Macbidouille.com offers some benchmarks of the PPC 970 platform running MacOSX in 32bit mode, against a P4 at 3 GHz and a dual G4 1.42 GHz. Keep in mind that OSNews can’t confirm the validity of the benchmarks as there is no official word from Apple or IBM (the article also doesn’t give information about the exact hardware configuration used for both platforms).
Mac OS X and PowerPC 970 Benchmarks
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Can anyone estimate when an iMac using this processor will become available?
And do you think it will have problems with heat and cooling, and also noise?
Because it would be nice if it was silent, but how, with this chip?
Is my browser on crack or does this link go nowhere?
It is slashdotted atm. Go there later.
I just found out that this article may be a hoax. It seems that some people on Slashdot.org claim that the site is unreliable, and there are questions concerning how the reviewer(s) were able to obtain the hardware and software to test (i.e. 64-bit versions.) It seems o.k. to me though.
Sorry, must be my connection on the fritz heh!
Thanks Eugenia, it just came up
Slashdotted… I like that hehe
In the article, is it compared to a p4 at 3ghz? If so, where?
Also, if macosX becomes 64bit, how much faster can it be expected to run, if at all?
>In the article, is it compared to a p4 at 3ghz? If so, where?
What do you mean “where”? In the pictures they include.
BTW, they show that their duals only are able to kick the P4 3GHz. However, we don’t have any information on the hardware used for the P4. Depending which mobo/memory they used for the test, P4 can be ‘fast’ or ‘faster’. As for their dual setup of 970, they should have also included a dual Xeon setup for fairness.
As it looks now, it seems that PPC 970 is “as fast”, but not extremely faster than the P4s in single cpu mode. When these machines come out, the new P4s will be out and Apple will still be playing catching up (even if for clock-per-clock Apple machines might be faster) and they will be pricier. And no one put Xeons in the mix yet…
It’s going to be interesting, that’s for sure.
FYI, Macbidouille.com has a mixed track record. It’s been way off with some predictions, but has also been spot-on on some occasions. As usual take it everything with a pinch of salt, but only a small pinch in the case of Macbidouille (imho).
Re MacOS X 64bit. It probably won’t go any faster at all. There’s really no reason it should go faster.
MacOS X 64bit will be faster then the 32bit version…
1) Most file systems use 64bit allocations to allow for partitions larger then 32G. I would give this about 5% faster for hard disk transactions.
2) IP6 and some network protocals also use 64bit numbers. Again only a small increase in speed.
3) If you place a lot of memory in the box (ie. 8G), then you can have many open applications and not see a paging slowdown. (Big speed impovement, but will be mostly seen in large servers; my computer at work has 8G of ram and normally has 1500 jobs running at any 1 time (it’s a POWER3 box)).
4) If Apple redesigns the ABI for the 64bit system, expect about a 10% speed increase. 64bit binaries don’t need to be 32bit compatable; so, Apple can fix some of the preformance problems caused by using a 68000/x86 style API on a PPC (just a bad idea; they’re emulating Instruction Pointer relative addressing on a machine with-out an accessable IP).
Eugenia, the PPC970 test they did was single. That dual is a G4 1.42.
sorry, my bad, i misread, you were right
How do you know that 970 will be more expensive than the top of the line intel processor? Do you have any clue as how much the 970 will actually cost?
If the figures are reliable then its very intresting that a 1.4 Ghz 970 is faster than a 3Ghz Pentium. The 970 will surely increase its Mhz quite fast in the beginning and is said to be available with speeds up to 1.8Ghz at launch.
This year will be good for apple.
The preliminary benchmarks reportedly reflect that the single 1.4GHz PPC 970 is anywhere from 87% to 254% faster than a current Dual 1.42GHz G4 machine depending on the specific test (Photoshop, Final Cut Pro, Alias|Wavefront Maya Render).
Ouch
Looks like a hoax to me… IBM has said that they won’t have SMP-capable 970’s (they will wait for 980s)
Looks made up. I don’t doubt that the PPC 970 will be a great chip, but I *highly* doubt it’ll be out by June. Perhaps by the end of the year.
This is definately looking good for Apple *if* true. Price remains the concern for me. I’ll wait for the Mac world in June. Afterall, we all know that Jobs can’t possibly introduce the music service as part of his keynote (its already been done)!
I get the gut feeling that because Apple seems to believe that their systems are “high end”, they will screw up and sell systems with the PPC 970 at a price that will be out of range for most buyers. I think the PowerMacs are overpriced now and I expect to see the 1st shipment of PPC 970’s cost even more.
I hope Apple surprises me and sells them at the same price as their current line. Maybe even a little less on the PowerMac’s would be really sweet.
Apple releases previews of Software, never of hardware. Until someone gets confirmation of wether this is real or not Im will just regard the article as trash.
Looks like a hoax to me… IBM has said that they won’t have SMP-capable 970’s (they will wait for 980s)
I think you’re confused. IBM hasn’t said anything yet regarding whether or not the PPC970 will have SMP support, and I think all indications point to yes. This processor is aimed at the low-end server market, an area where it would be foolish not to support SMP.
The Power5-derived successor to the PPC970 (which IBM hasn’t yet named “PPC980”, as far as I am aware) will, in all likelyhood, support Symmetric Multithreading (SMT) like its Power5 counterpart.
What I want to know is, will there be a Pegasos board using this chip? Or will there be a half-decent linux distro for it?
Now that’d be cool.
If it’s “obviously a hoax” then it is impossible for someone to confirm it. So why make that statement?
I find it difficult to believe that a 1.4 Ghz PPC, even in SMP configuration, can beat a P$ 3.0.. I seem to recall a similarly specced 64 bit x86 couldn’t…. even with a fully 64bit OS.
If the benchmarks are genuine ( doubt it ) I suspect they were achieved with altivec-enhanced PPC software and poorly-ported windows versions,… the way Apple always do it.
I will wait to see but dont get your hopes up
{ Looks made up. I don’t doubt that the PPC 970 will be a great chip, but I *highly* doubt it’ll be out by June. Perhaps by the end of the year. }
I believe so too. The PPC 970 will undoubtedly be a killer chip, Two things that make me believe they are fake is that Red Hat and SuSE would have released versions of their software for the 970 already. It never fails with IBM. Red Hat and SuSE ALWAYS get first dibs. Its hyperbole from mac fanatics trying to get site hits, run up them counter boys ( & Girls ) give em something to talk about at the end of the month
I find it difficult to believe that a 1.4 Ghz PPC, even in SMP configuration, can beat a P$ 3.0.. I seem to recall a similarly specced 64 bit x86 couldn’t…. even with a fully 64bit OS.
Rememeber, these aren’t just application tests, they’re application tests of applications that have been highly PPC optimized.
Now, I’m not saying this isn’t a hoax… procuring the hardware and software required for these tests alone would be quite a feat.
