Covergence?
Lame PM updates....rumoured iBook updates....maybe they are converging the two lines?
Lame PM updates....rumoured iBook updates....maybe they are converging the two lines?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
More PowerMac G5 Details...
- Thread starter MacRumors
- Start date
- Sort by reaction score
Maxiseller said:
You are quite right about most you said. But comparing the PBs with the PMs is just wrong. Your PB runs at 867MHz and 18 month later we are at 1.67 GHz. And the FSB ist really slow at 166 MHz. Thats much slower then the competition.
What would I like to see in future PowerMacs? It's this:
Systemwide: Move to PCI-Express. Maybe offer SLI-capability.
CPU:s: Dualcore and multithreading (two threads per core). IBM knows multithreading, so it can be done. Also, an integrated 128bit DDR2 mem-controller. This would give the system alot more effective bandwidth, since mem-access wouldn't eat in to FSB. FSB could be dedicated to PCI-E, HD's, USB/FW etc. etc. Also, having 2x CPU's (not cores) would double the effective mem-bandwidth.
Multithreading is good thing to have. While true multiprocessing is faster (MP gives 60-90% performance-boost in MP-aware apps) it's also alot more expensive (adding another CPU doubles the number of transistors). Multithreading would give 20-40% boost in MP-aware apps, and it would increase the number of transistors by about 5-15%. It makes sense.
As to the systems:
Low-end: 1x singlecore multithreading G5 @ 2.0GHz
When compared to the rumoured low-end PM (with 2x 2.0 GHZ G5's), this system would be faster in some cases and a bit slower in others. This system would have superior memory-subsystem, which could compensate for the lack of second CPU. Multithreading would also help. This system could handle two threads simultaneously. Price would be lower than on the thinksecret-machine, since this has only one CPU.
Mid-range: 1x Multithreading dualcore G5 at 2.4GHz.
Compared to the rumoured 2x 2.3GHz G5's, this system would be alot faster than the one Thinksecret reported on. Not only does it have better mem-subsystem (due to integrated mem-controller), it would have a bit more raw CPU-power. And it could handle four threads simultaneously (two cores, two threads per core) as opposed to 2 on the other model.
Hi-end: 2x multithreading dualcore G5's at 2.8Ghz
This system would kick ass and take names! It could handle 8 threads simultaneously and it would have twice as much mem-bandwidth when compared to the other PM-configurations (which would already have alot more bandwidth and alot lower latencies than current PM's do). This model might need liquid-cooling though.
What's notable in these systems is that the performance goes up quite fast as you move up in the model-line. Not only do better models feature faster CPU's, they could also handle more threads simultaneously than the lesser models do. iMac could use non-multithreading G5's at around 2Ghz, in order to differentiate between the two model-lines.
Would it take quite an effort to re-design PM in to these specs? yes. Is it doable? Yes! Would it make PM competetive again? Hell yes! Apple needs to move to PCI-Express quite soon, and integrated mem-controllers just make alot of sense.
Is this something you would like to see?
Systemwide: Move to PCI-Express. Maybe offer SLI-capability.
CPU:s: Dualcore and multithreading (two threads per core). IBM knows multithreading, so it can be done. Also, an integrated 128bit DDR2 mem-controller. This would give the system alot more effective bandwidth, since mem-access wouldn't eat in to FSB. FSB could be dedicated to PCI-E, HD's, USB/FW etc. etc. Also, having 2x CPU's (not cores) would double the effective mem-bandwidth.
Multithreading is good thing to have. While true multiprocessing is faster (MP gives 60-90% performance-boost in MP-aware apps) it's also alot more expensive (adding another CPU doubles the number of transistors). Multithreading would give 20-40% boost in MP-aware apps, and it would increase the number of transistors by about 5-15%. It makes sense.
As to the systems:
Low-end: 1x singlecore multithreading G5 @ 2.0GHz
When compared to the rumoured low-end PM (with 2x 2.0 GHZ G5's), this system would be faster in some cases and a bit slower in others. This system would have superior memory-subsystem, which could compensate for the lack of second CPU. Multithreading would also help. This system could handle two threads simultaneously. Price would be lower than on the thinksecret-machine, since this has only one CPU.
Mid-range: 1x Multithreading dualcore G5 at 2.4GHz.
Compared to the rumoured 2x 2.3GHz G5's, this system would be alot faster than the one Thinksecret reported on. Not only does it have better mem-subsystem (due to integrated mem-controller), it would have a bit more raw CPU-power. And it could handle four threads simultaneously (two cores, two threads per core) as opposed to 2 on the other model.
