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.
Dual-Core PowerMacs, New G4 PowerBook, and G5 Mini?
- Thread starter MacRumors
- Start date
- Sort by reaction score
ajampam said:
No. Yours is only the 4th reply on a comment that should be avoided on these forums. So let's not turn it into that.
Let's get back on topic: Someone on another site says a researcher gives his opinion on what he thinks Apple will release in the upcoming Apple Expo.
I hope it's the mother of all Powerbooks so I don't have to upgrade for a while. I also hope the Intel Powerbooks don't come out till late 2006/Early 2007 so I won't have to upgrade until the 2nd or 3rd revision Powerbook. I do hope to get an Macintel desktop in between those times.
Also, when's the 5G iPod coming out? I would get a 4G 60gb right now, but I just don't like the white. After I donate my Mini to my sister, I need a replacement.
My birthday is September 19 so a new Powerbook is a good birthday present to myself.
Also, when's the 5G iPod coming out? I would get a 4G 60gb right now, but I just don't like the white. After I donate my Mini to my sister, I need a replacement.
My birthday is September 19 so a new Powerbook is a good birthday present to myself.
Platform said:Even one of the Developer kits........then you could see which one was the fastest
I like the way you think....
Abercrombieboy said:
Piracy is a very sensitive subject for most people. Personally speaking, if you pirate something then sell it then you should go to jail. This is the worst type of piracy as you are making money off other peoples work. This also goes for people who purchase illegal material. You should go to jail for that too, because you are knowingly giving your money to someone who is robbing someone else.
If you pirate software you can clearly afford then well good luck to you in the courts. This is especially centered around companies who make money using the software they pirated while not making money by distributing it.
Although for personal use for people who cannot afford it there is a big gray area. Think of all the people who learnt photoshop on an illegal version and now have enough money to buy the full version and produce great works of art on it. People who pirated windows and now write excellent software for it. Also for microsoft people who pirated windows and office for so long that made it so popular which helped make it become the de facto standard
As for MP3 downloads, i rarely buy cd's and if i couldn't download an MP3 tomorrow i still wouldn't buy many cd's. I mean should you feel bad for not paying for music that you wouldn't have paid money for anyway? Should we all have felt bad recording from the radio all those years ago? I'm not taking money from the artist and i'm not making money from it either... It's a big gray area. I'm not saying it's right but i'm saying it's not the worst either. People who want to buy cd's will still buy cd's.
To look at it another way, the authorities should be going after people who are doing 100MPH through narrow roads and not the people doing 65MPH on a wide road (in a 60MPH zone). It's still wrong to be doing 65MPH but its much worse to be going 100MPH. However it's easier to catch the person going 65MPH because they will be law-respecting (if not fully law abiding) and will probably stop for the police, whereas the 100MPH person might just take a dash for it causing a big police chase and probably not have a license plate on to check. So the people who REALLY break the law get away with it and the people who on occasion break the law (in the gray area) go to jail.
septembers gonna be a hell of a month!!! if those G5 Mini's come out i will be first in line for one!!
AidenShaw said:
All true....
One further clarification though...
Unlike the PowerPC, Intel x64 64-bit FLOATING POINT SUCKS. just like it does in every X86 processor ever made.
A multiply add requires 2 clock cycles on an X86 based processor whereas on the Itanium/Power/Sparc/Mips/Cray/Alpha it only takes 1.
AidenShaw said:
Although they could also currently release a 2.5 to 2.7 GHZ dual core PowerPC chip too. That don't require the specialized cooling.
IBM will be offering 2.5GHZ dual cores on their blade server in Sept.
If 2 dual-cores can be cooled easily on a blade they sure as Hell can be cooled easily in an Xserve or In a PowerMac
AidenShaw said:
but is this in reference to ONLY the processor?
because from what i understand, the architecture G5's allow are it's main innovations. the scalability of the FSB being the most obvious.
taking the architeecture into account (for instance..the 7448's max Fsb of 200mhz v the G5's max of...(varies on the proc..but the LOWEST apple impliments is in the imac at 600..i think.) is the G5 still simply "on par" with the G4?
no, it's "whole system"
Note these charts from barefeats:
There certainly are benchmarks that show the G5 way ahead - some applications really benefit from the memory bandwidth of the G5 (but others are really hurt by the high latency of the G5 memory). (chart below from http://www.anandtech.com/mac/showdoc.aspx?i=2436&p=2)
No, it's from real timing of real applications....Surreal said:
Note these charts from barefeats:
There certainly are benchmarks that show the G5 way ahead - some applications really benefit from the memory bandwidth of the G5 (but others are really hurt by the high latency of the G5 memory). (chart below from http://www.anandtech.com/mac/showdoc.aspx?i=2436&p=2)
Attachments
AidenShaw said:No, it's from real timing of real applications....