Just keep in mind what sort of tests were proportedly conducted…
But rather the other architectural improvements IBM has made, while complying with the PowerPC instruction set. That’s what really gives the 970 muscle.
End of year for the 970? That’s the definition of wishful thinking. Apple employees are reportedly getting 30 percent discounts right now on the Powermacs. WWDC has been delayed to allow for release soon. Sure, we don’t know 100 percent, but “end of year” is a bad bet and not likely at this point. sorry, charlie.
You’re basing your opinion on a the performance of a CPU that hasn’t been released yet. As others before me have said, by the time the 970 is released, both AMD and Intel will have faster processors out. Apple’s problem hasn’t always been slower hardware, but timing. Better hardware will surely be released, Moore’s Law, blah blah blah, things just get better obviously that goes without saying. What’s worth noting is the pace at which Apple and their partners get these technologies out and into the hands of customers.
Rather than do something about that pace, Apple instead innovates on other technologies and features that buys them a quick “wow” effect on the crowd, but never anything concrete enough to sway even the non-believers.
You said: “You’re basing your opinion on a the performance of a CPU that hasn’t been released yet. As others before me have said, by the time the 970 is released, both AMD and Intel will have faster processors out.
Hey, the dual 970 (in 1.8 and 2.0 speeds) totally crushed the Pentium 3.0. Like in one test, the 970 duals took 1/3 the time of the Pentium. In another, it was less than half the time. The 970 machines look to be out within a month. Does Intel intend to ramp speed up to 9.0 gigahertz in the next month? Or even 6.0 gigahertz? Any can you run those in dual for any realworld applications?
also you said, “Apple’s problem hasn’t always been slower hardware, but timing. Better hardware will surely be released, Moore’s Law, blah blah blah, things just get better obviously that goes without saying. What’s worth noting is the pace at which Apple and their partners get these technologies out and into the hands of customers.”
We all know Moto screwed Apple. They may even be sued for it. If that’s a meritorious lawsuit (and you don’t know that it is not), then apple may have done nothing wrong, relied upon Moto representations and got screwed. That game is over.
Too bad, mac haters. In less than 4 weeks, you are going to have to come up with something other than “macs are slow.” Better start thinking.
“I seem to recall a similarly specced 64 bit x86 couldn’t…. even with a fully 64bit OS.”
can you clarify before i make assumptions and light you up?
-dually zeon owner
The benchmarks are fakes. There is a long thread at http://www.macrumors.com that explains this (well, some of us explain it…others are in denial). Basically, the problems are these:
(1) Bryce 5 shows a big gain from using dual processors in these benchmarks. But Bryce 5 does not use multiple processors, so that benchmark is clearly fake!
(2) The results for the G4 and Pentium 4 were clearly copied from http://www.barefeats.com (they get the exact same numbers, and the tests that BareFeats runs are NOT standard benchmark suites). So basically, what happened is someone just copied the P4 and G4 results from barefeats.com, and then made up the 970 results to compare them against.
everything thetruth said is right and these benchmarks are 100% FAKE.
mac user are so unsure and scared of how crappily their system performs when compared to x86 that they have to make up benchmarks to say that their processors are faster when they’re really just slower and more expensive.
then they use car analogy to make it seem like they’re getting luxury.
anyway obviously these benchmarks are total crap. ppc970 benchmarks show that 1.8ghz AMD opteron processor is much faster than 1.8ghz PPC970. but 3.4GHz prescott core p4 will be faster than them both.
ppc cannot keep up with x86, end of story. all macheads would like to think otherwise but this is the truth.
> 1) Most file systems use 64bit allocations to allow for partitions larger then 32G. I would give this about 5% faster for hard disk transactions
Why? 99% of the time spent using the hard disk are used by the disk itself. No one will see any speed increase in storage handling except for servers with large amounts of memory used as disk cache.
> 2) IP6 and some network protocals also use 64bit numbers. Again only a small increase in speed.
So small that it can’t be measured. Really.
> 3) If you place a lot of memory in the box (ie. 8G), then you can have many open applications and not see a paging slowdown. (Big speed impovement, but will be mostly seen in large servers; my computer at work has 8G of ram and normally has 1500 jobs running at any 1 time (it’s a POWER3 box)).
Yes this will get faster with a 64-bit OS.
> 4) If Apple redesigns the ABI for the 64bit system, expect about a 10% speed increase. 64bit binaries don’t need to be 32bit compatable; so, Apple can fix some of the preformance problems caused by using a 68000/x86 style API on a PPC (just a bad idea; they’re emulating Instruction Pointer relative addressing on a machine with-out an accessable IP).
Possibly, but the gains are not so big as you think. I would guess _at_most_ a 3% gain as most hotspots do not call other routines when using good compilers.
You have forgotten that a 64-bit OS also must manipulate pointers twice the size of a 32-bit system, this loads the memory subsystem a bit more and will make it a bit slower (DEC once wrote that the Alpha was 5-15% slower when running 64-bit code but with todays caches and memory interfaces I guess 1-2% slowdown is more realistic).
For the record, PPC970 most certainly does support SMP according to its manufacturer, IBM. See http://www-3.ibm.com/chips/techlib/techlib.nsf/techdocs/A1387A29AC1…
“Based on the award winning Power4 design, this processor is an 8-way superscalar design that fully supports Symmetric MultiProcessing.”
Possibly, but the gains are not so big as you think. I would guess _at_most_ a 3% gain as most hotspots do not call other routines when using good compilers.
You seem to be describing some sort of automatic inlining done by the compiler. This won’t occur unless the inline directive is specified (in C99 code).
Furthermore, this hampers not only the execution speed in terms of the activation record format, but also the speed of binding to dynamic libraries as well. The bizarre enforced prebinding system placed into Jaguar could be an attempt to counter the poor ABI design and its effect on application startup times.
The people who originally did analysis of the Mach-O ABI in place on Darwin/OS X suggested that it resulted in a 10-12% performance loss over a properly designed ABI.
For more information, see http://www.unsanity.org/archives/000044.php
Yes, these benchmarks are obvious fakes.
Still, I have no doubt that the 970 is going to be a MUCH faster processor than the current G4. My personal feeling is that the 970 will be a little slower than a top-of-the-line P4 when it is released. Apple will, of course, release certain AltiVec heavy benchmarks claiming that it is 5x the speed of a P4. In the end though, does it really matter. Is a 20% speed difference really that big? Given the Mac’s loyal following, I’d guess that OSX is worth more than 20%. The new processor will revitalize Apple’s stagnant product lines regardless of whether it is fastor or slower than the P4.
64-bit vs 32-bit: I haven’t seen any major ISA improvements with 64-bit POWER vs 32-bit PowerPC (unlike AMD’s extra registers in x86-64) – correct me if I’m wrong. Unless large amounts of memory are used, users will probably see an extremely modest increase (or possibly even a small decrease) in performance.
