New & Old dual 1 Gig. PowerMac Compared.
- Thread starter Samuel
- Start date
I don't see a measurable difference between the two honestly (as long as it's the faster DDR Bus and RAM)...
We're getting the new DP 1Ghz vs the Old because it fit the bill perfectly for what we where looking for (we where NOT concidering _ANY_ of the SDR systems).
No one's saving the world in a few nanoseconds (unless, ofcourse you're Superman)
We're getting the new DP 1Ghz vs the Old because it fit the bill perfectly for what we where looking for (we where NOT concidering _ANY_ of the SDR systems).
No one's saving the world in a few nanoseconds (unless, ofcourse you're Superman)
Oscar Castillo
Registered
I think we need to look at benchmarks in a different light. There are benchmarks which measure CPU performance and those which measure I/O. I can't say which if any of these measure I/O since I haven't looked at them closely, but I keep hearing how crippling the 1MB L3 cache is.
Did anyone stop to consider that these tests may mainly measure CPU performance so comparing an 1Ghz G4 with 2MB L3 cache versus new one with 1MB L3 cache you'd expect to get the same results or close to it. And if that is the case, it tells me the 2MB L3 cache doesn't do much in terms of performance.
I'd expect these test to be the same or close to it if that was the case.
What we need to see are memory and hard drive benchmarks to determine if the 133 or 167 system bus and new DDR RAM gives any improvement there.
Did anyone stop to consider that these tests may mainly measure CPU performance so comparing an 1Ghz G4 with 2MB L3 cache versus new one with 1MB L3 cache you'd expect to get the same results or close to it. And if that is the case, it tells me the 2MB L3 cache doesn't do much in terms of performance.
I'd expect these test to be the same or close to it if that was the case.
What we need to see are memory and hard drive benchmarks to determine if the 133 or 167 system bus and new DDR RAM gives any improvement there.
serpicolugnut
OS X Supreme Being
Hmmm. What it tells me is that this upgrade really is no upgrade in any sense of real world speed.
This round of PowerMacs will probably only be bought by people who absolutely need a new machine, or haven't upgraded in a long while. People (like me) who have a G4/800DP machine or better will probably wait for awhile, as these machines offer no real reason to upgrade.
Oscar Castillo
Registered
That's not what I want to imply. Comparing a dual 1GHz 2MB cache with a new 1GHz 1Mb cache and see the same results tells me, if the tests are solely CPU dependent, that the exta 1MB cache in the older duals does nothing for performance.
What I would like to see are subsystem tests all I/O independent of CPU to see if indeed there are any gains.
Disk I/O and memory speed tests in the similar fashion as the SiSoft Sandra benchmarks for the PC.
I'm sure there must be some Unix benchmark programs for OS X by now. Anyone know of any?
Oscar Castillo
Registered
If I had a old dual 1GHz I wouldn't bother ugrading either. For the gains you may see in I/O still aren't known. If you are running a file server then perhaps you may see the difference with the ATA100 drives, but it doesn't make much sense to go through the upgrade.
If the old dual 1GHz were priced better I would have gotten one of those from MacWarehouse to whereever. And the prices on Ebay for a used one is pretty ridiculous. $3000 cause it comes with installed software "you will need to get a license for" as it says in many listings. Some offer nothing more than the basics.
Those that are trying to get rid of 933s are having a hard time deciding on price cause they'e almost squeezed out of the picture.
If the old dual 1GHz were priced better I would have gotten one of those from MacWarehouse to whereever. And the prices on Ebay for a used one is pretty ridiculous. $3000 cause it comes with installed software "you will need to get a license for" as it says in many listings. Some offer nothing more than the basics.
Those that are trying to get rid of 933s are having a hard time deciding on price cause they'e almost squeezed out of the picture.
serpicolugnut
OS X Supreme Being
Well, it does something for performance... It's just that what it does is equally replicated by the inclusion of DDR SDRAM in the system. If DDR SDRAM is a cheaper fix for Apple than a 2MB backside cache, then it makes sense.
Personally, I think if they would have kept the 2MB backside cache along with the inclusion of DDR SDRAM it would have given it a pretty good boost over the previous dual ghz model.
These systems are really just a stepping stone. Apple is on record as being a member of the Hypertransport consortium. The Hypertransport bus is a reality (already employed in the XBOX, I hear), and Apple will use it - it's just a matter of when. With the current bottlenecks being 1) the system bus speed, and 2) the CPU and the speed at which it talks with the system, it's a no brainer that Apple will implement these as soon as it can... And if this round of Powermacs don't sell big (which I doubt they will), then you can count on Apple rushing these improvements through as quickly as it can....
Backside cache really isnt that important, why do you think Athlons and P4's dont bother with it.
Also the reason there isnt really a difference between the DDR and SDR systems is because the FSB is still the same on the Ghz models hence the fsb can still only take PC133 bandwidth yet its given double that using DDR Memory.
If the G4 FSB was DDR aswell like the Athlon it would then be better placed to take advantage of the new DDR memory.
Im affraid this new powermac range was simply another example of Apple using flashy names taking for granted that user would just believe it is faster. I mean it must be faster its DDR right? Give me a brake. It's actually an insult to the inteligence of its user base.
Also the reason there isnt really a difference between the DDR and SDR systems is because the FSB is still the same on the Ghz models hence the fsb can still only take PC133 bandwidth yet its given double that using DDR Memory.
If the G4 FSB was DDR aswell like the Athlon it would then be better placed to take advantage of the new DDR memory.
