Actually, they're below 200k, and they recently said that they don't believe it will ever get above 200k.
On a related note: [this was linked to from the AI home page, but I've copied it here for convenience].
Analyst: 11% of Windows iPod users to buy a Mac
Feb 24 - 12:00 pm EST In a research note released to clients on Thursday and obtained by AppleInsider, Needham & Co. analyst Charlie Wolf raised his target price on Apple Computer to $104 a share. The analyst reiterated his assumption that 11% of Windows users owning iPods will purchase a Mac, and notes that sales at the iTunes Music Store have accelerated to a $450 million annual run rate. "The increase in music sales in combination with a decrease in peripheral sales in our revised model raises Apple's overall gross profits because music sales are much more profitable than peripheral sales," Wolf said. In addition to raising its Apple price target from $83 to $104, the firm also raised its earnings per share estimates from $1.85 to $2.00 in fiscal 2005, and from $2.25 to $2.50 in 2006. Needham maintains a 'Buy' rating on AAPL.
I didn't know that quarterly sales of iPods are now at 860K! If 80% of iPod users are on Win32 and 11% of them buy a mac [last year, the iPod attributed switch rate was 6%]--that's about +75k Mac/quarter.
That sales boost on hardware would definitely be nice--and with the Macs "Staying Power", that increases the units-in-field nicely.
Actually, they're below 200k, and they recently said that they don't believe it will ever get above 200k.
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Apples keynote on the G5 was probably one of the most purposely misconstrued, and dishonorable representations of ones self I've ever seen. It was shameful. I found it hard to believe they would stoop to that level. A few friends told me that the keynote alone lost them as a Mac user.
I do hold out hopes, but from what most people think that Apple intends on for the next PowerMac upgrades it doesn't look good. They should have some success with the new OS, and I expect sales will be great for most Macs because of it's popularity, but the machine (powermac) itself will probably still be a second rate pro machine after it's updated, and sales for the PowerMac will probably keep falling because of it.
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Apples keynote on the G5 was probably one of the most purposely misconstrued, and dishonorable representations of ones self I've ever seen. It was shameful. I found it hard to believe they would stoop to that level. A few friends told me that the keynote alone lost them as a Mac user.
I do hold out hopes, but from what most people think that Apple intends on for the next PowerMac upgrades it doesn't look good. They should have some success with the new OS, and I expect sales will be great for most Macs because of it's popularity, but the machine (powermac) itself will probably still be a second rate pro machine after it's updated, and sales for the PowerMac will probably keep falling because of it.
I'm sorry to say I completely agree with you.
As is stands, Apple had better damn well absolutely stun me with their next Powermac, or I'm buying a mac mini, a nice LCD, and then piece together a killer linux box.
It's really gotten that bad. I used to be really excited that apple could push the market (yeah yeah, I know it's all about IBM's supply) but the point still remains: Apple is (almost) always playing second fiddle in the hardware game. The marketing value of Apple being able to (truthfully) say that "We have the fastest desktop in the world" is very important. The OS is unparalleled, this is true, but you have to have the hardware to go with it. These two together would be a killer combination. Apple might make a living off of the low-end switchers for a while, but they can't make the mistake of losing their high-end customers as well.
I agree with onlooker about the "power" argument... however I DO NOT agree with the g5 keynote.
There was nothing misleading in that keynote. All the tests they presented during that keynote are 100% correct. Floating point integer tests smoked the current line up of amd and intel chips. DNETC will prove that even today.
You're beginning to sound like a broken record, powermacs aren't powerful enough... their cpu speeds are comparible if not faster than current x86 chips! Do your own benchmarking instead of relying on others to do it for you. And do real software independent benchmarking... That means, no photoshop, no cinema, no maya. Just PURE RAW CALCULATING / CPU INSTRUCTION EXECUTION benchmarking. That means programs like DNETC, GCC, etc. You will plainly see that the 970 family is more often faster than almost any other chip on the x86 side.