1) One has to wonder why Apple hasn’t pulled that info from the site yet. They tend to be quite quick with their “cease and desist” letters…
2) MHz does not matter. Feel free to check Intel Centrino benchmarks for some “laptop eats workstation for breakfast” figures (Pentium M 1.6 GHz >= P4 Xeon 2.2 GHz)
3) The success rate for MacBidouille regarding rumors has been quite good lately.
Yes, those benchmarks smell fishy… but do not be too surprised if the PPC 970 happens to be a *good* machine.
Me, I’ll wait some weeks before purchasing my new computer (an Intel Centrino laptop, most probably
just in case Apple finally gets it right…
exactly which market(s) would the 970 address?
Would apple push 64 bit computing into the mainstream yet?
What would make sense, i suppose, is to use 64 bit in the powermacs and servers and use a G4 in the eMAC, and iMAC.
The lesson from sun, SGI and others is that you really need to capture the low end and the high end to be successful. The consumer, general business segments give you volumes and you really need those for a semiconductor efficiencies.
The argument seems to be that because the numbers for the pentium and G4 come from a prior test, then the 970 numbers are not real. but what if somebody just took the prior numbers and then did the same test on the 970?
Sure, the likelihood is that they are fake. but any way you slice and dice it, macs are going to get a whole lot faster soon and a number of PC diehards are unhappy about this.
Never trust Apple zealots and their benchmarks even if the current machine is signifcantly slower than anything else on the market Apple fans will always maintain that it is faster than intel/amd and osx is so much better
You said: “Never trust Apple zealots and their benchmarks even if the current machine is signifcantly slower than anything else on the market Apple fans will always maintain that it is faster than intel/amd . . .”
Actually, we admit all the time that PCs are faster on the desktop. This is obvious. Just like it’s obvious that OS X and the iApps/FCP, etc. are better technically and from a useability/ease of use standpoint than windows and MS or 3d party app equivalents. You only have to spend 30 minutes with each system to realize this.
Windows is about compatibility, speed and cheap cost. Those are its advantages. See, it’s not so hard to admit things.
To all the Mac haters – give me a break. The PPC was clearly faster than the Intel x86 from ~94-2000 or so. Just because Intel pulled ahead doesn’t mean that IBM can’t catch back up, or even surpass Intel. IBM isn’t a small company, and if they feel that Linux/PPC is a great future for them (as opposed to Linux/x86) then they would be fools to not push this architecture to its max possible performance. Its the mid 90s all over, except they are backing an OS that is gaining not falling apart (Linux versus OS2).
The fact is that the Mac market is peripheral to IBM’s real plans for this chip – and thats a good thing in this case. Low end servers, workstations, blades, and Linux boxes galore are where IBM wants to go with this. Make money off the hardware, and the consulting, and tell MS where to go.
Apple may be a cool machine, has cool design and the hardware/software integration is good. Its something I would give my mom to use, but I’d never find an use for it, and so many others who like to code or just hack software. Its also a good machine for artists and people who work with documents. But then again, not something for the mainstream. Me, I use *ONLY* linux. Slackware Linux 9.0 to be exact, on my Athlon 1.8 Ghz/512 mb ram/100gb/ATI Radeon 9000 Pro. How much would I pay for an apple with this configuration that can match up the speed? Allot. I run games, movies, graphics and much more. Not that Slackware is something extraordinary, its just an i686 optimized (and i386 compatible) distribution on top of which things just work because it doesn’t have any patches or modifications. So my point is:
1) Don’t hate Apple, cause the company is good, and their machines are good for some people, even I would like to have one.
2) They are overprized, and the design is really not interesting for many, or that crap with titanium notebooks wich are made of alluminium… titanium is very expansive.
3) Why I don’t use Windows? Cause its buggy… even XP. And its slower in many ways than my Slackware Linux.
4) Back to apple: They have “great” marketing, but repeating things gets just borring. And buying an apple and then being dissapointed (I met people), its not good. Most people who burned themselvs are lyeing to themselvs that apple is faster, when in reality its just a computer that has a little better quality than the rest (well, the other are doing well also), and of course the pride that not every Bob Joe Six Pack has one. Its something like it was with Linux in the begging when the first Bob Joes tried it and where proud that they use Linux, its hard to use, and not everyone uses it, and then the stupid thing: Its not M$, so its better. M$ doesn’t make such bad software, they are just … pigs….
They are overprized, and the design is really not interesting for many, or that crap with titanium notebooks wich are made of alluminium… titanium is very expansive.
The Titanium PowerBook G4 (the 15.2″ model) *was* made of Titanium, not Aluminum. Apple moved to Aluminum for the new models (the 12″ and 17″) and is expected to release a 15.4″ (1.2GHz?) PPC970-based Aluminum PowerBook later this year. The current 15.2″ TiBook doesn’t come with FireWire 800 or integrated Bluetooth, when even the 12″ sports integreated Bluetooth.
For what you get, the Apple laptops are very nice. While PC fanboys are quick to paste the prices of Dells with all the latest bells and whistles, they seem to forget how horribly those Dell laptops are constructed. Part of the price overhead you’re paying for with an Apple laptop is quality construction.
Of course the powerbooks are made of aluminum. That’s what it states on the product page. They’re not titanium (the old ones were though)
“Housed in a lightweight and durable aluminum alloy enclosure, the new 17-inch PowerBook G4…”
“IBM isn’t a small company, and if they feel that Linux/PPC is a great future for them (as opposed to Linux/x86) then they would be fools to not push this architecture to its max possible performance. Its the mid 90s all over, except they are backing an OS that is gaining not falling apart (Linux versus OS2).”
Ibm yearns to displace intel in the processor world. They have for many years. Intel and the pc drones may soon face an onslaught like they never could have been programmed to expect. Things could get ugly here, for the pc crowd that is.
I like hearing about Apple moving to a new processor bla bla bla but I would like to see Apple’s announcement
stating they are using the PPC970 chip.
“Bascule:
expected to release a 15.4″ (1.2GHz?) PPC970-based Aluminum PowerBook later this year.”
Where does it say that? Apple’s website?
If these are real (I am tempted to agree with those who say they aren’t) I can see some excellent systems coming from this. The thing makes me most interested is the thought of building homemade powerhouse PPC systems.
Maybe this will spur some of the GCC guys to start focusing on some PPC optimizations (gcc loves x86 right now, but not much of anything else).
Though I doubt that they actually have a working PPC970 dual processor 1.8 ghz or 2.0 ghz Mac system, it wouldn’t surprise me if one would seriously wipe the G4 and of course the old architecture pentium 4 system. Moto is dead and Intel has been lengthening the pipeline for the P4 for years to eek out greater and greater mgz/ghz numbers. It’s a trade off and no sizeable speed increases have been made. It won’t be catch up any longer, unless you are talking about P4 trying to catch up with PPC 970. I, for one, would like to see Apple put a Power 4 or even Power 5 into a high end or server system for the real speed demons out there. But hey, it can’t hurt to dream.
Where does it say that? Apple’s website?