Note these charts from barefeats:
There certainly are benchmarks that show the G5 way ahead - some applications really benefit from the memory bandwidth of the G5 (but others are really hurt by the high latency of the G5 memory). (chart below from http://www.anandtech.com/mac/showdoc.aspx?i=2436&p=2)
...but those benchmarks show the G4 Powermac with the extra L3 cache, when you remove that extra cache, performance gets a big hit. So I think surreal has a point there. The G4 itself can complete more instructions per clockcycle, but the system architecture is a bottleneck.
~loserman~ said:
I don't know too much detail about this but taking two clock cycles to multiply would only be bad if there was only one multiply unit. Also the the P4 ALU runs at twice the core clock speed so on a 3.0GHz P4 it is running at 6.0GHz. Two clock cycles doesn't seem all that bad now does it
The Athlon on the other hand has a massively powerful FPU (ALU) unit, a lot faster than the P4 but i don't know how fast compared to the G5.
TBi said:I don't know too much detail about this but taking two clock cycles to multiply would only be bad if there was only one multiply unit. Also the the P4 ALU runs at twice the core clock speed so on a 3.0GHz P4 it is running at 6.0GHz. Two clock cycles doesn't seem all that bad now does it
The Athlon on the other hand has a massively powerful FPU (ALU) unit, a lot faster than the P4 but i don't know how fast compared to the G5.
Not so on any account.
Explanation of Floating Point Units on X86 Processors... whether Intel or AMD
All X86 floating point units behave identically because of compatibility reasons(Their design and instructions are the same)
All of the Intel and AMD CPUs in the last 5 years have 2 FPU per CPU
These FPUs all run at CPU core freq.
Both the AMD and Intel can perform 2 Floating Point Operations per clock cycle because they have 2 floating point units.
Now the difference between them and Itanium, PowerPC, Alpha etc
All of these CPUs have 2 FPU units per core but all except the X86 CPUs can perform up to 4 Floating point operations per clock cycle because if they need to do a multiply + add they can perform it in 1 cycle.
X86 CPUs require 2 cycles to perform it.
Of course all code doesn't need to do Fused multiply+adds but codes that typically do a lot of double precision floating point do.
So in summary for double precision floating point operations
Theoretical MAX
INTEL P4 = 2 per cycle
INTEL Xeon =2 per cycle
Intel Itanium =4 per cycle
AMD Athlon = 2 per cycle
AMD Opteron = 2 per cycle
PowerPC = 4 per cycle
Watts the problem?
The Mac Mini will get the G5 upgrade in September for two reasons...
1. This isn't your daddy's G5. the 970FX comes in at a cool 13 watts.
2. After spending so much time and money trying to overcome the Mother of all Thermal Challenges, do you expect Apple and IBM to walk away from the low power 970? No way. They have too much invested in the new fabrication process etc. The Mac Mini (and maybe the e-Mac) is the only computer in the Apple line-up that the 970FX makes sense for. The PowerMacs are going Dual-core with the MP version (they're not looking back).
While we are on the topic of processor performance the question begs to be asked...
What's better (same clock speed), a 64-bit G5 or a 32-bit dual core Yonah? You hear so much about there being no real benefit for dual core without a multi-threaded app to go with it. Same goes for 64-bit too. That is the decision Apple customers will have to make next May when Yonah comes out.
Maybe Apple should just overclock the new 7448 G4?
The Mac Mini will get the G5 upgrade in September for two reasons...
