wide world of Wintel, it's pretty rare.


Maybe "rare" in the $399 PC world, but 512MiB to a couple Gig is pretty common in the rest of the world.

My company puts 512MiB in a "secretary's desktop". Even though that would seem to be overkill, there's a big advantage. Win2K/XP are very good about taking unused memory for the filesystem cache. This means that those email/web desktops will be caching a lot of files - which reduces the load on the networks and file servers.

Workstations for engineers start at 1GiB, with a second GiB available just by ticking the "2 GiB" box on the internal requisition.

Standard laptops are 1GiB Dell Latitude D400s. My group just got a half dozen D600s with 2GiB - 1GiB isn't enough.


And in those types of applications (especially now with HDTV and 5.1 surround) you actually need about 20 GiB of work space. Having this much RAM would eliminate the need to make heavy use of virtual memory.

Ummm, you're not using virtual memory for 20 GiB of work space today - the 32-bit machine only has 4GiB virtual memory!

Maybe the programs are shuffling the data around in work files, or only keeping part of the source file in virtual memory - but today they're not keeping it all in virtual memory.

(Apologies if you knew this and your post wasn't clear enough - but FTR wanted to make sure that nobody read your post to think that 20 GiB VM apps were running today on the Mac.)

[JJTiger1]

22 DAYS TO GO.

Then we shall know who is right and who will have egg on their face,

... and who pays the big bucks to be the first on the block with the new toy.

:D

JJ

[ZeeOwl]

Originally posted by AidenShaw
Maybe "rare" in the $399 PC world, but 512MiB to a couple Gig is pretty common in the rest of the world.

My point exactly. Most of the PCs being sold now are in the 399$ to 799$ range. That's why most people buy PCs, because they're cheap. And the vast majority of the PC owners I know are using them for eMail, Web surfing and games, with some occasional word-processing. Who needs more than 2 GiB of RAM to do that? I only know one PC owner who does high-end (audio) stuff with her's. In the Mac world, the majority of PowerMac owners are using their machines for high-end work.

Ummm, you're not using virtual memory for 20 GiB of work space today - the 32-bit machine only has 4GiB virtual memory!

Maybe the programs are shuffling the data around in work files, or only keeping part of the source file in virtual memory - but today they're not keeping it all in virtual memory.

(Apologies if you knew this and your post wasn't clear enough - but FTR wanted to make sure that nobody read your post to think that 20 GiB VM apps were running today on the Mac.)

Yup you're right. It was kinda misleading. "Shuffling" and virtual memory are not the same thing, technically speaking. Right now, the work I do requires about 3 GiB of memory. And my machines only have 1 GiB of RAM. So I'm not sure if OS X is handling this as virtual memory or "shuffling", as I don't know what the logical memory limit is under OS X. All I can tell you is it works, but it's slow. :D So I'd enjoy having a machine with 4+ GiB of physical RAM.

[fourthtunz]

Originally posted by SuzanneA
Well, at the risk of being pedantic, most of the best *Audio* software started life on the Atari ST, but MATURED on the Mac.

(Cubase, Logic (Notator), etc)

Still, the PC is still the 'last platform it was ported to' in these cases, and in most of them, its obvious when you try to use them (Logic and Cubase always feel like mac apps, even on PC)

Yeah I knew I could be opening up a can of worms! I almost bought an Atari but bought an Amiga 2000 with a video toaster in 92', that was an awesome machine! I was speaking of Protools, Avid and the Adobe stuff, I think was on the Mac first.
Peace

[AidenShaw]

Originally posted by ZeeOwl
My point exactly. Most of the PCs being sold now are in the 399$ to 799$ range. ... In the Mac world, the majority of PowerMac owners are using their machines for high-end work.


A couple of points.

First of all, you're comparing bargain basement PCs with Powermacs - when perhaps a better comparison would be Powermacs with Intel workstations (single P4 or duals). Compared price-for-price, there's probably much less difference in memory and whether the user is "high-end" or not. Or, compare the $700 PC with the eMac or iMac...

Second, it isn't clear how much memory those bargain PCs will really have - 128MiB is the min, but memory upgrades are probably the most common option that people would take. Your comment that 512MiB is rare in the PC world is not easy to prove or disprove.