No, don’t expect Apple to announce anything about the PPC970 until they have saleable PPC970-based products, as it would hinder sales of their current product line.
The statement was prefixed by “expected” because of that. Apple is not releasing any official information regarding the PPC970. Any information you read regarding Apple and PPC970 is going to come off various rumour sites, some of them more reputable than others.
Information regarding the 15.4″ Aluminum PowerBook is coming from MacWhispers. Here is the most recent information:
In June, there are two development tracks we have been watching closely that will have reached a point where an actual product release is possible. More likely, these products will be released in July, following a June announcement. Both the much anticipated new mid-size PowerBook and the new PowerMac towers are on very closely aligned development tracks, since the PowerBook was diverted into what was a major redesign process in late January. We have a number of (admittedly untested) sources in the Apple OEM channel telling us that both products will use the IBM PPC 970 chip, with the PowerMac debuting with both single and dual processor configurations.
The new PowerBook is being built around an all-new 15.4-inch LCD from Chi Mei Optoelectronics. We do not know the pixel resolution, but do know that the panel will have specifications “equal to or better than” the recently introduced 20.1-inch Cinema Display panel, also built by Chi Mei.
Those benchmarks look fake to me, I doubt we will see any real PPC 970 based Macs for quite a while.
But personally I really hope that the PPC 970 blows away anything by Intel and AMD. I probably still wouldn’t buy a Mac, but it would encourage Intel and AMD to make even faster and cheaper CPUs.
Plus it would probably boost Apple’s sales and I think it’s important that they remain a healthy company. Competition is always good and a lot of important innovations have come from Apple. It would be a disaster for the industry if the last real alternative to Wintel (on the desktop) disappeared.
Here’ hoping Amiga Inc and Genesi produce some boards with these processors on them.
Screw Linux/OSX. They wouldn’t be able to compete on the speed front on the same hardware.
The article seems fake to me. They may have even extrapolated the results. What we can count on is a faster bus and faster processors. You can’t really speculate beyond that.
lol @ mdma, did you get your t-shirt yet?
Amiga is what the world needs right now. A MacOSX/Linux/m$ alternative that doesn’t scream useless flashiness/bloatware/and user unfriendliness all at once. IBM cpus with Amiga will harness all the new found power and herald in a new age in computing. Things can only get better from here on.
I look forward to an evergrowing PPC market with Amiga to destabilise the stagnant and innovation starved x86 market.
hey palm, maybe this is an opportunity for you to jump back into the OS game. beos on the 970. People do prefer computers over pdas. Yeah i know it’ll never happen. Palm is going to die their way.
Apple may be a cool machine, has cool design and the hardware/software integration is good. Its something I would give my mom to use, but I’d never find an use for it, and so many others who like to code or just hack software. Its also a good machine for artists and people who work with documents. But then again, not something for the mainstream.
You seem to be misinformed at the least. A lot of UNIX geeks actually like OS X. There is a lot to code and hack there. Also, 99.5% of computer user “work with documents”.
Me, I use *ONLY* linux.
Then, you are not in the mainstream.
2) They are overprized, and the design is really not interesting for many, or that crap with titanium notebooks wich are made of alluminium… titanium is very expansive.
You don’t even know what you are talking about…not to mention your spelling. How old are you? 9?
Here’ hoping Amiga Inc and Genesi produce some boards with these processors on them.
Screw Linux/OSX. They wouldn’t be able to compete on the speed front on the same hardware.
Yes, and Amiga has no professional level applications.
Furthermore, you’re talking about running a vaporware OS on a vaporware processor… and expect to have some idea of how it will compare against new versions of other operating systems running on the same platform?
Amiga is what the world needs right now. A MacOSX/Linux/m$ alternative that doesn’t scream useless flashiness/bloatware/and user unfriendliness all at once.
And you’re comparing the user friendliness of MacOS X to… a vaporware operating system. I can’t imagine from your comments that your experience with OS X is extensive.
Looking over the AmigaOS 4 feature set, I see a lot of “bloat” as well. But “bloat” is a word that zealots tend to use for an extensive set of features, which is certainly not a bad thing, unless you’re on lower-end hardware at which point it might affect application startup time adversely. However we’re talking about next generation 1.8GHz PPC processors here, ones with an 800MHz bus. They certainly don’t qualify as low-end hardware.
IBM cpus with Amiga will harness all the new found power and herald in a new age in computing.
You’re dreaming. For one the main feature of the PPC970 over previous PPC processors will be 64-bit support, and due to the extensive amount of 32-bit PPC assembly present in AmigaOS 4, a 64-bit build of the kernel will be quite impossible for the short term. XNU, written with portability in mind (and thus largely in C) can easily be built on these processors.
Amiga will revolutionize nothing, and will have a userbase a fraction of the size of the current Mac userbase, which is in turn a fraction of the size of the Windows userbase. I’m afraid Amiga will be forever relegated to a sideshow of mainstream computing.
Development of PPC970-based Amiga boards is certainly a long time off, but more likely will never happen.
Typing this on an x86 ThinkPad running Redhat.
Can we expect a ThinkPad running PPC Linux any time soon?
C’mon IBM, you know you want to.
(And a 970 JVM too!)
RJDohnert Apple releases previews of Software, never of hardware. Until someone gets confirmation of wether this is real or not Im will just regard the article as trash.
Last year Apple released/announced the Xserve at WWDC : P
yep IBM make pretty good “stickers” that they shove out for the consumers.
this article shows a more realistic view of the PPC 970
http://macbuyersguide.com/editorials/editorial-ppc970.htm
Conclusion: PPC 970 is fast but by the time its released it will be slower then then the AMD Opteron and Intel processors.
So it will be eating the dust like always of the PC CPU’s.
The funny thing is that PPC CPU’s actually risc CPU’s were meant to run at higher speeds then Intel (CISC) CPU’s as they are a cleaner design so they should have been able to be able to run at a higher clock with not to much fuzz.
I think Apple should support the .NET from MS or do its own. This way applications from now can run on any platform Apple likes with little fuzz porting them.
just my two euros
> yep IBM make pretty good “stickers” that they shove out for the consumers.
I’m content with Linux. Centrino ThinkPads many $AUS100s less than a 12″ powerbook are available. (No bluetooth or airport extreme, but these aren’t yet must-haves for most.)
However if the benchmarks are to be believed, it would make sense for IBM to sell a few with its own CPU. Particularly if they marketed it as an all-IBM mobile J2EE development platform (DB2/Websphere).
Of course it’ll never happen for fear of Apple’s ire over OSX on non-Apple machines.
so just use Linux.
> so just use Linux.
I will. My point being that Linux is a software platform. If the PPC970 blows x86 variants away, wouldn’t it make sense to run it on optimal hardware?
And without subsidising Apple through their typically more expensive hardware? (I happen to believe IBM could produce PPC ThinkPads cheaper than the inflated Australian prices Apple retails here)
I will. My point being that Linux is a software platform. If the PPC970 blows x86 variants away, wouldn’t it make sense to run it on optimal hardware?