So you'll say, well that isn't real world. You're right its not. And developers are 100% to blame for that... I am one I know how it goes, programmers don't really care when porting to make it 5 ticks of a second faster unless there is money in that. But after benchmarking you will plainly see that saying powermacs aren't fast enough, is plainly false.
They are plenty fast, just need better funding and more time for development (software).
Most people do not need a powermac, why? Because they are TOO MUCH POWER for them to justify. I never once maxed out my dual 2.0 g5 unless I was benching or rendering. There is something to be said for that. So why pay that much money on a machine when you don't need that much power and can get away with using iMac power? (Of course people don't know this until after they buy the machines... but this is my experience).
There was nothing misleading in that keynote. All the tests they presented during that keynote are 100% correct. Floating point integer tests smoked the current line up of amd and intel chips. DNETC will prove that even today.
"There was nothing misleading in that keynote"
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
If you want to split hairs (by not using the exact same compiler ie gcc) than compare intels compiler with ibm's 64bit compiler... you will surely see the ibm compiler is nearly 2 times as fast comparing a dual 2.5 and a p4 3.6 (I haven't ever owned a xeon so I can't speak for them).
That's really expensive from most manufacturers. Like $2,500 just for the board. It's more than double the price of a dual. I think an Xserve w/4 processors is something that only Pixar, and the rest could afford. But I think a Xserve blade sounds cooler if your going to go that route.
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
This was argued endlessly on many boards, but what the heck, I checked out the website you linked to. I stopped reading when I got to his statement that hyperthreading was disabled delibrately to hedge results in PPC's favor. Even Intel admits that for Spec testing, disabling hyperthreading IMPROVES Spec scores. And no, I'm not about to search down the link to Intel's site that mentions this, but I did read it and it was on Intel's website.
As far as hedging tests, they all do it. For the most part comparing Spec test results between different architectures is difficult at best and down right useless at worst. Spec should be used by engineers comparing the results between similar cpu architectures for analysis of design changes. Compilers have WAY TOO much effect on the results to infer which architecture is better or worse, or CPU will perform better in real world apps.
I stopped reading when I got to his statement that hyperthreading was disabled delibrately to hedge results in PPC's favor.
You shouldn't have stopped reading. Maybe then you would understand the level of deceit used in the tests. That was just one example page, and the first I found on the subject. The author admittedly wasn't aware of what all the tests were, but to be fair reported everything changed in the tests. Also there are results on that page that show the differences in performance after things were change back to normal. I don't think there was one test that the G5 actually came out on top in.
I don't think there was one test that the G5 actually came out on top in.
Blast? Yes, I know, most people don't run Blast at home. But still one for the home team.
I don't want to get into a rehash of this old topic. Yes, Apple skewed test results. This article blew much of it way out of proportion and a lot of Mac zealots reacted very poorly in responding to this article.
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Ok, forgive me if I have to laugh at this.
You say you've had three PowerMacs. Presuming that at least one or two of them were in the G3 or G4 range you're telling me that you were happier with the old 133-167 mhz front side buses than you are with the current 900mhz-1.25ghz ones?
I'm not about to start comparing high end Mac's and PC's because there's too much to factor in for both sides to make a valid comparison that will satisfy everyone, but I will say that the G5 was a quantum leap forward in hardware architecture for the PowerMac compared to earlier versions and I can't quite comprehend how you were happy with old PowerMac performance, but are disappointed by the current G5's.
Onlooker is heavy into 3D and he's right. If you want the fastest 3D then right now PCs are going to win hands down. Apple could ameliorate this situation by actively promoting ATI and Nvidia's higher end cards and helping with driver development. Maybe that'll happen with Tiger..who knows.
My particular needs for a Powermac G5 would be Audio and Video so I'm not in the same boat. The G5 kicks enough butt there to keep me happy.