Don't take offense, but your "high-end" comments do come across as a bit elitist. There are lots of high-end PC users as well, but since you aren't comparing Intel workstations to PowerMacs you miss them.

This is like the "Mac users are better educated" stories - when of course that's obvious. Since Macs are on average more expensive, and since income is correlated with education level, simple statistics would predict that more expensive computers would tend to be bought by better educated people.

(And before anyone shouts that "Macs are not more expensive", please show me the $399 Mac to match the PCs that we're talking about here!)

[MetallicPenguin]

Originally posted by JJTiger1
22 DAYS TO GO.

Then we shall know who is right and who will have egg on their face,

... and who pays the big bucks to be the first on the block with the new toy.

:D

JJ

Yeah.........


I am willing to sell my iMac, flat screen (it's the new one with the dome thing).

Along with pro speakers, and an epson printer.

And a Sony Clié (NX70V, it the one with the swivel screen and built in camera and video recorder), and with that you get 4 games (on an average of $15 dollars worth each). I will also include 4 memory sticks:

One 128MB
2 32MB
1 16MB

I will request about $2100 for all of that, who wants it?

[visor]

Originally posted by joelc
I thought my 512 megs was pretty good...*looks sheepishly at Indigo iMac*
This is crazy stuff. How common is what ZeeOwl described? Do people actually need that kind of capacity?

Well, depends on what you do really. For home use, you'll probably not need it to quickly - because basically you dont know how to handle the software that does need the memory. However, I was recently attached to a little project that stuggled with a 200GB textureized Data.
Nice if you can render without swapping to HD.

[Rincewind42]

Originally posted by fourthtunz
Ok, so for the slower ones in our viewing audience,(me) you're saying that the ram disk would no longer be needed because the os is doing the same thing behind the scenes?
Setting one up in OS 9 and previous was easy but 1 gig isn't enough at least for audio and video to matter.
So I guess my next questions would be how much ram can an app use under os X and can you set up a ram disk?
I would still love to have 12 gigs or so of ram disk to capture to, seems like it would have to speed up renders? peace
daniel

Under MacOS X all disk transfers (read & write) are cached in memory. If nothing else is using the memory, then in theory your program could take up all available memory. In practice, the OS tends to leave a small reserve around and comits disk writes very quickly to avoid data loss, but by caching in memory it allows programs to retain as much as possible in memory (up to 4GB per process) where it is faster to access. I'm pretty certain you can see this with an Audio CD in iTunes. When you play a track the first time, you will hear it reading from the CD - on subsequent times the track will play from memory even allowing the CD to spin down completely (I know I've done this playing audio cd programatically, but I don't know if iTunes turns off the disk caching for it's reads).

[ZeeOwl]

Originally posted by AidenShaw
A couple of points.

First of all, you're comparing bargain basement PCs with Powermacs - when perhaps a better comparison would be Powermacs with Intel workstations (single P4 or duals). Compared price-for-price, there's probably much less difference in memory and whether the user is "high-end" or not. Or, compare the $700 PC with the eMac or iMac...

That wasn't my point. Of course a low-end PC isn't comparable to a high-end Mac. In performance or price. My point was that the ratio of high-end to low-end users in the Mac community is much higher than in the PC community (of course, that started shifting when the iMac came out). I'm sorry if this comes off as "elitist", but it's an undeniable fact. What are the traditional Mac strongholds? Publishing, film, music, academia, scientific research. What are the traditional PC strongholds? Home computing, games, mainstream business, internet. That's just the way things are... :) That's why I was pointing out that for a PowerMac, 32 GiB maximum RAM capacity is far from overkill. It's a necessity, as high-end apps will need this within the next 5 or so years. And high-end apps is what most PowerMac owners are using on their machines.

[Rincewind42]

Originally posted by AidenShaw
Huh?? Just over twice as fast?

First of all, the P4 bus is 64-bits wide, not 32-bits like the 970 bus. So, "one transfer" is 64-bits at 800 MHz, vs 32-bits at 900 MHz (assuming 1.8GHz CPU).