PPC970 will not “blow x86 variants away”
It might in application benchmarks of things like Photoshop, but that’s entirely irrelevant in the Linux world as Linux does not support these applications.
If you want to run Linux on a laptop, there is absolutely no reason not to go x86
Some people has thought that bench available in macbidouiile web site, are a fake because the bryce results can not be possible, saying the Bryce is not optimized for multiprocessing. Well i don’t see any problem with those results, the difference between the powerpc970 1.4 ghz and the dual-1.8 are even small. Juste look in the barefeats results and you will see the results for the dual-G4 at 1.0, 1,25 and 1,42 ghz, and you get 30, 24, 21. Well the difference of performance is due to the difference of frequency, and more the diffrence of frequency is large more the gain in performance is high. The diffrence between the dual-1.0 ghz and the dual-1.25 ghz (difference of 250 mhz) is higher than between the dual-1.45 ghz and the dual-1.25 ghz (difference of 192 ghz).
So conclusion ; the difference in the macbidouille bench is normal, the two models have a difference of 400 mhz, so the difference of performance is just due to the difference in the frequency, and this diffrence seems correct regarding the difference of frequency.
I am not saying that those bench are right or not, everyone has to make his own opinion, i just wanted to say that the bryce argument, to say that its a fake, is wrong.
“Conclusion: PPC 970 is fast but by the time its released it will be slower then then the AMD Opteron and Intel processors.
So it will be eating the dust like always of the PC CPU’s. ”
I think you try to convince yourself that the x86 plateform is faster than the powerpc. For my opinion, its just depends on the applications and the level of optimization. The G4 shows very good performance (faster than x86), in video, audio, and science applications. Just see what FinalCutPro4 can do on the G4, do you know any software that can offer the same level of performance on x86. No, there is not my friend.
And i think that the powerpc970 will be really a x86 killer, not only on Apple computers, but also on the IBM hardware.
But in any case, we have to wait and see, …..isn’t it?
The article in “macbuyersguide” (a notorius anti-mac site incidentally) has a number of flaws.
First, it focuses on SPECint and SPECfp tests for single processors (P4 vs 970). But the P4 and realworld wintel apps can’t really run dual, can they? While the 970 can and OS X apps can. Also, right now you don’t get double the performance with G4 dual systems because of the limitations of the bus, etc. But Macbidulle is saying with the new motherboard architecture, we will get 90% increase with the second 970 processor (versus 50 % with the additional G4)
Second, as the article itself conceded, “SPECint (integer) and SPECfp (floating point) tests don’t bear much resemblance to real-world tasks with Altivec-enabled applications, where we expect the PPC970 to shine.”
I think these 970 systems will be announced at WWDC June 23-27, same time that the 64 bit OS X is introduced. But could be earlier.
Much faster Pentiums are not in the cards this year (the article says 3.2 and then later 3.6 second half 2003 – that won’t be fast enought to keep up with dual 970 systems it looks like). And Athlon 64 is not coming till Oct, says the article.
Put dual 970s with Altivec on a good bus and modern m-board, and what will happen? I think it looks like the mac systems will be “blowing away x86 variants,” at least for the time being.
F A K E !
Plain and simple.
“They’re fake, aren’t they? They really must be fake, right? Have to be fake, right everybody?”
Ask yourself this, what if they aren’t?
If I could get a commodity PPC970 (ie, like what you can do with any consumer-level x86 hardware), I wouldn’t mind running Linux on it. The PPC really does have a far nicer architecture than x86, at least in terms of assembly language programming (which you shouldn’t do anymore, but such is the life of a CS student). 32 GP registers, yummy, I like. However some people here need to not confuse “nicer architecture” with “better performing”. A nice, clean RISC chip is still going to get the crap kicked out of it by by an ugly pseudo-CISC chipset running at 10x the clock speed. Unfortunately, clockspeed is not really determined by who has the cleanest chip, but rather who has the most hacks to let it ramp faster.
The benchmarks are definitely faked, that much I believe. My personal impression from what I’ve read is that the G5 is going to run “somewhat” faster than the G4. A lot of it is going to depend on clock speed and how fast it can ramp. However, it is extremely premature to say either way whether a PPC970 is faster or cheaper than an Opteron or Xeon, because _no one here has seen a PPC970 or its pricing scheme_. So, Mac Zealots: shut up about the PPC970, and PC Zealots: shut up about the PPC970.
Thanks!
-Erwos
“Ask yourself this, what if they aren’t?”
Ask yourself this, what if they are?
In the scientific world, we would have peer review. You publish the results of your experiments, and then everyone tries to replicate them. This happens in the computer hardware world, too. GeforceFX is released, everyone+dog goes out and reviews it, and comes up with roughly the same numbers.
The PPC970 cannot be subject to this process since no one has one, except for a very few people. Therefore, there’s no reason to care about these benchmarks, since they cannot be verified. So, unless you can independently verify these benchmarks for us, I see very little reason to argue about them.
So, sure, they might be true, but there’s an equal chance it’s not true. In fact, I’d argue that it’s more likely to _not_ be true because it’s extremely unusual for any CPU to make performance jumps like the one claimed. And we haven’t even taken into account that the reviewer has an obvious bias for the Apple, putting all their results in doubt before we even bother with the scientific process.
Does that make my position any more clear? I grew up on Macs, but Apple’s software prowess does not mean that IBM can make miracles. If it’s true, it’ll be nice, but try being a _little_ more circumspect about the whole thing.
-Erwos
you said, “In fact, I’d argue that it’s more likely to _not_ be true because it’s extremely unusual for any CPU to make performance jumps like the one claimed.”
it seems that one of the biggest reasons for the performance gains will be the bus improvements, so talking about a “CPU jump” exclusively isn’t quite capturing the whole story.
also don’t forget altivec. Again, it’s the bus, stupid. Macbidulle is reporting an 80 percent increase in altivec performance because of the bus and interconnects on the new board.
And If the benchmarks aren’t true, then macs will still be a whole lotta faster than they were. It’s all plus for apple. The only issue is how much of a plus.
Sorry it took me so long to answer but beyond popular belief I have other things to do than to sit here and roam the forums all day. So lets get started:
{The argument seems to be that because the numbers for the pentium and G4 come from a prior test, then the 970 numbers are not real. but what if somebody just took the prior numbers and then did the same test on the 970?
Sure, the likelihood is that they are fake. but any way you slice and dice it, macs are going to get a whole lot faster soon and a number of PC diehards are unhappy about this. }
Who cares about benchmarks. I dont care if a Mac is 10 ghz and the PC is 3.0 ghz, Im buying the PC. First off, I dont have a need for a 64 bit Processor and if I ever do I will go with the AMD Opteron, it is fast enough. Why? because I hate companies that control the hardware and the software, you think integration is nice? I think integration is crap, all it does is lock you into a proprietary platform. If I was a diehard PPC Apple user, then if I got pissed off at Apple, I have nowhere to go. If HP pisses me off I go to Dell, if Dell pisses me off I can go to Gateway. I dont have to put up with crappy service and support and I dont have to worry about waiting for my company to put its foot in its mouth and not be able to sell me the system I want, I dont have to wait for backorders or put up with excuses.