Onlooker is heavy into 3D and he's right. If you want the fastest 3D then right now PCs are going to win hands down. Apple could ameliorate this situation by actively promoting ATI and Nvidia's higher end cards and helping with driver development. Maybe that'll happen with Tiger..who knows.
My particular needs for a Powermac G5 would be Audio and Video so I'm not in the same boat. The G5 kicks enough butt there to keep me happy.
Hmurchison,
I'm not disagreeing that PC's certainly have definite speed advantages over Mac's in certain areas, 3D being one of them.
The part of Onlooker's post that confused me was his satisfaction with the performance of his three previous PowerMacs, I can only assume he was satisfied since he kept buying them, but his disappointment with the current G5's.
As I said in my previous post, PC's and Mac's have certain strengths and weaknesses in performance when compared to one another and there's no one test that will satisfy everyone.
So I guess what I'm saying is that if Onlooker is disappointed with PowerMac performance in comparison to PC's then that is understandable, but if his disappointment is with the performance of current PowerMacs to previous ones, then that I don't understand and from the way his post was worded that certainly seemed to be one of the things he was complaining about.
P.S. I also do Video work and am waiting patiently for the next rev. G5's to upgrade my 867 G4 Quicksilver, though to be honest the 867 still does a damn fine job with FCP and DVDSP.
You shouldn't have stopped reading. Maybe then you would understand the level of deceit used in the tests.
Not really.
You assume that gcc for the Mac is as optimized for the hardware as ICC is for x86. And you'd be horribly wrong. All the benchmarks that show "normality" demonstrate is that a wildly well-optimized compiler will beat a poorly optimized compiler. But most applications are compiled with MS' compiler, not Intel's, so it's not necessarily a real-world test to use ICC.
On top of that, at the time Apple ran the tests, gcc was better optimized for x86 than for PPC, by a margin, simply because people had been working on the problem.
Nevertheless, what Apple effectively did was take the compiler out of the equation by using the same one for both platforms. This is standard practice for benchmarking: You compare Premiere on both platforms. You don't compare Premiere to FCP. Why should compilers be any different?
If you want to use ICC for the Intel side, you'd have to use something comparably optimized for the PPC on the Mac side to have a fair comparison.
P.S. I also do Video work and am waiting patiently for the next rev. G5's to upgrade my 867 G4 Quicksilver, though to be honest the 867 still does a damn fine job with FCP and DVDSP.
Sweet! Then you're eagerly waiting for NAB 2005 as I am. I'm most excited to see the first Final Cut Pro version that I will buy. I've used FCP very sparingly since version 2
Oh noooooooooooooooooooooooooooooooooooooooooo
not the infamous Haxial garbage. As Rickag say..that bozo was debunked a long time ago. Amorph is right, the guy ranted about compilers but Apple chose GCC because at least you had the same compiler to test. I doubt GCC was even close to be optimized for PPC as it was for X86 hence the improvement we saw with the IBM XLC/XLF compilers. If high scores are what you value the most then you use a tweaked compiler. If a fair test is what you want you use a compiler that is available on both platforms.
You know we've heard this myth for years about how fast PCs are. That'll last about as long as it takes to build a PC and say "well it doesn't fell 'that' much faster than a fast Mac"
I felt duped myself..I noticed that my PCs did well when I was letting them crunch one task at a time but as soon as I wanted to start tossing in new tasks they degraded pretty fast.
You want to benchmark a computer right..develop a "real world" task where you utilize the computer in a scenario that is likely to happen. Something like compressing video while surfing and burning a CD. The one thing about Macs that people love is that they degrade gracefully under load. Unix baby!
I love the iPods success because I don't have to care about whether someone "gets it". If a PC is so great then go buy one. But you have to wonder why the most exciting thing happening in PCs nowadays is modding them with plexiglass windows adding colors and lights. Everything "but" computer performance. IMO.
Comments
Originally posted by the cool gut
Actually, they're below 200k, and they recently said that they don't believe it will ever get above 200k.