I don't think we know enough about the 970 bus and memory controller to compare latencies. Does the 970 have to use 2 transfers on the 32-bit bus to pass the 64-bit memory address? That will hurt latency.

Does the 970 (or the P4) immediately forward the first chunk of data up through the cache levels to the registers? Or does it wait to fill a cache line (32 to 64 bytes) before making the data available - big latency issue here.

It was stated that the 970 bus is a transaction based bus, which I read as a packetized bus. I would expect that a memory request would be more that just sending the memory location in either system, but I don't know anything about bus protocols themselves. In order to avoid retransmitting a memory address I would expect the 970's bus protocol to transmit at least 32 bytes for any memory read - a 4-byte 'read' command, the 42-bit zero padded memory address, and a 4-byte 'tag' value. The tag value would be used for refering to the transaction from there out. Writes would be similar, but instead of the tag value the actual data would burst. Of course that little bit is just speculation :D.

And 2 back-to-back transfers actually won't hurt latency - bursts do arrive at the advertized (800 or 900Mhz) rate. It's only the first event - the lead transfer - that suffers latency. But the fact that every transaction must have a lead transfer means that the latency inherent in the bus will reduce it's efficiency when double or more pumped. In the past latency was kind of ignored in busses because they were all single pumped - Joe's 200Mhz bus had the same latency as John's 200Mhz bus. But when Jack came out with a 200Mhz bus that worked by double pumping a 100Mhz bus latency suddenly became an issue of comparison.

I don't know exactly how the cache affects this issue. I think that previous PowerPCs would have to at least obtain the critical word before it could be forwarded. SDRAM systems I've read always forward the critical word first, so that would mean that there is less wait. RDRAM systems do not always send the critical word first, and thus an increase in latency. So I would expect this issue to only hurt P4 systems running with RDRAM.

Damn I wish there were more info out on the 970 - it's hard waiting for that users manual :D

[Rincewind42]

Originally posted by ZeeOwl
Yup. And actually operating systems (including Mac OS X, you-hoo Apple!) incorrectly display it as 172.3 GB, instead of 172.3 GiB. They're misleading the user by not using the correct SI prefix. The number is right, it's the letters that are wrong. Sort of like when Americans come up here... If our speed limit signs said 100 mph, that would be misleading, because the limit is actually 100 km/h.

I would think this is mearly inertia. I have yet to see any computer industry literature use these SI prefixes (and given they have been out since 1999 and it seems people are just starting to use them 4 years later it doesn't seem like this will be coming soon).

Personally, I think that we could have avoided the whole issue if the storage industry had just used x-byte prefixes like everyone else in the computer industry does instead of first coming out with the whole 'formatted capacity' bull and then finally admitting that they are just measuring 1GB=1 billion bytes. That is what I think confuses people more than GB or GiB.

[fourthtunz]

Originally posted by AidenShaw
A couple of points.
since income is correlated with education level

Well these damn numbers, the richest man in the world is a college dropout and not a Mac user!(or is he??)
I do agree with what you say though, we Mac users are smarter:D

[Jeff Harrell]

Originally posted by ZeeOwl
Yup. And actually operating systems (including Mac OS X, you-hoo Apple!) incorrectly display it as 172.3 GB, instead of 172.3 GiB. They're misleading the user by not using the correct SI prefix.Sorry, but I have to take issue with this one. As has already been mentioned elsewhere on this board, the prefixes refer either to powers of two or powers of ten depending on the context. (I'm not talking about what the SI says they mean. I'm talking about how they're actually used in practice.) The fact that some random, unelected somebody came along, for reasons unknown, and decided to arbitrarily decide that giga, which previously meant either 10^9 or 2^30 depending on context, should mean only 10^9, and that using giga to mean 2^30 should suddenly be wrong, is irrelevant. What matters is how the terms are actually used in the real world, for that's from whence they derive their meanings.

Words don't mean what some guy in an ivory tower says they mean. They mean what the consensus agrees that they mean. So in this case, in practical terms, the world is right and the SI is wrong.

If you want to refer to 2^30 bytes by a word that is meaningless to the vast majority of humanity and that, when pronounced correctly, sounds like baby-talk, then be my guest. When you do, though, be prepared to launch into your explanation of how everybody else uses the term incorrectly because some French dude says so. And be prepared to be scoffed at in return.