Im affraid this new powermac range was simply another example of Apple using flashy names taking for granted that user would just believe it is faster. I mean it must be faster its DDR right? Give me a brake. It's actually an insult to the inteligence of its user base.
I've been skeptical for months over this DDR thing...asking this board and others for proof that there is any real advantage, at least today. Finally, proof that DDR is a lot of hype. What I like about those tests is that they are real world tests, not some stupid arcane behind the scenes data rate speed tests, but real life "How long does Photoshop take..." tests. Good job.
erhm... *cough* the new 1 GHz is _not_ the high end machine, i repeat: the new 1 GHz is _NOT_ the high end machine.
please stop comparing those two! if all you see is that you can now get the same (more or less, benchmarks lie) for less, why aren't you happy? and if you want more and have the dough anyway, why not order the 1.25 GHz?
I'm more in the 867 MHz region, if I would buy now, because I'd get a SCREAMING FAST dual processor machine for the money that would have bought me a single processor machine some month ago!
let's look at the benchmarks, when the high end machine drops by, okay?
please stop comparing those two! if all you see is that you can now get the same (more or less, benchmarks lie) for less, why aren't you happy? and if you want more and have the dough anyway, why not order the 1.25 GHz?
I'm more in the 867 MHz region, if I would buy now, because I'd get a SCREAMING FAST dual processor machine for the money that would have bought me a single processor machine some month ago!
let's look at the benchmarks, when the high end machine drops by, okay?
fryke
we know its not the high end machine, we are just critising the fact that Apple have added DDR to there powermacs leading people to beleive it adds the performace benefits seen in the X86 world to the Mac world which it clearly doesnt.
This is shown by comparing one machine of the same speed with another with the difference being the memory.
we know its not the high end machine, we are just critising the fact that Apple have added DDR to there powermacs leading people to beleive it adds the performace benefits seen in the X86 world to the Mac world which it clearly doesnt.
This is shown by comparing one machine of the same speed with another with the difference being the memory.
Who's talking about the high end machine? I'm talking about two DP 1 gigs with and without DDR ram. There is no performance advantage in that scenario. Nuff said.
As for the "high end" machine I can guarantee you it will be the same situation. Yes, it wil be faster than the 1 gigs, duh, but I"m still not convinced DDR is getting me much of anything.
I'll bet the 1.25 gig machine will be, surprise surprise, 15-20% faster and that DDR won't have much of anything to do with that jump. Of course, from this point forward, we'll never be able to compare a 1.25 with and without DDR so all arguments will be pointless.
I'm glad Apple is moving forward with DDR, I guess, if at some point it actually means something, but right now I ain't seeing it.
As for the "high end" machine I can guarantee you it will be the same situation. Yes, it wil be faster than the 1 gigs, duh, but I"m still not convinced DDR is getting me much of anything.
I'll bet the 1.25 gig machine will be, surprise surprise, 15-20% faster and that DDR won't have much of anything to do with that jump. Of course, from this point forward, we'll never be able to compare a 1.25 with and without DDR so all arguments will be pointless.
I'm glad Apple is moving forward with DDR, I guess, if at some point it actually means something, but right now I ain't seeing it.
boi
official breaker of macs.
it seemed just a month ago that people were whining because apple didn't have DDR RAM. now people are whining that it's not any faster than the old computers. hmm... maybe that's why apple didn't really rush it out the door, eh? a few of you are even attacking apple for deceiving the customer, etc.
that is when this conversation turned from just showing that DDR RAM isn't as good as it's hyped up to be to Apple's faults and fallacies. so if you want to look at blaming apple, you have to look at the big picture. you can now get a 1.25ghz machine for a tad less than you could get the dual 1 ghz machine just a few weeks ago.
anyway, thanks for the findings- they're quite interesting.
i thought of another thing: If you did this 'real world test' to find the difference between a pentium IV 1.6 ghz machine and a p IV 2.2 ghz machine, the numbers you get will be really close. why? because this day of ultra quick processors and a boatload of RAM, each subsequent upgrade to computers gets less and less noticable in the 'every day world'. let's do some protein folding on these macs and get some real numbers
that is when this conversation turned from just showing that DDR RAM isn't as good as it's hyped up to be to Apple's faults and fallacies. so if you want to look at blaming apple, you have to look at the big picture. you can now get a 1.25ghz machine for a tad less than you could get the dual 1 ghz machine just a few weeks ago.
anyway, thanks for the findings- they're quite interesting.
i thought of another thing: If you did this 'real world test' to find the difference between a pentium IV 1.6 ghz machine and a p IV 2.2 ghz machine, the numbers you get will be really close. why? because this day of ultra quick processors and a boatload of RAM, each subsequent upgrade to computers gets less and less noticable in the 'every day world'. let's do some protein folding on these macs and get some real numbers
boi
Im afraid the Mac G4's tend to suck at stuff like seti and protein folding, ive compared them to my athlon XP many times and they get smoked hopelessly. I dont think there is any altivec code in the clients but the gaps is really that big that even altivec couldnt close the gap to anything near being respectful.
I too thought they would be fast but nope, they arent.
Im afraid the Mac G4's tend to suck at stuff like seti and protein folding, ive compared them to my athlon XP many times and they get smoked hopelessly. I dont think there is any altivec code in the clients but the gaps is really that big that even altivec couldnt close the gap to anything near being respectful.
I too thought they would be fast but nope, they arent.