On a related note: [this was linked to from the AI home page, but I've copied it here for convenience].
Analyst: 11% of Windows iPod users to buy a Mac
Feb 24 - 12:00 pm EST In a research note released to clients on Thursday and obtained by AppleInsider, Needham & Co. analyst Charlie Wolf raised his target price on Apple Computer to $104 a share. The analyst reiterated his assumption that 11% of Windows users owning iPods will purchase a Mac, and notes that sales at the iTunes Music Store have accelerated to a $450 million annual run rate. "The increase in music sales in combination with a decrease in peripheral sales in our revised model raises Apple's overall gross profits because music sales are much more profitable than peripheral sales," Wolf said. In addition to raising its Apple price target from $83 to $104, the firm also raised its earnings per share estimates from $1.85 to $2.00 in fiscal 2005, and from $2.25 to $2.50 in 2006. Needham maintains a 'Buy' rating on AAPL.
I didn't know that quarterly sales of iPods are now at 860K! If 80% of iPod users are on Win32 and 11% of them buy a mac [last year, the iPod attributed switch rate was 6%]--that's about +75k Mac/quarter.
That sales boost on hardware would definitely be nice--and with the Macs "Staying Power", that increases the units-in-field nicely.
Originally posted by the cool gut
Actually, they're below 200k, and they recently said that they don't believe it will ever get above 200k.
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Apples keynote on the G5 was probably one of the most purposely misconstrued, and dishonorable representations of ones self I've ever seen. It was shameful. I found it hard to believe they would stoop to that level. A few friends told me that the keynote alone lost them as a Mac user.
I do hold out hopes, but from what most people think that Apple intends on for the next PowerMac upgrades it doesn't look good. They should have some success with the new OS, and I expect sales will be great for most Macs because of it's popularity, but the machine (powermac) itself will probably still be a second rate pro machine after it's updated, and sales for the PowerMac will probably keep falling because of it.
Originally posted by onlooker
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Apples keynote on the G5 was probably one of the most purposely misconstrued, and dishonorable representations of ones self I've ever seen. It was shameful. I found it hard to believe they would stoop to that level. A few friends told me that the keynote alone lost them as a Mac user.
I do hold out hopes, but from what most people think that Apple intends on for the next PowerMac upgrades it doesn't look good. They should have some success with the new OS, and I expect sales will be great for most Macs because of it's popularity, but the machine (powermac) itself will probably still be a second rate pro machine after it's updated, and sales for the PowerMac will probably keep falling because of it.
I'm sorry to say I completely agree with you.
As is stands, Apple had better damn well absolutely stun me with their next Powermac, or I'm buying a mac mini, a nice LCD, and then piece together a killer linux box.
It's really gotten that bad. I used to be really excited that apple could push the market (yeah yeah, I know it's all about IBM's supply) but the point still remains: Apple is (almost) always playing second fiddle in the hardware game. The marketing value of Apple being able to (truthfully) say that "We have the fastest desktop in the world" is very important. The OS is unparalleled, this is true, but you have to have the hardware to go with it. These two together would be a killer combination. Apple might make a living off of the low-end switchers for a while, but they can't make the mistake of losing their high-end customers as well.
There was nothing misleading in that keynote. All the tests they presented during that keynote are 100% correct. Floating point integer tests smoked the current line up of amd and intel chips. DNETC will prove that even today.
You're beginning to sound like a broken record, powermacs aren't powerful enough... their cpu speeds are comparible if not faster than current x86 chips! Do your own benchmarking instead of relying on others to do it for you. And do real software independent benchmarking... That means, no photoshop, no cinema, no maya. Just PURE RAW CALCULATING / CPU INSTRUCTION EXECUTION benchmarking. That means programs like DNETC, GCC, etc. You will plainly see that the 970 family is more often faster than almost any other chip on the x86 side.