Never forget: the purpose of language is to communicate. When you deliberately choose to use a vocabulary that is different from, and semantically incompatible with, the vocabulary of the person or people you're trying to communicate with, you're not communicating. You're actively hindering communication.

[Jeff Harrell]

Originally posted by fourthtunz
Well these damn numbers, the richest man in the world is a college dropout and not a Mac user!(or is he??)
I do agree with what you say though, we Mac users are smarter:D I thought the richest man in the world was the Sultan of Brunei? Last I heard, his personal net worth was measured in the hundreds of billions of dollars, but I can't recall the precise figure.

Anybody know what kind of laptop the Sultan of Brunei uses? ;)

[ZeeOwl]

Originally posted by Rincewind42
I would think this is mearly inertia. I have yet to see any computer industry literature use these SI prefixes (and given they have been out since 1999 and it seems people are just starting to use them 4 years later it doesn't seem like this will be coming soon).

Personally, I think that we could have avoided the whole issue if the storage industry had just used x-byte prefixes like everyone else in the computer industry does instead of first coming out with the whole 'formatted capacity' bull and then finally admitting that they are just measuring 1GB=1 billion bytes. That is what I think confuses people more than GB or GiB.

You're right. It is inertia (you're talking about the GB thing, not the mph thing, right? lol). Give it time. The SI system as we know it today has been out for over 80 years, and some countries are still using imperial. There's Liberia, Burma, and another one, what's it called again?... hehe

You're right things would have been simpler if HD manufacturer's had stuck to the old implied binary convention. But on the other hand, I'm glad they messed up. Because it forced the SI commitee to clear up an ambiguity in the system. And ambiguity is not a good thing to have in a measurement standard.

[AidenShaw]

Originally posted by ZeeOwl
You're right things would have been simpler if HD manufacturer's had stuck to the old implied binary convention. But on the other hand, I'm glad they messed up. Because it forced the SI commitee to clear up an ambiguity in the system.

On the other hand, mega and giga have always been base 10 in networking. Your gigabit ethernet is 1,000,000,000 bits per second, the 100baseT is 100,000,000 bits per second. Your modem, your fax, your ASR-33 teletype....

It wasn't "clearing up an ambiguity" - the SI prefixes were always base 10. The standards groups are trying to disambiquify the situation by creating a "base 2" prefix so that the "base 10" prefix can once again be clear.

[maradong]

Originally posted by ZeeOwl
You're right. It is inertia (you're talking about the GB thing, not the mph thing, right? lol). Give it time. The SI system as we know it today has been out for over 80 years, and some countries are still using imperial. There's Liberia, Burma, and another one, what's it called again?... hehe

lol ?
i think i guessed it...
well there are quite many :D like australia, us ( officially adopted it i think but pratically nobody uses it ) than there is south africa ( which i m not sure about ) ..
list to be completed.

[maradong]

Originally posted by AidenShaw
It wasn't "clearing up an ambiguity" - the SI prefixes were always base 10. The standards groups are trying to disambiquify the situation by creating a "base 2" prefix so that the "base 10" prefix can once again be clear.
not really true....
infact, for instance the acid grade can be calculated with the natural logarithms, and the non - natural ones. the non - natural ones, are not based on the "10 system "
but i might be wrong.

[rjwill246]

Originally posted by maradong
lol ?
i think i guessed it...
well there are quite many :D like australia, us ( officially adopted it i think but pratically nobody uses it ) than there is south africa ( which i m not sure about ) ..

and I am wayyyy off topic.
Australia has been metric for more than 20 years: currency first, then measurements a few years later. Even Canada and the UK still mix pounds/ounces and miles with Kg and Km. The US is not likely to change, as it is not clearly in its interest to do so, Mars Lander notwithstanding. I have lived with both systems... I hate Celcius and hectares but like kilograms. Milligrams make sense: I loathe Mmols and other ISO difficulties that makes blood sugars and cholesterol values much less clear than the larger older metric units. Now, isn't it interesting that the US (yeah! I said it) is so innovative and productive compared to so many of those countries that LOVE telling the US how backward it by not using the only logical system in the Universe... ISO 2000 +. Now, isn't it? Et, messieurs, c'est absolument impossible, n'est-ce pas? Imaginez cela!