{watching dohnert squirm is real fun}
Who is squirming? The point is i could give a crap lesshow many systems Apple sells, or how fast the systems are. The PC is fine for me. It works, it is fast enough and Im not wasting my money on hardware that I dont need. Im a realist unlike you, I know there is no way Apple will ever knock Intel or AMD out of the top slots. Apples are a niche market, thats it. The only people that are buying Macs are the people that have had them for years, you arent seeing any kind of significant Marketshare increase on the Mac side, the people that switch, what few there are, are people who buy into the propaganda about Apple producing the best OS and the best hardware. I am productive with Linux, and I can be productive with Windows. All I care about is being able to play what few games I have, development tools and accessing my iPod which I can do on Linux BTW. The OS doesnt make the Computer, the hardware doesnt make the computer. The user makes the computer. The perfect example is. Can you run Mac OS X on an unmodified 9600? No, can I run Linux on an unmodified 266? Yes. My girlfriend runs Win XP on a 500 mhz P3 with 196 mb of RAM and it runs beautifuly. Try running Mac OS X on a G3 it is slow as dirt. Apple locks you into its hardware, it sells hardware. That is Apples money maker. If Apple goes with the 970 which I seriously doubt they will Mac users will have to follow a 64 bit upgrade plan way before I will even think about it. Grow up, Apple is not a family, its not a community, Apple doesnt know who you are nor do they care, all they care about is being able to run to the bank and cash the checks you people write them. You want a 64 bit system? Seriously tell me what you will be able to do on it that you cannot do on a 32 bit platform.
End of preach
Have a nice day
Vio con Dios
Never trust MS zealots and their benchmarks even if the current machine is signifcantly slower than anything else on the market MS fans will always maintain that it is faster than Moto/IBM and XP is so much better
For RJDohnert, sorry kids, Windows Xp is very slow on a P3 500 mhz, you can tell its fast to someone who don’t know windows, but to me, you make me laugh,….
Windows Xp is incredibly slow on a 1.3 ghz P3, everything is slow, the interface, the IO, internet explorer, and its unstable.
Have you tried osX on a G3? For sure not, i have a Imac first generation that run 10.2 very well. Try tu run Windows Xp on a 233 mhz machine. You will be amazed how its impossible ti use. I don’t even know if winWp can run on a Pentium 233 mhz. Osx runs perfectly on a G3 500 mhz machine.
It’s all about configuration. You’ll need about 256Megs of ram, but winxp will run quite nicely on a p3 450Mhz.
I don’t know if the benchmarks are fake but are pretty close to expected performance. And you should expect so much FUD from the x86 world to make you sick. But I don’t care about f@<&!#g benchmarks, contrary to all those idiots who spend all day at TomsHardware.
PPC970 will certainly be good but it’s the half part of the equation. You need expensive peripherals too to build a powerful system (how about a SCSI320 RAID and 23″ TFT?). Since I do not run benchmarks, do 3D modeling or play games on my home computer I am about to convert my lame peecee into a monitor-less Linux server, putting it 2 rooms away, and start saving money for a powermac (iMac is also a possibility) to go after with my Cube.
How many minutes have you had the chance to witness XP speed? 5 Minutes? In fact, all Windows (W9x, W2K, XP) are almost identically fast — given enough RAM. When you load some apps in XP, 256 still might not be enough, better to go with 512. Stateing that it crawls on 1,3 GHz is a tell-tale about your insight into the matter…
XP runs perfectly fine on my 3-years-old 900MHz Athlon with 256 MB of RAM. (Even when I bought it, it wasn’t top of the line.)
A friend of mine runs XP on a 300 MHZ Compaq notebook with 128 MB RAM.
I am currently running a 1GHZ Athlon. It has used the following RAM/OS combinations (it uses 16/32 MB) shared video memory:
48 MB/Win98 slow
48 MB/Win2000 very slow
288 MB/Win98 fast
288 MB/Win2000 fast
288 MB/WinXP Pro medium fast
480 MB/Win2000 fast
480 MB/Win2000 fast
480 MB/WinXPPro fast
NT based OSes (Win2000, XP) are very strongly affected by available RAM. In fact CPU usage rarely exceeds 20% on a 1GHz system except for processor intensive tasks (games/multimedia). Buy a cheap CPU 2GHz and spend more on 1 GB RAM and a fast HD.
The article in “macbuyersguide” (a notorius anti-mac site incidentally) has a number of flaws.
An anti-mac Mac-buyers Guide. ROFL
First, it focuses on SPECint and SPECfp tests for single processors (P4 vs 970). But the P4 and realworld wintel apps can’t really run dual, can they?While the 970 can and OS X apps can.
SPECint and SPECfp are good tests for cache, memory and bus speeds (despite their claim to test other things) and is a rather bad system benchmark but can give you an overall indication of a cpu. “Realworld” x86 apps sure as hell can run on SMP machines (given that they are running on an SMP OS) but if the apps have not been written to be multithreaded then you will only see minor improvements in application but huge increases in system responsiveness. Do you really think application vendors only would make the MacOS version of a program multithreaded and not the x86 version? Are you that ignorant?
Also, right now you don’t get double the performance with G4 dual systems because of the limitations of the bus, etc. But Macbidulle is saying with the new motherboard architecture, we will get 90% increase with the second 970 processor (versus 50 % with the additional G4)
You have obviously no knowledge in performance increases associated with dual CPU systems so please refrain from making things up! Look at the newly released Opteron and in database scaling tests you see a 23% increase over a single CPU system and this is for a system designed to scale vell ( http://www.aceshardware.com/read.jsp?id=55000262 )! Most applications thus only see only small benefits by a second CPU unless they have a multithreaded design for the CPU intensive tasks. Most apps do not fit in this category.
Second, as the article itself conceded, “SPECint (integer) and SPECfp (floating point) tests don’t bear much resemblance to real-world tasks with Altivec-enabled applications, where we expect the PPC970 to shine.”
Um, I thought the G4 also had an Altivec unit? You can say the same about the P4 and Athlon64-Opteron and applications that are “SSE2-enabled”.
I think these 970 systems will be announced at WWDC June 23-27, same time that the 64 bit OS X is introduced. But could be earlier.
Expect 64bit MacOS to be announced but not available for quite some time as it takes time to rewrite drivers etc for 64 bit. 64bit OS is not needed at the start anyway so Apple can take it rather cool about 64bit MacOS for the time being, no rush.
Much faster Pentiums are not in the cards this year (the article says 3.2 and then later 3.6 second half 2003 – that won’t be fast enought to keep up with dual 970 systems it looks like). And Athlon 64 is not coming till Oct, says the article.