Hi-end: 2x multithreading dualcore G5's at 2.8Ghz
This system would kick ass and take names! It could handle 8 threads simultaneously and it would have twice as much mem-bandwidth when compared to the other PM-configurations (which would already have alot more bandwidth and alot lower latencies than current PM's do). This model might need liquid-cooling though.
What's notable in these systems is that the performance goes up quite fast as you move up in the model-line. Not only do better models feature faster CPU's, they could also handle more threads simultaneously than the lesser models do. iMac could use non-multithreading G5's at around 2Ghz, in order to differentiate between the two model-lines.
Would it take quite an effort to re-design PM in to these specs? yes. Is it doable? Yes! Would it make PM competetive again? Hell yes! Apple needs to move to PCI-Express quite soon, and integrated mem-controllers just make alot of sense.
Is this something you would like to see?
10 month for a 200MHZ upgrade????
Come on IBM, you can do better!!!!
We are gonna see the same situation as before with the PB line, a .17Ghz boostie!!
Lets hope its not true, vey dissapointing!!
256MB ram Standard!!!!!
GO APPLE GO!
Come on IBM, you can do better!!!!
We are gonna see the same situation as before with the PB line, a .17Ghz boostie!!
Lets hope its not true, vey dissapointing!!
256MB ram Standard!!!!!
GO APPLE GO!
AidenShaw said:Apple: DDR 400 MHz dual channel (PC3200)
Dell: DDR2 533 MHz dual channel (PC4200)
So, Apple is using an 800 MHz memory bus, and Dell a 1066 MHz.
What's the point in bragging about an FSB that is sitting idle waiting for 800 MHz memory?
(ps: and the 2.7 might have a 900 MHz FSB - which would be just as fast at waiting for 800 MHz memory
DDR2-533 Memory has much higher latencies than DDR memory. For a lot of applications that will hurt DDR2 utilising systems. AMD have said that they will not use DDR2 until 667MHz/800MHz speeds are standard and good value, because the latencies are simply too high until 667MHz.
Each core in a current PowerMac has full bandwidth to the northbridge, enough to carry memory traffic and cache coherency traffic and more. The dual-core P4 shares the FSB, meaning an effective 533MHz FSB per core (400MHz for the standard Pentium D). The P4 dies a miserable death of performance when its bus bandwidth is cut.
AMD have done it right by integrating the northbridge into the processor and bypassing the FSB bottleneck completely. At least IBM have a bus per processor. Intel dual-core means sacrificing performance, both because of heat (hot P4 processors - two of them means slower processors) and bus bandwidth sharing (although this isn't different from current dual processor Xeon systems, it is a general inefficiency that only aids cache snooping).
nexusfx said:
You talk about x86 (instruction-set), and then talk about how eliminating northbridge was a great idea. What does northbridge have to do with the CPU's instruction-set? And then you talk about specific CPU-architectures (pipeline-stages). Again, that has nothing to do with the instruction-set. And, IIRC G5 has pretty long pipeline as well. Number of pipelines is more a feature of particular CPU-families and not the instruction-set. Athlons have alot less pipeline-stages than P4 does, yet they are both x86 (A64 is x86-64 as well)
I don't care how "crappy" x86 is. I care about the performance. why is it that Pentium/Athlon64 (with their "crappy" instruction-set), compete just fine with G5 (with it's "superior" instruction set)?
You talk about bandwidth. Funny, Athlon64 is alot smarter when it comes to bandwidth that G5 is. In G5, when the CPU wants to talk with the RAM, it first talks with the northbridge (sharing the FSB-bandwidth with other devices), northbridge talks with the RAM, RAM then talks with the northbridge which then talks with the CPU. That wastes bandwidth and it takes a long time. Contrast that to Athlon64, where the CPU talks directly with the RAM using a dedicated channel, leaving the FSB to other devices.
LMAO
man, i really wanna know what jobs you guys do, that requires a dual core 3ghz powermac with blu-ray
man, i really wanna know what jobs you guys do, that requires a dual core 3ghz powermac with blu-ray
Sharky II said:
Well, faster CPU's would make thing like rendering, compiling and so forth faster. Most people do not NEED uber-fast machines. But if you use the computer for a living, you want it to be as good as possible. And if you have to occasionally wait around for the computer to finish it's task, faster CPU does make sense.
Market effects
Alot of you are saying "oh well, we will just have to wait" as a soothing reply to all of the people complaining that the rumored updates suck.