[AidenShaw]

Originally posted by maradong
for instance the acid grade can be calculated with the natural logarithms, and the non - natural ones


????

The discussion has been focussed on the SI prefixes "kilo", "mega", "giga", etc.

What are you talking about - pH?

[AidenShaw]

Originally posted by fourthtunz
the richest man in the world is a college dropout and not a Mac user
This is specifically why I used the word "education" and not the word "intelligence". :p

The non-Mac using rich guy seems pretty smart, even if he got bored with university!

[fourthtunz]

Originally posted by Jeff Harrell
I thought the richest man in the world was the Sultan of Brunei? Last I heard, his personal net worth was measured in the hundreds of billions of dollars, but I can't recall the precise figure.

Anybody know what kind of laptop the Sultan of Brunei uses? ;)

Forbes #1 last year was Gates, I'm not sure what laptop he uses:D
daniel

[Jeff Harrell]

Originally posted by fourthtunz
Forbes #1 last year was Gates

Forbes only ranks the richest people in America.

[mathiasr]

Originally posted by rjwill246
Now, isn't it interesting that the US (yeah! I said it) is so innovative and productive compared to so many of those countries that LOVE telling the US how backward it by not using the only logical system in the Universe... ISO 2000 +. Now, isn't it? Et, messieurs, c'est absolument impossible, n'est-ce pas? Imaginez cela!
I like the way IBM innovates: the PowerPC 970 has a 121 mm2 die, is based on a 0.13 µm (130nm) CMOS SOI technology and is packaged in a 25x25 mm CBGA module that counts 576 pins on 1 mm pitch ;)

[fourthtunz]

Originally posted by Jeff Harrell
Forbes only ranks the richest people in America.

Here's their world listings: http://www.forbes.com/people/2001/06/21/billionairesindex.html
:D

[hacurio1]

Originally posted by Furious Tiger
would it hurt apple so much if they tricked out the mobos like intel an other mobo manufacturers do to get the edge. 800MHz FSB has just been released
Yup, they were already released, but they will be limited to the DDR400 throughput. If MB's rumor is true, the new PowerMacs will be a lot faster that any PIV, even if they are at 3.6Ghz and have only DDR400. The reason: The individual memory bank per processor.

[AidenShaw]

Originally posted by hacurio1
Yup, they were already released, but they will be limited to the DDR400 throughput.

Maybe if you'd bother to check your info, you'd see that the Intel DDR 400 is dual channel - two DDR400 DIMMs clocked in parallel for DDR 800MHz performance....

http://developer.intel.com/design/motherbd/bz/index.htm

Don't look at the specs of the individual DIMMs, and think that you can infer the specs of the mobo from that!

In fact, some of the Intel server boards choose to use PC1600 in sets of 4 to get PC6400 performance. The slower speed PC1600 DIMMs run cooler and are therefore more reliable.

[AidenShaw]

Originally posted by hacurio1
The reason: The individual memory bank per processor.


Think again about this one too.

NUMA is good when your program needs data that's in the memory attached to the CPU that you're currently running on. If the data happens to be in the other CPU's memory bank, then you have added latency to get to the data, and it runs slower.

I'd be *very* surprised if the Apple mobo uses NUMA for a dual processor.... Typically, NUMA systems have groups of 4 to 8 CPUs per memory domain (traditional SMP memory semantics). You build 32 to 64 to nnn CPU systems by linking these 4 to 8 CPU modules together, and memory in the local group is fast, and memory attached to other CPU groups is slower.

Think about it - lots of extra overhead if your program runs on CPU 1, and all of its data is in the RAM attached to CPU 0.

[hacurio1]

Originally posted by AidenShaw
Maybe if you'd bother to check your info, you'd see that the Intel DDR 400 is dual channel - two DDR400 DIMMs clocked in parallel for DDR 800MHz performance....

http://developer.intel.com/design/motherbd/bz/index.htm

Don't look at the specs of the individual DIMMs, and think that you can infer the specs of the mobo from that!