The P4 successor Prescott is coming this year. Athlon 64 will ramp quickly in speeds and they will have to in order to compete with intel. For both AMD and intel is the 970 an no-factor. Apple simply have a too small marketshare to matter anymore and sadly this wont change anytime soon.
Put dual 970s with Altivec on a good bus and modern m-board, and what will happen? I think it looks like the mac systems will be “blowing away x86 variants,” at least for the time being.
The bus is what will give the 970 a boost but I believe the 970 will not be faster in the clear majority of benchmarks. That wont stop fantastic-supercomputer-Jobs from saying that is blowing the competition away and the believers will preach Amen and sing halleluja because Steve Jobs simply cannot be lying, can he?!
I also always have to point out IBM in all this. The thing is not about Apple and the 970 it is about IBM and IBMs plans for the 970. This is an unknown factor and my guess is that they will use the 970 in servers and cheap enterprise desktops using Linux to get into the emerging Linux-on-the-enterprise-desktop market. This (if true) will not be good for Apple.
I hope the new Apple music store is a success and that this gives Apple enough freedom from hardware sales that they can take the leap and fully become a software company because this is where Apples strenght lies and to have this hampered by proprietary hardware is never good. Why have x86 hardware succeded? Price and choice!
MacBidouille posts regarding concerns about the similarities between the benchmarks and Barefeats’ numbers as well as the concern of Bryce numbers (English Translation by Nonoche):
“We contacted our source about the too big similitudes between the benchs we received and those published by Barefeats, and were told that Apple took the numbers of the P IV and dual 1.42 from Barefeats and used their testing protocol. For Bryce [nonoche:not supporting multiple processors], here’s the explanation we got:
“I should have told you, the benchmarks with Bryce… were done on a beta of a forthcoming version of Bryce, which should be released in July or August this year. Version 6 will support multiple processors… Bryce 6 beta is available on a few P2P sites, notably on hotline. The real file is Corel Bryce 6 beta.sit, and takes 91.2 MB”
Read this: http://slashdot.org/comments.pl?sid=63152&cid=5886775
Jon “Hannibal” Stokes knows what he’s talking about and have done nice comparisons of different CPU architechtures at ArsTechnica.
{ WinXp on a P300, …..what?}
I said Pentium 3 500mhz, I knew font rendering in Safari was a little messed up, but I didnt know it was that messed up.
{ For RJDohnert, sorry kids, Windows Xp is very slow on a P3 500 mhz, you can tell its fast to someone who don’t know windows, but to me, you make me laugh,….
Windows Xp is incredibly slow on a 1.3 ghz P3, everything is slow, the interface, the IO, internet explorer, and its unstable.
Have you tried osX on a G3? For sure not, i have a Imac first generation that run 10.2 very well. Try tu run Windows Xp on a 233 mhz machine. You will be amazed how its impossible ti use. I don’t even know if winWp can run on a Pentium 233 mhz. Osx runs perfectly on a G3 500 mhz machine.}
Oh man dont make me laugh, I know Windows better than probably Bill Gates himself and it runs fine on a 500 mhz machine. I had Mac OS X 10.2 installed on a Snow iMac and the thing was so slow I wanted to throw it out the Window. Please save the Mac Zealotry for someone who doesnt know any better.
Never trust Apple zealots and their benchmarks even if the current machine is signifcantly slower than anything else on the market Apple fans will always maintain that it is faster than Intel/AMD and Mac OS X is so much better
You said: “Realworld” x86 apps sure as hell can run on SMP machines (given that they are running on an SMP OS) but if the apps have not been written to be multithreaded then you will only see minor improvements in application but huge increases in system responsiveness. Do you really think application vendors only would make the MacOS version of a program multithreaded and not the x86 version? Are you that ignorant?”
All I know is that I don’t see or hear people talking about dual P4 systems. Some dual Athlons are out there. But overall, dual processors just don’t seem to be a viable thing right now on the desktop x86 systems and desktop wintel apps.
You said: “Look at the newly released Opteron and in database scaling tests you see a 23% increase over a single CPU system and this is for a system designed to scale vell ( <http://www.aceshardware.com/read.jsp?id=55000262> )! Most applications thus only see only small benefits by a second CPU unless they have a multithreaded design for the CPU intensive tasks. Most apps do not fit in this category.”
What Opteron can do does not necessarily equal what these new 970 powermacs can do. This is what Macbidulle is saying, not me. They are saying that with the bus improvements and other system-type improvements, the new mac 970 systems are getting 90 percent increase in performance out of the second process. Maybe this is true, maybe it’s false. We’ll know very soon.
You said: “Um, I thought the G4 also had an Altivec unit? You can say the same about the P4 and Athlon64-Opteron and applications that are “SSE2-enabled”.
Yes, the G4 has altivec, but it’s bus-starved. Macbidulle is reporting that Altivec performance is jumping 80 percent on the new bus. Also SSE2 does not seem to me to be taken advantage of to the same extent as Altivec on the Mac. The G4 are so bad that the mac platform really had to suck every last ounce of juice out of Altivec just to stay remotely credible compared to x86. The same imperative obviously has not existed on the x86 side.
You said: “The P4 successor Prescott is coming this year. Athlon 64 will ramp quickly in speeds and they will have to in order to compete with intel. For both AMD and intel is the 970 an no-factor. Apple simply have a too small marketshare to matter anymore and sadly this wont change anytime soon.”
Could you give me some more specifics? Like speed and month of release. Because the 970 macs are being introduced on June 23-27, or possibly earlier.
You said, “The thing is not about Apple and the 970 it is about IBM and IBMs plans for the 970. This is an unknown factor and my guess is that they will use the 970 in servers and cheap enterprise desktops using Linux to get into the emerging Linux-on-the-enterprise-desktop market. This (if true) will not be good for Apple.”
As I have explained before, I do not see how the emergence of Linux hurts apple. It helps apple because: (1) the perception will be that there’s multiple valid choices in a platform, not just one (windows); (2) linux promotes open standards that apple can adopt, so the “incompatibility” issue always presented by windows will become less and less of a factor; (3) linux apps can be modified slightly to run on OS X. Sure, at some point in time Linux could threaten apple in its markets (consumer, creative pro, edu). But that is a long way off (except edu, that could happen).
You said, “I hope the new Apple music store is a success and that this gives Apple enough freedom from hardware sales that they can take the leap and fully become a software company because this is where Apples strenght lies and to have this hampered by proprietary hardware is never good. Why have x86 hardware succeded? Price and choice!”
Don’t you see that making hardware and software is part of what allows apple to excel and create new and interesting stuff that takes BOTH hardware and software? Look at Microsoft and these dull hardware conferences that they have – another one is coming up. Is anybody looking forward to it? No. Look at those tank-sized and expensive “slate” PCs that Billie boy came up with. Yuck. Price and choice is great, but it comes at a cost. You act like having open hardware is a cost-free, total upside situation. Few things in life are that way, and this is not one of them. Look who has succeeded in that environment, Dell. Is there any hardware innovation there? Any innovation there?