But a very real problem here is if people dont buy powermacs this quarter, apple will have to come up with sales elsewhere.
In case any of you dont follow the market, Apple last week announced really impressive sales last quarter http://www.apple.com/pr/library/2005/apr/13results.html - including 43% more CPU units than last year. And now of course everyone expects them to do better. Apple estimated slightly higher sale $$ for this quarter. If their guess on Powermac sales is way off, it will hurt the stock price and make life less fun for management. If noone buys them, the margins will have to go down to push them out the door. Thats really bad.
Based on the relatively low guidance Apple gave on sales increases this coming quarter, its possible they know how bad this update sucks, and are prepared to not sell many at all (as they know people will wait - the powermac business for them is highly upgrade-dependent). Momentum is important, and if everyone waits and all the sales get pushed to the next quarter, then they wont have enough and everyone will grumble about having to wait a month for their dual core G5. Its a lose-lose.
anyway - just so everyone realizes that this is actually a pretty important issue - and that people inside Apple certainly realize the implications of it. It also means those of you who are trying to "stop the rush of pitchforks to the castle" might not realize how bad a blip like this can get.
Alot of you are saying "oh well, we will just have to wait" as a soothing reply to all of the people complaining that the rumored updates suck.
But a very real problem here is if people dont buy powermacs this quarter, apple will have to come up with sales elsewhere.
In case any of you dont follow the market, Apple last week announced really impressive sales last quarter http://www.apple.com/pr/library/2005/apr/13results.html - including 43% more CPU units than last year. And now of course everyone expects them to do better. Apple estimated slightly higher sale $$ for this quarter. If their guess on Powermac sales is way off, it will hurt the stock price and make life less fun for management. If noone buys them, the margins will have to go down to push them out the door. Thats really bad.
Based on the relatively low guidance Apple gave on sales increases this coming quarter, its possible they know how bad this update sucks, and are prepared to not sell many at all (as they know people will wait - the powermac business for them is highly upgrade-dependent). Momentum is important, and if everyone waits and all the sales get pushed to the next quarter, then they wont have enough and everyone will grumble about having to wait a month for their dual core G5. Its a lose-lose.
anyway - just so everyone realizes that this is actually a pretty important issue - and that people inside Apple certainly realize the implications of it. It also means those of you who are trying to "stop the rush of pitchforks to the castle" might not realize how bad a blip like this can get.
As I also stated in the other powermac thread (which goes on for another 20+ pages) : I don't think there will be any upgrade in the next 2 weeks.
New Powermacs will ship with Tiger, so why announce them now ? By the way : on the Apple Store every Mac is still listed as shipping "same business day" : I guess the stocks also have to be depleted...
New Powermacs will ship with Tiger, so why announce them now ? By the way : on the Apple Store every Mac is still listed as shipping "same business day" : I guess the stocks also have to be depleted...
Don't know if this has been mentioned, so far ...
Seems alot of people are hoping that Apple have been feeding TS misinformation about PM updates, as part of some retribution on TS leaks in the past ...
Isn't there a law against a company deliberately misinforming the market about it's business in this way????
Seems alot of people are hoping that Apple have been feeding TS misinformation about PM updates, as part of some retribution on TS leaks in the past ...
Isn't there a law against a company deliberately misinforming the market about it's business in this way????
sw1tcher said:
Err, the first one was Intel bringing HyperThreading down the entire P4 line from higher clock speeds.
Yep. From http://www.intel.com/pressroom/archive/releases/20021114comp.htm (November 2002) Intel had a 3.06GHz HyperThreading enabled processor on the market.
So you could say that in 2 and a half years, Intel has merely increased their clock speeds by 733MHz to the Pentium 4 570J 3.8GHz, or 24%. In fact, any bus speed improvements and cache size increases in that time have been negated by moving to the more inefficient Prescott core. That is truly pitiful.
Since June 2003 Intel have had a 3.2GHz HT enabled P4 processor. So in pretty much the same time that Apple have been selling G5 machines, Intel have managed a 19% clock increase, whilst Apple have gone from 2GHz to 2.7GHz, a 35% increase.
Things are a little better if you start talking about Xeons though, they've gone from 2.8GHz to 3.6GHz, or a 29% clock speed increase with significant bus and cache improvements. Still, Apple have remained competitive in the 2P arena.
ACED said:
I don't think so. I mean, Apple is not officially telling anyone anything. If some third-party decides to leak information that ends up being false, I fail to see how that is Apples problem.