In fact, some of the Intel server boards choose to use PC1600 in sets of 4 to get PC6400 performance. The slower speed PC1600 DIMMs run cooler and are therefore more reliable.

I'm aware of the 6.4GB/s theoretical throughput of the Dual Channel DDR400, But since the rumor mentions a dual processor dual memory bank motherboard with, what it has been mentioned to be, two north bridges, I assumed better performance. Nobody knows specifics regarding the motherboards, nor memory controllers, but as far as I've read the 970 is a excellent at SMP, hence my assumption. As IBM describes it, the 970's FSB is an "elastic interface" allowing easy multiprocessor synchronization, memory sharing, and memory coherency. Two 970s with two bridges and two memory banks allows a 6.4GB/s throughput per processor. So it's all up to how does Apple design the Motherboard. If they do a poor job, then yes, the performance (throughput related) between a PIV in an 875 chipset will be similar; but, if Apples does a great job designing the Motherboard and memory controllers, then a PIV in an 875 chipset will not stand the chance (Assuming dual 970s). But yes, it was my mistake to state a rumor as a fact because none of us know specifics about the Motherboards and memory controller’s design of the next PowerMacs.

[Phil Of Mac]

Originally posted by ZeeOwl
Yup. And actually operating systems (including Mac OS X, you-hoo Apple!) incorrectly display it as 172.3 GB, instead of 172.3 GiB. They're misleading the user by not using the correct SI prefix. The number is right, it's the letters that are wrong. Sort of like when Americans come up here... If our speed limit signs said 100 mph, that would be misleading, because the limit is actually 100 km/h.

Oh, quit your whining. 1GB=1024MB, 1MB=1024K, 1K=1024 bytes. That's what it means in the real world. The SI-monkeys can call it whatever they want, but as rational individuals, we have the right not to go along with them when they're being stupid.

They're being stupid.

And changing to the metric system isn't worth the effort needed to change.

[mathiasr]

Originally posted by Phil Of Mac
1K=1024 bytes. That's what it means in the real world.
I don't know where you are from, but in my universe 1 K = -272.16 °C ;)

[AidenShaw]

Originally posted by Phil Of Mac
Oh, quit your whining. 1GB=1024MB, 1MB=1024K, 1K=1024 bytes. That's what it means in the real world.

In the "real world" K=1000, M=1,000,000, G=1,000,000,000 long before some computer types started to misuse the terms.

Not even "all" computer types, the storage and networking people still use the original base 10 meaning.

The "stupidity" is to continue the current ambiguous use of the long-standing SI prefixes. Learn to type the little "i" when using a binary prefix - you'll fit into the mainstream better!

[solvs]

GiB/GB

1000/1024

17'/15'

Come on kids, does it really matter that much?

I shoulda gone to Karaoke (ha - look at me with my bad grammar, and spelling and misusing words. Bet that's driving y'all nutz n'stuff).

[groovebuster]

Originally posted by rjwill246
Now, isn't it interesting that the US (yeah! I said it) is so innovative and productive compared to so many of those countries that LOVE telling the US how backward it by not using the only logical system in the Universe... ISO 2000 +.

:D

You should be careful to open that door... ;) Innovative... maybe, but not more than other industrialized countries. Maybe also because in science they use the metric system anyway? ... world-wide! ;)

When it comes down to productivity especially the US is not #1... Get your facts straight.

In my opinion all this has more to do with the need of feeling special. In the given situation the US has the economical power to reject the metrical system. Like they rejected almost any standard in the past 50 years to "protect the local market"... It's like: "Hey, we do our own stuff, who cares about the rest of the world!"

But nothing lasts forever and you'll see how fast the US will join the party of the "metric countries"! ;)

And now back to the topic... :rolleyes: :cool:

groovebuster

[ZeeOwl]

Originally posted by maradong
lol ?
i think i guessed it...
well there are quite many :D like australia, us ( officially adopted it i think but pratically nobody uses it ) than there is south africa ( which i m not sure about ) ..
list to be completed.

Can't comment on South Africa, but Australia has been metric since the 60's (about the same time as Canada). I lived there from 1973 to 1976, and everything was 100% SI.