1. This isn't your daddy's G5. the 970FX comes in at a cool 13 watts.
2. After spending so much time and money trying to overcome the Mother of all Thermal Challenges, do you expect Apple and IBM to walk away from the low power 970? No way. They have too much invested in the new fabrication process etc. The Mac Mini (and maybe the e-Mac) is the only computer in the Apple line-up that the 970FX makes sense for. The PowerMacs are going Dual-core with the MP version (they're not looking back).
While we are on the topic of processor performance the question begs to be asked...
What's better (same clock speed), a 64-bit G5 or a 32-bit dual core Yonah? You hear so much about there being no real benefit for dual core without a multi-threaded app to go with it. Same goes for 64-bit too. That is the decision Apple customers will have to make next May when Yonah comes out.
Maybe Apple should just overclock the new 7448 G4?
This is funny
It is a great misconception when people talk about the heat of a G5 Processor. The G5's do NOT get that hot, they run an average of 12 degrees farenheight warmer than the G4 Processor. The G4 in my iBook easily reaches 150+ degrees and my Pentium 4 desktop can go into the low 200s. The idle temperature of the G5 chip is 138 degrees. This is really not that hot.
The only problem with a G5 is the power consumption, that is why they can't put it into a laptop. They could easily put one into a mac mini, as along as it had an updated powersupply.
yellow said:A G5 Mini? I find that to be a dubious rumor at best. Maybe if it was a regular mini sandwiched between 2 colosal fans.
It is a great misconception when people talk about the heat of a G5 Processor. The G5's do NOT get that hot, they run an average of 12 degrees farenheight warmer than the G4 Processor. The G4 in my iBook easily reaches 150+ degrees and my Pentium 4 desktop can go into the low 200s. The idle temperature of the G5 chip is 138 degrees. This is really not that hot.
The only problem with a G5 is the power consumption, that is why they can't put it into a laptop. They could easily put one into a mac mini, as along as it had an updated powersupply.
csjk789 said:
Power consumption = heat. The reason the G5 runs "cooler" is because it has a much bigger heatsink. Have you seen the size of the heatsink that goes into the G5 tower and it is still noisy. Compare that to the miniature one on the G4 in the mini (or the iBook). Also think about it, why would you need to water cool a cool running chip? The newer 970FX is a lot cooler running though.
My P4 ran cooler than my Athlon XP, but thats only because the P4 has an XP-120 on it whereas the Athlon had a small little copper heatsink on it that just about covered the socket A socket.
To ~loserman~, i didn't know that and thanks for the information. I won't make the mistake in future. Maybe apple are now relying on SSE to make up for the X87's lack of processing power compared to the G4 and G5. Then again, SSE probably isn't as fast as Altivec. Too bad they can't just implement Altivec in new Pentium Processors (SSE A).
G5 Mini not likely, makes me sceptical of the rest
Releasing a G5 Mini right now makes no sense. If they do, it would be up to par, if not very close, with the iMac G5. Follow me here.
iMac 17" 2GHz: $1,500
Mini ($600) + 20" display ($800) : $1,400
So you are saving $100, plus you get the sick 20" display. They would have to either improve the iMac, increase the price of the Mini or decrease the price of the iMac. None of these things seems likely. Any thoughts?
Releasing a G5 Mini right now makes no sense. If they do, it would be up to par, if not very close, with the iMac G5. Follow me here.
iMac 17" 2GHz: $1,500
Mini ($600) + 20" display ($800) : $1,400
So you are saving $100, plus you get the sick 20" display. They would have to either improve the iMac, increase the price of the Mini or decrease the price of the iMac. None of these things seems likely. Any thoughts?
Intel mini in Jan
What I think is that Steve surprises us with a Intel Mini in Jan on MWSF06! Remember this post!
What I think is that Steve surprises us with a Intel Mini in Jan on MWSF06! Remember this post!
If they made the Mac Mini a G5, they would have to make the eMac a G5, because the eMac would be the only G4 left and it would sell like crap..
I do think there will be PowerBook updates. That's a pretty safe bet in my book. I dunno about PowerMacs. I would think they would be the first Intel Macs, and there won't be anymore PPC updates for them, but that's just me.
I do think there will be PowerBook updates. That's a pretty safe bet in my book. I dunno about PowerMacs. I would think they would be the first Intel Macs, and there won't be anymore PPC updates for them, but that's just me.