Hattig said:
Well, you just repeated what I said 2 pages ago
If you look at AMD by the way : they introduced their FX51 @ 2.2Ghz in august-september 2003, and are today @ 2.6Ghz (soon to be 2.8Ghz). Also "only" a 27% increase in 2 years (doing the math with the 2.8Ghz FX-57).
Evangelion said:
... but people are in such disbelief about the rumoured unwhelming PM update, that they nelieve Apple must have 'feed' them misinformation, for them to release.
This, I believe, would be a very serious breach of corporate law, since it very well would influence its own share price, and mislead investors.
Doesn't matter, if its official leaks, (or not)!!!
ACED said:
If you deliberately misinform staff who are bound by an NDA to not go outside the company then I think you would find it very hard to prosecute the company, especially if the intention is to uncover persons known to be leaking secrets to the market.
ts1973 said:Well, you just repeated what I said 2 pages ago
If you look at AMD by the way : they introduced their FX51 @ 2.2Ghz in august-september 2003, and are today @ 2.6Ghz (soon to be 2.8Ghz). Also "only" a 27% increase in 2 years (doing the math with the 2.8Ghz FX-57).
Nothing wrong in reiterating stuff! Some people here seem to require it to get real facts drilled into their heads even. They haven't realised that it isn't only Apple that are stuck in a clock speed increase rut, everyone is.
Actually, I did forget to mention that the real speed increased on the P4 line have actually come from software optimisation for SSE2. Hence the good media processing benchmarks the P4 shows quite regularly. I suppose that Tiger and more recent Mac software will start to fight back against these optimisations however.
The issue isn't these new PowerMacs, it is whether or not the next PowerMac will be in 6 months time, or in 10 months time, and what they will include. I certainly think that this is just a standard specification upgrade that Apple should probably have done a couple of months ago, and the real news will be in October/November which is a more reasonable time to expect IBM to be able to release dual-core processors.
Look on the bright side, by the time consumer windows software actually takes advantage of the second core, the 970MP should be out.
cjeukens said:What is going to happen to the single processor G5
Is it going to be discontinued
I was on the verge of buying the G5 single 1.8 when I read the negative buyer advise. What is going to be the "logical" equivalent (when I want to have Mac separate from screen).
Coen
I don't know. The single 1.8 sold pretty poorly. Then again the entire PowerMac line sold pretty poorly. Hopefully they release a smaller single processor PowerMac Mini or cube, but then again those would potentially cut into iMac sales.
Most likely there is someone who feeds ThinkSecret and I'm sure Apple wants to know who's the culprit. One way to find that person is to "leak" different configs to the persons who you think it could be and see which config will finally find it's way on TS. Pure speculation but it's a possibility.
ts1973 said:Well, you just repeated what I said 2 pages ago
If you look at AMD by the way : they introduced their FX51 @ 2.2Ghz in august-september 2003, and are today @ 2.6Ghz (soon to be 2.8Ghz). Also "only" a 27% increase in 2 years (doing the math with the 2.8Ghz FX-57).
You could look at it differently as well. A64 was introduced at 2Ghz in august-september, and today it runs at 2.4GHz (3800+, using the new Venice core). But besides Mhz, they have improved the mem-controller, added SSE3-support, doubled the mem-bandwidth, increased the FSB to 1GHz (from 800Mhz).
So, while the actual increases in Mhz has been modest, the actual performance has been going up. But on the G5 there has been no other performance-improvements, besides increased clock-speeds.
And if you look at Opteron.... It started at 1.4Ghz in April 2003 and it's today at 2.6Ghz
. And it too has received mem-controller improvements and the like (although no SSE3 yet IIRC). So it has had increase of 85% in two years.
For me and I suspect those who are frustrated by this I guess its just a matter of timing. I haven't been able to afford a desktop for 3 years, my main computer is a Tibook, I bought that because I wanted both a laptop and a desktop and also didn't really want a PM which sounded like concorde taking off in the house! previous to that I had a 7500 with a G3 card in it! It took me a long time to get the cash together, and might be another god knows how many more years until I could buy another PM so while of course I'd be able to do much more intensive 3D work than with my current kit it would be kinda gutting to get a current liquid cooled behemoth when something much better, faster next gen and most importantly - more future proof is not far off. I cannot wait forever, so if I have to get a minor speed bumped PM then I will probably go for the cheapest I can instead of my originally anticipated top of the line - so instead of being hugely exciting after all this time will be a bit of an anticlimactic "oh well there you go I have another computer". If quads are announced before or at WWDC then my prayers would be well and truly answered.