[ZeeOwl]

Originally posted by rjwill246
Even Canada and the UK still mix pounds/ounces and miles with Kg and Km.

Can't comment on the UK (never been there), but Canada has been completely SI since the early 70's. The only place where the old imperial system is still commonly refered to is when people state their height and weight. Officially, we use SI for that too, but for some odd reason, people just can't seem to get used to the new system for their personal specs. :) For everything else, we're perfectly comfortable using metres and kilograms. Weird huh? As for miles, I don't think anyone up here under the age of 60 even knows how long a mile is! lol

[jettredmont]

Originally posted by AidenShaw
Maybe "rare" in the $399 PC world, but 512MiB to a couple Gig is pretty common in the rest of the world.


Pedantic note: If you are going to be pedantic about using "GiB" instead of "GB", the correct chort name for "Gibibyte" would be "Gib", not "Gig".

:)

[jettredmont]

Originally posted by AidenShaw
Maybe "rare" in the $399 PC world, but 512MiB to a couple Gig is pretty common in the rest of the world.


Pedantic note: If you are going to be pedantic about using "GiB" instead of "GB", the correct chort name for "Gibibyte" would be "Gib", not "Gig".

:)

[AidenShaw]

Originally posted by jettredmont
Pedantic note: If you are going to be pedantic about using "GiB" instead of "GB", the correct chort name for "Gibibyte" would be "Gib", not "Gig".


Ooops - got me. I'll practice.

Gib. Gib. Gib. Gib. Gib. Gib.

Oooh, that's hard! ;)

[maradong]

Originally posted by AidenShaw
Ooops - got me. I'll practice.

Gib. Gib. Gib. Gib. Gib. Gib.

Oooh, that's hard! ;)
make it GiB not Bib.
B has always represented the Byte ; b the bit...

[AidenShaw]

Originally posted by maradong
make it GiB not Bib.
B has always represented the Byte ; b the bit...
But "Gib" the diminuitive for "GibiByte", not the abbreviation "GiB".

[Rincewind42]

Originally posted by AidenShaw
But "Gib" the diminuitive for "GibiByte", not the abbreviation "GiB".

Heh... looks like things are getting confused again :D! Better come up with more abbreviations!

[Pale Rider]

Originally posted by Chobit
<SNIP> Before the prefixes were changed (1998) there could be either 1024 bytes in a kilobyte, or 1,000 and you rarely would know which one. Now you only know for sure what someone's talking about if they use the new base two prefixes as so many use the base 10 prefixes either way. <SNIP>


Maybe this is another SI thing, but abbreviations, titles, or informational indicators which come after the item they are describing fit the category of suffix. A prefix comes before the item it is describing, defining or modifying.

There is something to be said for precision in language as well as in measurement.:D

[Chryx]

Originally posted by Rincewind42
The Pentium will be at an advantage (though unlikely used one) of being able to completely focus it's bus on either reading or writing, but will have a lower actual throughput. The 970 will have the (likely rarely used) advantage of being able to read and write to memory at the same time.

Actually, anything that involves chucking large amounts of data in and out of main memory (video work springs to mind) the dual memory channels might well end up saturated, depending on just how capable the Altivec hardware on the 970 turns out to be.

[ZeeOwl]

Hey, that stuff you guys are on looks good... Can I have some too? :p

You're totally bored out of your skulls, aren't you? lol

Geez, if I'd known I would have started an international incident by using the correct SI prefixes, I would have shut my trap!

Ah, the life of the pioneer is a tough one... :rolleyes:

[Phil Of Mac]

Originally posted by groovebuster
You should be careful to open that door... ;) Innovative... maybe, but not more than other industrialized countries. Maybe also because in science they use the metric system anyway? ... world-wide! ;)

When it comes down to productivity especially the US is not #1... Get your facts straight.


In the stats I've seen, US workers are the most productive per worker. Per hour, the French are actually the most productive, but considering that they spend a third of the year on strike, a third of the year on vacation, and the other third working 20 hours per week, they don't really produce much in a year.

Originally posted by groovebuster
In my opinion all this has more to do with the need of feeling special. In the given situation the US has the economical power to reject the metrical system. Like they rejected almost any standard in the past 50 years to "protect the local market"... It's like: "Hey, we do our own stuff, who cares about the rest of the world!"