So you'll say, well that isn't real world. You're right its not. And developers are 100% to blame for that... I am one I know how it goes, programmers don't really care when porting to make it 5 ticks of a second faster unless there is money in that. But after benchmarking you will plainly see that saying powermacs aren't fast enough, is plainly false.
They are plenty fast, just need better funding and more time for development (software).
Most people do not need a powermac, why? Because they are TOO MUCH POWER for them to justify. I never once maxed out my dual 2.0 g5 unless I was benching or rendering. There is something to be said for that. So why pay that much money on a machine when you don't need that much power and can get away with using iMac power? (Of course people don't know this until after they buy the machines... but this is my experience).
Originally posted by emig647
Most people do not need a powermac, why? Because they are TOO MUCH POWER for them to justify.
I don't think that's it. They are too expensive to justify.
Originally posted by emig647
There was nothing misleading in that keynote. All the tests they presented during that keynote are 100% correct. Floating point integer tests smoked the current line up of amd and intel chips. DNETC will prove that even today.
"There was nothing misleading in that keynote"
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
This is just one of the various summaries of how many ways apple was misleading in those tests
but keeping it that way was extremely time consuming and expensive.
One of these days I'd like to see a more meaningful test that compares
the overall extended reliability of a machine running CPU intensive programs for weeks on end without failure.
This is where I feel Apple would prove it's worth.
It's one thing to take a lap and another to finish the race.
Originally posted by onlooker
"There was nothing misleading in that keynote"
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
This is just one of the various summaries of how many ways apple was misleading in those tests
If you want to split hairs (by not using the exact same compiler ie gcc) than compare intels compiler with ibm's 64bit compiler... you will surely see the ibm compiler is nearly 2 times as fast comparing a dual 2.5 and a p4 3.6 (I haven't ever owned a xeon so I can't speak for them).
EDIT
In floating point calculations
Originally posted by benzene
It's really gotten that bad. I used to be really excited that apple could push the market (yeah yeah, I know it's all about IBM's supply)
I don't think it really is about IBM's supply. The disappointment I sometimes feel with the PowerMac line is the lack of additional features like:
- Another internal hard drive bay
- Hardware RAID on the mobo for drives 2 & 3 (see above)
- Additional built-in FireWire channels (not ports, channels)
- More advanced ethernet (Jumbo Packet support)
- More aggressive GPU options, and prices
- PCI-E
- A quad CPU workstation model
I'd rather see a couple of these options offered than a speed bump to 3 GHz. But still, I'd rather create on a Mac than a Windows machine any day.
Originally posted by - J B 7 2 -
- A quad CPU workstation model
I'd
That's really expensive from most manufacturers. Like $2,500 just for the board. It's more than double the price of a dual. I think an Xserve w/4 processors is something that only Pixar, and the rest could afford. But I think a Xserve blade sounds cooler if your going to go that route.
Originally posted by - J B 7 2 -
- Another internal hard drive bay
- Hardware RAID on the mobo for drives 2 & 3 (see above)
- More aggressive GPU options, and prices
I'd settle for any two of these three [assuming that it could boot from the hardware raid drives 1+2].
for these upgrade options in their top of the line workstation.
The technology is already there, so it's more a matter of connectivity
to enable the user to expand as needed.
How difficult can it be for them to make these connectivity options available in a $3000 aluminum box?
We can only speculate how long it will be before Apple configures
a tower with a pair of dual core 970MP or Power 5 processors giving us the "Precious" QuadraMac.
A 3rd HD bay would be nice, but I would rather see better connectivity
support for external storage if space is too much of a premium.
PCI-e is a must, until Apple figures out a way to take advantage of CELL based GPU's.
My personal favorite would be a new audio interface card taking full advantage of 64 bit technology enabling Apple to leave ProTools
and their friggin $7000 HD requirements in the dust.
It's a nice wish list, but how many of these improvements can we realistically expect in the next revision?
