MacRonin's joke aside I'd like to lend a prediction, or more of a what seems possible now. It could obviously be better, but my faith in Apple putting together a top performing machine has been shot to hell over he past few years, and I'll keep it reasonably tame because Apple only makes tame computers, and I don't think they will enter the highend market.
I'll only do the top model.
2x 970MP (dual core) 2.2GHz processors
1x PCI-E Slot
No QuadroFX Availability announcement from anyone.
Graphics cards: PCI-E versions of current models.
3x PCI-X Slots
2x HD bays
1x liquid cooling
To sum it up the same computer with 2x 2.2GHz Dual core processors,
and the added PCI-E graphics. That's what I predict.
Honestly I don't see the 970MP coming out next. I see the 970GX coming out before the MP. I think all these rumors and hints from CHUD are a matter of testing purposes... though I could be wrong. Just seems easier to make the GX over the MP. I hope onlooker is right though.
Honestly I don't see the 970MP coming out next. I see the 970GX coming out before the MP. I think all these rumors and hints from CHUD are a matter of testing purposes... though I could be wrong. Just seems easier to make the GX over the MP. I hope onlooker is right though.
What I think is the GX is being tested for Laptops, iMacs, and such. I think the MP has been in development for some time longer than we have been fully aware of it, and at WWDC it will appear.
I think either would suffice for 2D work. 2D isn't as taxing on a system as 3D where is you constantly prefer the fastest, and most powerful processors, and graphics cards available. I've never heard of a 2D render taking 2 to 3 days before. Where as in 3D depending on your equipment your scene could easily take much longer.
emig647, you're using the power question? How much power do you need? Since when have people ever cared about how much power they need? People buy cars with some insane power under the hood not because they need it but because they want it. When buying a Power Mac, we're not asking for a Ford Escort, you can get a Mac Mini for that. When buying a Power Mac (once the fastest desktop on the planet) we're expecting performance equal to the fastest car on the planet. Sure, we don't need all that power right this minute (aside from the folks using Maya and such) but we all know that we may want that power sometime down the road. We buy expensive cars for the same reason we buy expensive computers. Because we want them, and there is never too much power.
Many people can say that they don't want all the power, and they are the ones usually using really old computers, and not likely to be buying a Power Mac, but rather an eMac or iMac. But if you're going for what was once the fastest desktop on the planet, you're well beyond the realm of a Ford Escort, and we all know it. I want all the power, knowing I'll probably never max out a machine like that aside from when I'm compressing video in Final Cut Pro, or playing a video game. That's my reality, and I've got over $8k waiting for a new one and won't think twice about dropping the cash on a new high-end system. We buy what we want, not necessarily what we need.
as far as my question on which (dual 3 or quad 2.2) wouldnt apps that are non mp aware perform better on the dual 3??? Like Indesign and Illy are not multiprocessor aware I do not believe.
A bottleneck in my workflow right now is ripping open 1 gig+ eps files created/exported in Quark into Photoshop for example. You'd be surprised how slow that can be on my dual 1 gig! I wonder which machine would be better for this type of task??? And massive batch processes in Photoshop, like converting a few thousand psd files into flattened web ready jpgs.
Since i run so many apps all the time, i'm more inclined to go for the quad since I assume it would be the finest multitasking system ever made (factoring in OSX).
I stopped reading after page 7.but most of u guys have no real clue what PCi E is about.PCIE is a complete upgrade for the pci bus.PCi-X is a juice up longer slot pci bus but it has to have switches(extra chips to control each lane) for each slot/ingrated part's on the south brigh like the hdd controllers,ingrated network cards.PCI-X is mostly used in servers,macs,High end workstations becuase anything that need MORE than 133mb's max shared PCi bus.Example A scsi 320 raid array of 4 drives or your highend audio work addion card.
PCi 2.X all slots and anything linked though the pci bus to the south brige has a max ban of 133mb and it a parallel inerface so it needs many wire and traces to to the task.Just think the diffrents between pata ribbion cable and sata cable.
PCIE Has a speed of 250mb per wire/lane so PCIE x1 slot has 500mb's of i/o bandwidth.PCIe x4 has 2gigs of total i/o bandwidth and PCiE x8 is 4gigs.Now looking at that do u really think its just for video cards?Then u can use X many lanes or hypertranport to like the south bridge to the north bridge.