But nothing lasts forever and you'll see how fast the US will join the party of the "metric countries"! ;)


Actually, the reason we haven't switched to the metric system is that out government doesn't impose stupid regulations on industry like the metric system that, while they may be an improvement, aren't enough of an improvement to warrant the effort of changing the country over.

Originally posted by AidenShaw
In the "real world" K=1000, M=1,000,000, G=1,000,000,000 long before some computer types started to misuse the terms.

Not even "all" computer types, the storage and networking people still use the original base 10 meaning.

The "stupidity" is to continue the current ambiguous use of the long-standing SI prefixes. Learn to type the little "i" when using a binary prefix - you'll fit into the mainstream better!

I'm drowning in the mainstream already. The difference between 1000 and 1024 is trivial. If anything, storage and networking should change to the binary system because computers access data in binary chunks, not decimal.

[maradong]

Come on people, lets give it a little back to topic please ...

[Rincewind42]

Originally posted by Chryx
Actually, anything that involves chucking large amounts of data in and out of main memory (video work springs to mind) the dual memory channels might well end up saturated, depending on just how capable the Altivec hardware on the 970 turns out to be.

Actually, this has relatively little to do with Altivec in the end. You could saturate the busses relatively easily - but you wouldn't be doing any real work, just shuffling memory from one place to the other. But it also depends on what the memory system is capable of handling. If the memory system can actually read and write the full 3.2GB/s each way then this could do it. Otherwise you would get less.

The Pentium has a similar problem - if it wants to fully saturate the bus it would have to do some combination of complete read/write. It may be able to get some work done in between (if only due to having a higher cpu/bus clock ratio and double pumped ALUs) but the difference would be trivial. And I suspect that switching the bus from read to write mode has it's own penalties involved (since it must communicate this fact to the north bridge).

All in all, I think that both CPUs will show what a much faster bus can do, that neither of them will commonly show the maximum performance possible, and that the 970 will have a higher combined read/write transfer rate than the P4, but that neither pure reading nor pure writing will be as fast.

[AidenShaw]

Originally posted by Phil Of Mac
The difference between 1000 and 1024 is trivial.
Maybe, but the difference between 185 GB and 185 GiB is about 12 GB! :eek:

[maradong]

is there anybody who thinks there could become some 4 cpu configurations aviable in the near future? more and more x86 workstations are ready for 4 cpu s.. it s rare, but the share of those mobos is raising.

[jettredmont]

Originally posted by Phil Of Mac
If anything, storage and networking should change to the binary system because computers access data in binary chunks, not decimal.

Not sure about networking, but storage is measured more naturally using base-10 units ... well, not really so much "more" naturally, but there's no reason to use base-2 as there is with memory and bandwidth. So, they could use base-10 and way their drive is 2GB large, or use base-2 and say it is 1.5GB ... wonder why they chose base-10?

There really are very (very very) few places where "1K" should mean 1024 of something relative to the places where it should mean 1000 of something. As such, the binary-based worlds should be treated as the exception, not the model upon which others should rely.

[Rincewind42]

Originally posted by jettredmont
Not sure about networking, but storage is measured more naturally using base-10 units ... well, not really so much "more" naturally, but there's no reason to use base-2 as there is with memory and bandwidth. So, they could use base-10 and way their drive is 2GB large, or use base-2 and say it is 1.5GB ... wonder why they chose base-10?

There really are very (very very) few places where "1K" should mean 1024 of something relative to the places where it should mean 1000 of something. As such, the binary-based worlds should be treated as the exception, not the model upon which others should rely.

Hard disks were at one point sold in base-2 size. Unfortunately, this was when hard disk sizes were predominatly measured in the MB range. Somewhere around the early 90s (when HDs were starting to approach half a GB) someone switched over to base-10 to make their HDs look larger. Since consumers choose the 'larger' drives everyone else started doing it. By the time HDs reached half a gig everyone was doing it - this is why an old old limit on HD size was 540MB - it was about 536 thousand bytes of storage which equals 512MB (this issue was either in DOS or in the IDE controllers of the time).

So in the end, this evil was a marketing decision :D