Originally posted by onlooker
"There was nothing misleading in that keynote"
There surely was. They compiled things for the PC using GCC instead of their usual intel compiler that they are normally compiled with. Apple said (lied) it was to be fair to the PC, but they new that when compiled with their usual intel compiler the test showed the process to be faster on the intel machine.
As was shown later in independent tests all over the internet.
"SPEC FAQ, SPECfp2000 contains 10 Fortran programs, and 4 C programs. In other words, SPECfp is mostly Fortran, and NAGWare is the Fortran compiler, so therefore it is most likely NAGWare that is the bad compiler for Intel, not GCC.)" from lower article. My bad.
"All the tests they presented during that keynote are 100% correct."
Yes, they were correct, but they were purposely obscured by Apple. That's why it was misleading. I don't forget when someone's honor has fallen until they make a mens to restore the honor they have lost.
This is just one of the various summaries of how many ways apple was misleading in those tests
This was argued endlessly on many boards, but what the heck, I checked out the website you linked to. I stopped reading when I got to his statement that hyperthreading was disabled delibrately to hedge results in PPC's favor. Even Intel admits that for Spec testing, disabling hyperthreading IMPROVES Spec scores. And no, I'm not about to search down the link to Intel's site that mentions this, but I did read it and it was on Intel's website.
As far as hedging tests, they all do it. For the most part comparing Spec test results between different architectures is difficult at best and down right useless at worst. Spec should be used by engineers comparing the results between similar cpu architectures for analysis of design changes. Compilers have WAY TOO much effect on the results to infer which architecture is better or worse, or CPU will perform better in real world apps.
Anyway, that is my humble opinion.
Originally posted by rickag
I stopped reading when I got to his statement that hyperthreading was disabled delibrately to hedge results in PPC's favor.
You shouldn't have stopped reading. Maybe then you would understand the level of deceit used in the tests. That was just one example page, and the first I found on the subject. The author admittedly wasn't aware of what all the tests were, but to be fair reported everything changed in the tests. Also there are results on that page that show the differences in performance after things were change back to normal. I don't think there was one test that the G5 actually came out on top in.
Originally posted by onlooker
...
I don't think there was one test that the G5 actually came out on top in.
Blast? Yes, I know, most people don't run Blast at home. But still one for the home team.
I don't want to get into a rehash of this old topic. Yes, Apple skewed test results. This article blew much of it way out of proportion and a lot of Mac zealots reacted very poorly in responding to this article.
Originally posted by onlooker
LOL.. No kidding. It's no wonder why either. If they don't start making it true to it's name (PowerMac) they probably wont. I don't think I'll be buying one again. Not unless they start making one that can compete with the high end offerings from their PC competitors. I've had three PowerMacs, and I don't think I'll be buying a 4th at this rate.
Most PowerMac users I new bought an AMD PC after the independent G5 speed comparisons came out.
Ok, forgive me if I have to laugh at this.
You say you've had three PowerMacs. Presuming that at least one or two of them were in the G3 or G4 range you're telling me that you were happier with the old 133-167 mhz front side buses than you are with the current 900mhz-1.25ghz ones?
I'm not about to start comparing high end Mac's and PC's because there's too much to factor in for both sides to make a valid comparison that will satisfy everyone, but I will say that the G5 was a quantum leap forward in hardware architecture for the PowerMac compared to earlier versions and I can't quite comprehend how you were happy with old PowerMac performance, but are disappointed by the current G5's.
Onlooker is heavy into 3D and he's right. If you want the fastest 3D then right now PCs are going to win hands down. Apple could ameliorate this situation by actively promoting ATI and Nvidia's higher end cards and helping with driver development. Maybe that'll happen with Tiger..who knows.
My particular needs for a Powermac G5 would be Audio and Video so I'm not in the same boat. The G5 kicks enough butt there to keep me happy.
Originally posted by hmurchison
Ravenpen.