Now reading the last page lol ppl are asking about the GF6200 its is a bragin bin card it has a max of 64mb of ram on a 32bit or 64bit memory bus and it shares upto 256mb of system ram and if u want to know it speed same as any pc video cards goin into a mac look up at a pc review site and with tomshardware he is amd bias and a been has been a big lair a few time.Just look it up if u think i am a liar about tom
THT your explaing of ddr2 vs ddr1 is majory flawed
quote"DDR-II is effectively quad pumped SDRAM. So DDR-II, PC2-4000 memory is 125 MHz SDRAM that can transfer 32 bytes/cycle (64 bits * 4/cycle) for a data rate of 4 GByte/s. DDR-I is double pumped SDRAM. DDR-I, PC4000 memory is 250 MHz SDRAM that can transfer 16 bytes/cycle (64 bits * 2/cycle) for a data rate of 4 GByte/s."
Double data rate sdram @400mhz is 200mhz X 2bits X64bits=3.2 gigs/sec.DDR sends data at the rise and falling of a clock pulse.SDRam was 1 bit per clock pulse.DDR2 is the same tech as DDR1 other than It uses less power (1.8volts) than standerd DDR1 upto ddr400 (2.5 volts).Your XXXmhz X 4bits is Rdram
DDR2 increases the memory prefetch to four from 2 in ddr1.
But The reason why ddr2 sux is have a cas speed of 4 or 5.U might see so ultra highend ram market that alot of gammes and built your own rig in the pc market @cas 3 but i think its only that low of cas timeing at ddr2 533.Even still u get DDR400 cas 2 samsung awsome TCCD ram chips that can hit DDR600 with better timing than DDR2 can and still cost less and after all AMD goin to support it for awhile longer.
Man u easly slap back don't u.I bet u didn't even read my post did u? just the 1st line.Ya why do u need 6 pciE slots well if u look at PC motherboards there just not all PCIE slots they have your x16 slot or x8 ,a few X4 and a few x1 slots plus your reg pci slots.Apple can put the pciE slots and pci and pci x if they want to thats no prob.
The push of PCiE is it finally some that will replace PCi\\X that is better and goin to cost less in the end for hardware maker because it is serial based.A x1 slot is like 1/3 the size of a reg pci slot and it gives u 250mb in and out=500mb when pciX 133mhz X 32bits= 533mb and 2 wires carry that other than a crap load of them.
Man u easly slap back don't u.I bet u didn't even read my post did u? just the 1st line.Ya why do u need 6 pciE slots well if u look at PC motherboards there just not all PCIE slots they have your x16 slot or x8 ,a few X4 and a few x1 slots plus your reg pci slots.Apple can put the pciE slots and pci and pci x if they want to thats no prob.
The push of PCiE is it finally some that will replace PCi\\X that is better and goin to cost less in the end for hardware maker because it is serial based.A x1 slot is like 1/3 the size of a reg pci slot and it gives u 250mb in and out=500mb when pciX 133mhz X 32bits= 533mb and 2 wires carry that other than a crap load of them.
#1 what mother board has 6 PCI-E card slots? I've never seen one. I've never seen one with more than two.
#2 How many types of cards are using PCI-E right now?
#3 your article is from May 13, 2004 when many thought PCI-E was going to be used by every card made. It hasn't been. And they probably thought that back then because of that same article.
"Ya why do u need 6 pciE slots well if u look at PC motherboards there just not all PCIE slots they have your x16 slot or x8 ,a few X4 and a few x1 slots plus your reg pci slots "
look at where i put "not" in there.There is pci reg slots.
Plus when u look a good retail Pc board alot of stuff is onboard and they are mostly linked to a pciE link or reg pci.There becomeing less need on the pc side for alot of stuff in card format.u got on board lan or 2,have a good audio chip at times ,sata raid support,bunch of ubs 1.1/2 ports a few firewire ports.heck even hardware firewall in one chipset.
PCIE is goin to be used pretty much everything when the time comes. PCI reg is old it will be supported for quite awhile longer but in the end it needs to die like isa and nubus.The same with agp it been good to use for many years but with video cards getting faster,SLi,muilt gpu's the agp can't do sli and with more than 1 gpu on the card it turns the agp slot into a 66mhz pci slot.
Oh with u never seen a motherboard more than 2 pciE slots u never looked too hard.Most PciE motherboards that not the cheapy ones have 4 pcie slots and 2 pci.
Comments
With Dual dual-core PowerPC970MPs running @ 3GHz, 16GB of DDR2/667MHz RAM, dual optical drives, four SATAII HDDs & six (yes,six!) PCIe slots...
Lets make sure to space those slots to allow for all sizes of cards, okay?
To fill the slots, well...
The aforementioned dual QuadroFXs,
A pair of ART PURE rendering cards (imagine if these could also be used to accelerate CoreX functions!!!)
A Fibre Channel card & a HD A/V I/O card...
Run a XRAID off of the FC, and an Apple 30"/Cintiq 21" off of the QuadroFXs...
Oh yeah, and an Apple 23" off of the A/V card for a HD monitor...
Sweet! And expensive! Call it about 50K, with software (Apple Production, Adobe Creative, Maya Unlimited & Shake) & support (Maya & Shake)...
;^p
Originally posted by hmurchison
Didn't Nvidia just announce a new 6200 series?
.... yup
This card should be the G5 PM lowend card now.
And hopefully the standard G5 iMac graphics chip too. Although the last rumors don't leave much hope for that.
I'll only do the top model.
2x 970MP (dual core) 2.2GHz processors
1x PCI-E Slot
No QuadroFX Availability announcement from anyone.
Graphics cards: PCI-E versions of current models.
3x PCI-X Slots
2x HD bays
1x liquid cooling
To sum it up the same computer with 2x 2.2GHz Dual core processors,
and the added PCI-E graphics. That's what I predict.
Anyone else feel like trying?
Originally posted by onlooker
... Anyone else feel like trying?
That sounds about right. The graphics will likely start with Nvidia 6200, 6600 and 6800 BTOs.
The lower-end PMs will likely have single 970MP processors, in order to make the chip supply go farther.
Maybe the top-end iMac will have a 970MP, but it would be hard to keep cool. (23"?)
Originally posted by emig647
Honestly I don't see the 970MP coming out next. I see the 970GX coming out before the MP. I think all these rumors and hints from CHUD are a matter of testing purposes... though I could be wrong. Just seems easier to make the GX over the MP. I hope onlooker is right though.
What I think is the GX is being tested for Laptops, iMacs, and such. I think the MP has been in development for some time longer than we have been fully aware of it, and at WWDC it will appear.
Originally posted by mikenap
What would you guys choose, a Dual proc 3 or a quad 2.2 for heavy 2D work? (photoshop/illy/indesign)
Quad all the way.
I just add up the megahurts.
6GHz verus 8.8Ghz is no contest.
Originally posted by hmurchison
Quad all the way.
I just add up the megahurts.
6GHz verus 8.8Ghz is no contest.
Very true but how much power do you really need for desktop development. My dual 2.0 was plenty fast for that stuff.
Many people can say that they don't want all the power, and they are the ones usually using really old computers, and not likely to be buying a Power Mac, but rather an eMac or iMac. But if you're going for what was once the fastest desktop on the planet, you're well beyond the realm of a Ford Escort, and we all know it. I want all the power, knowing I'll probably never max out a machine like that aside from when I'm compressing video in Final Cut Pro, or playing a video game. That's my reality, and I've got over $8k waiting for a new one and won't think twice about dropping the cash on a new high-end system. We buy what we want, not necessarily what we need.
Originally posted by onlooker
MacRonin's joke aside...
Joke...?!? Who's joking...!?!
Think about it, the specs are good, the pricing is correct, the parts are all do-able,,,
This could be a serious DCC workstation configuration, yet still provide WAY more bang for the buck than anything out there...
I remember the 50K price being about the same for a Hi Impact & a full cut of PA, WAY back in 'The Day'...
Originally posted by onlooker
No QuadroFX Availability announcement from anyone.
I really think that nVidia/Apple has GOT to get off the stick sometime on this whole 'pro-level' workstation graphics card crap...
With Tiger & dual-core CPUs, this could be the time to trot out 'pro-level' OpenGL cards...
A bottleneck in my workflow right now is ripping open 1 gig+ eps files created/exported in Quark into Photoshop for example. You'd be surprised how slow that can be on my dual 1 gig! I wonder which machine would be better for this type of task??? And massive batch processes in Photoshop, like converting a few thousand psd files into flattened web ready jpgs.
Since i run so many apps all the time, i'm more inclined to go for the quad since I assume it would be the finest multitasking system ever made (factoring in OSX).
or better than the ati 9600XT purposed in BTO ?
PCi 2.X all slots and anything linked though the pci bus to the south brige has a max ban of 133mb and it a parallel inerface so it needs many wire and traces to to the task.Just think the diffrents between pata ribbion cable and sata cable.
PCIE Has a speed of 250mb per wire/lane so PCIE x1 slot has 500mb's of i/o bandwidth.PCIe x4 has 2gigs of total i/o bandwidth and PCiE x8 is 4gigs.Now looking at that do u really think its just for video cards?Then u can use X many lanes or hypertranport to like the south bridge to the north bridge.
http://www.nwc.com/showArticle.jhtml?articleID=20000286
Here a good link for info.