Onlooker is heavy into 3D and he's right. If you want the fastest 3D then right now PCs are going to win hands down. Apple could ameliorate this situation by actively promoting ATI and Nvidia's higher end cards and helping with driver development. Maybe that'll happen with Tiger..who knows.
My particular needs for a Powermac G5 would be Audio and Video so I'm not in the same boat. The G5 kicks enough butt there to keep me happy.
Hmurchison,
I'm not disagreeing that PC's certainly have definite speed advantages over Mac's in certain areas, 3D being one of them.
The part of Onlooker's post that confused me was his satisfaction with the performance of his three previous PowerMacs, I can only assume he was satisfied since he kept buying them, but his disappointment with the current G5's.
As I said in my previous post, PC's and Mac's have certain strengths and weaknesses in performance when compared to one another and there's no one test that will satisfy everyone.
So I guess what I'm saying is that if Onlooker is disappointed with PowerMac performance in comparison to PC's then that is understandable, but if his disappointment is with the performance of current PowerMacs to previous ones, then that I don't understand and from the way his post was worded that certainly seemed to be one of the things he was complaining about.
P.S. I also do Video work and am waiting patiently for the next rev. G5's to upgrade my 867 G4 Quicksilver, though to be honest the 867 still does a damn fine job with FCP and DVDSP.
Originally posted by onlooker
You shouldn't have stopped reading. Maybe then you would understand the level of deceit used in the tests.
Not really.
You assume that gcc for the Mac is as optimized for the hardware as ICC is for x86. And you'd be horribly wrong. All the benchmarks that show "normality" demonstrate is that a wildly well-optimized compiler will beat a poorly optimized compiler. But most applications are compiled with MS' compiler, not Intel's, so it's not necessarily a real-world test to use ICC.
On top of that, at the time Apple ran the tests, gcc was better optimized for x86 than for PPC, by a margin, simply because people had been working on the problem.
Nevertheless, what Apple effectively did was take the compiler out of the equation by using the same one for both platforms. This is standard practice for benchmarking: You compare Premiere on both platforms. You don't compare Premiere to FCP. Why should compilers be any different?
If you want to use ICC for the Intel side, you'd have to use something comparably optimized for the PPC on the Mac side to have a fair comparison.
P.S. I also do Video work and am waiting patiently for the next rev. G5's to upgrade my 867 G4 Quicksilver, though to be honest the 867 still does a damn fine job with FCP and DVDSP.
Sweet! Then you're eagerly waiting for NAB 2005 as I am. I'm most excited to see the first Final Cut Pro version that I will buy. I've used FCP very sparingly since version 2
Oh noooooooooooooooooooooooooooooooooooooooooo
not the infamous Haxial garbage. As Rickag say..that bozo was debunked a long time ago. Amorph is right, the guy ranted about compilers but Apple chose GCC because at least you had the same compiler to test. I doubt GCC was even close to be optimized for PPC as it was for X86 hence the improvement we saw with the IBM XLC/XLF compilers. If high scores are what you value the most then you use a tweaked compiler. If a fair test is what you want you use a compiler that is available on both platforms.
You know we've heard this myth for years about how fast PCs are. That'll last about as long as it takes to build a PC and say "well it doesn't fell 'that' much faster than a fast Mac"
I felt duped myself..I noticed that my PCs did well when I was letting them crunch one task at a time but as soon as I wanted to start tossing in new tasks they degraded pretty fast.
You want to benchmark a computer right..develop a "real world" task where you utilize the computer in a scenario that is likely to happen. Something like compressing video while surfing and burning a CD. The one thing about Macs that people love is that they degrade gracefully under load. Unix baby!
I love the iPods success because I don't have to care about whether someone "gets it". If a PC is so great then go buy one. But you have to wonder why the most exciting thing happening in PCs nowadays is modding them with plexiglass windows adding colors and lights. Everything "but" computer performance. IMO.