Now reading the last page lol ppl are asking about the GF6200 its is a bragin bin card it has a max of 64mb of ram on a 32bit or 64bit memory bus and it shares upto 256mb of system ram and if u want to know it speed same as any pc video cards goin into a mac look up at a pc review site and with tomshardware he is amd bias and a been has been a big lair a few time.Just look it up if u think i am a liar about tom
THT your explaing of ddr2 vs ddr1 is majory flawed
quote"DDR-II is effectively quad pumped SDRAM. So DDR-II, PC2-4000 memory is 125 MHz SDRAM that can transfer 32 bytes/cycle (64 bits * 4/cycle) for a data rate of 4 GByte/s. DDR-I is double pumped SDRAM. DDR-I, PC4000 memory is 250 MHz SDRAM that can transfer 16 bytes/cycle (64 bits * 2/cycle) for a data rate of 4 GByte/s."
Double data rate sdram @400mhz is 200mhz X 2bits X64bits=3.2 gigs/sec.DDR sends data at the rise and falling of a clock pulse.SDRam was 1 bit per clock pulse.DDR2 is the same tech as DDR1 other than It uses less power (1.8volts) than standerd DDR1 upto ddr400 (2.5 volts).Your XXXmhz X 4bits is Rdram
DDR2 increases the memory prefetch to four from 2 in ddr1.
But The reason why ddr2 sux is have a cas speed of 4 or 5.U might see so ultra highend ram market that alot of gammes and built your own rig in the pc market @cas 3 but i think its only that low of cas timeing at ddr2 533.Even still u get DDR400 cas 2 samsung awsome TCCD ram chips that can hit DDR600 with better timing than DDR2 can and still cost less and after all AMD goin to support it for awhile longer.
Originally posted by MacRonin
Joke...?!? Who's joking...!?!
..
I figured you had to be.
6 PCI-E slots? What can you put in them? The max you'll need is 2 for SLI, and the rest would need to be PCI, PCI-X, or a combination of both.
I stopped reading after page 7.but most of u guys have no real clue what PCi E is about.
Give it up already.
The push of PCiE is it finally some that will replace PCi\\X that is better and goin to cost less in the end for hardware maker because it is serial based.A x1 slot is like 1/3 the size of a reg pci slot and it gives u 250mb in and out=500mb when pciX 133mhz X 32bits= 533mb and 2 wires carry that other than a crap load of them.
Originally posted by Thomaz
Man u easly slap back don't u.I bet u didn't even read my post did u? just the 1st line.Ya why do u need 6 pciE slots well if u look at PC motherboards there just not all PCIE slots they have your x16 slot or x8 ,a few X4 and a few x1 slots plus your reg pci slots.Apple can put the pciE slots and pci and pci x if they want to thats no prob.
The push of PCiE is it finally some that will replace PCi\\X that is better and goin to cost less in the end for hardware maker because it is serial based.A x1 slot is like 1/3 the size of a reg pci slot and it gives u 250mb in and out=500mb when pciX 133mhz X 32bits= 533mb and 2 wires carry that other than a crap load of them.
#1 what mother board has 6 PCI-E card slots? I've never seen one. I've never seen one with more than two.
#2 How many types of cards are using PCI-E right now?
#3 your article is from May 13, 2004 when many thought PCI-E was going to be used by every card made. It hasn't been. And they probably thought that back then because of that same article.
"Ya why do u need 6 pciE slots well if u look at PC motherboards there just not all PCIE slots they have your x16 slot or x8 ,a few X4 and a few x1 slots plus your reg pci slots "
look at where i put "not" in there.There is pci reg slots.
Plus when u look a good retail Pc board alot of stuff is onboard and they are mostly linked to a pciE link or reg pci.There becomeing less need on the pc side for alot of stuff in card format.u got on board lan or 2,have a good audio chip at times ,sata raid support,bunch of ubs 1.1/2 ports a few firewire ports.heck even hardware firewall in one chipset.
PCIE is goin to be used pretty much everything when the time comes. PCI reg is old it will be supported for quite awhile longer but in the end it needs to die like isa and nubus.The same with agp it been good to use for many years but with video cards getting faster,SLi,muilt gpu's the agp can't do sli and with more than 1 gpu on the card it turns the agp slot into a 66mhz pci slot.
Oh with u never seen a motherboard more than 2 pciE slots u never looked too hard.Most PciE motherboards that not the cheapy ones have 4 pcie slots and 2 pci.