It’s good to have variety and different models in any ecosystem, including the PC one. x86 brings a lot to the table – but so does a company like apple that makes hardware and software.
You said, “I dont care if a Mac is 10 ghz and the PC is 3.0 ghz, Im buying the PC. . . .Why? because I hate companies that control the hardware and the software, you think integration is nice? I think integration is crap, all it does is lock you into a proprietary platform.”
I think this is the essence of this issue. What PC diehards find so infuriating or disconcerting (to the point many of them just bury their head in the sand) is that
You said, “Im a realist unlike you, I know there is no way Apple will ever knock Intel or AMD out of the top slots. Apples are a niche market, thats it.”
Are you saying there’s no way that Apple’s dual 970 systems will be faster when released in the next 60 days than the fastest, then available, desktop x86 systems? Because you can’t possibly know that. We don’t know yet. A reasonable person has to admit it’s at least possible. Me, it looks like that’s where we are headed, but nobody knows yet.
You said, “I am productive with Linux, and I can be productive with Windows. All I care about is being able to play what few games I have, development tools and accessing my iPod which I can do on Linux BTW. The OS doesnt make the Computer, the hardware doesnt make the computer. The user makes the computer.
But what if you are not a very good “user”. You know, not technically knowledgeable like you. That’s 90 percent of the population. Even for technical people, why should things be unnecessarily difficult like they are in windows (and still for Linux although people are working on that). When things aren’t as user friendly as possible, it’s just bad design.
You said, “Apple locks you into its hardware, it sells hardware. That is Apples money maker.”
Apple sells packages of hardware and software. Sure, you “buy” a computer. That’s what you lug home. But inside is the software. The software is really the computer. People overrate the importance of hardware. The software is what we interact with. The hardware is just the fuel. The software is the machine.
You said, “If Apple goes with the 970 which I seriously doubt . . .”
This is pretty dumb. The likelihood of Apple using the 970 is about 98 percent.
You said, “Grow up, Apple is not a family, its not a community, Apple doesnt know who you are nor do they care, all they care about is being able to run to the bank and cash the checks you people write them.”
I think Apple is a company that actually has to work to stay alive. So far I judge they deserve my money. Will they make it, I don’t know. I do question whether enough people care about the quality and the innovation. But it will have been a fun ride in any event.
There is only one thing that matters to me. And that is that the PPC 970 be able to run OS X and its applications really fast. With the Mac, arguments about benchmarks compared to Intel, etc. are useless because the Mac doesn’t run on those processors. The only benchmarks I’m interested in are those that truly show the differences between the PPC 970 and previous PPC’s.
You said, “I dont care if a Mac is 10 ghz and the PC is 3.0 ghz, Im buying the PC. . . .Why? because I hate companies that control the hardware and the software, you think integration is nice? I think integration is crap, all it does is lock you into a proprietary platform.”
I think this is the essence of this issue. What PC diehards find so infuriating or disconcerting (to the point many of them just bury their head in the sand) is that they really are advantages to making both, you can innovate and implement new things in ways that companies making just hardware or software simply can’t. But then you are “locked in” So the PC people react by just denying the advantages, pretending there’s no downside at all to the open hardware model. It’s just not the case, but some people just can’t accept it.
All I know is that I don’t see or hear people talking about dual P4 systems. Some dual Athlons are out there. But overall, dual processors just don’t seem to be a viable thing right now on the desktop x86 systems and desktop wintel apps
Perhaps that is because it’s not needed as a single CPU does the job fine. The second CPU in the PowerMacs helps Apples marketing department more than it is useful for Mac users.
What Opteron can do does not necessarily equal what these new 970 powermacs can do. This is what Macbidulle is saying, not me. They are saying that with the bus improvements and other system-type improvements, the new mac 970 systems are getting 90 percent increase in performance out of the second process. Maybe this is true, maybe it’s false. We’ll know very soon
No the Opteron is no 970 (thank god). But the inprovments in respect to the bus is similar and in the case of the memory system is more advanced than the 970. Remember that IBMs engineer have in the past been about brute force and not grace. Motorola designs with grace (the AltiVec unit is a work of art) but that did not help against the competition.
Yes, the G4 has altivec, but it’s bus-starved. Macbidulle is reporting that Altivec performance is jumping 80 percent on the new bus. Also SSE2 does not seem to me to be taken advantage of to the same extent as Altivec on the Mac. The G4 are so bad that the mac platform really had to suck every last ounce of juice out of Altivec just to stay remotely credible compared to x86. The same imperative obviously has not existed on the x86 side.
Um this is one of those things that indicates that the “Benchmarks” are false. My guess is that they just took the current system bus speed and found that if you slow down the bus of the 970 (900) by 80% you get the speed of the G4 (166)!
As I have explained before, I do not see how the emergence of Linux hurts apple. It helps apple because: (1) the perception will be that there’s multiple valid choices in a platform, not just one (windows); (2) linux promotes open standards that apple can adopt, so the “incompatibility” issue always presented by windows will become less and less of a factor; (3) linux apps can be modified slightly to run on OS X. Sure, at some point in time Linux could threaten apple in its markets (consumer, creative pro, edu). But that is a long way off (except edu, that could happen).
Linux is about cheap and disliking MS. About no 1, this is where Apple will never win because those looking elsewhere from MS will not want to go to a new proprietary (even more so than MS as Apple have complete control over hardware) platform. Linux will effectively be the “only” other choice (sad but true). Remember that Apple is still dependent upon MS apps.
About no 2, incompatibility have not been a factor that have stopped Apple in the past as MS have ported the #1 important app, Office, to the Mac. Windows Media is also unimportant as MP3 (sadly, as Ogg rules
) the most used media format.
About no 3, If you want to run Linux apps why walk the extra mile (and double the cash) when Linux is there. As the OpenSource alternatives becomes viable alternatives (and they in many cases are already) the need for proprietary systems will decrease rapidly. Just because Linux isn’t quite there yet for Average-Joe doesn’t mean it never will be. Everything points at Linux becoming more user friendly (hint to Mr Torwalds: “leave SysV go BSD”
)
There are also many projects currently underway in the creative/pro space that soon will bear fruit. None of this bodes well for Apple in the long run unless Apple wants to stay as the BO of computers.
appleforever: Sure, the likelihood is that they are fake. but any way you slice and dice it, macs are going to get a whole lot faster soon and a number of PC diehards are unhappy about this.
Sure, new Macs would be faster than old Macs. *big suprise*. But would it be faster than PCs? Come on, even if this benchmark is real, which is highly doubtful, what makes you think that Intel (as well and especially AMD) would sit still until these processors go mass produce and actually somehow find their way on super-expensive highest-end PowerMacs?
Unlikely. If Intel doesn’t reach 5GHz, they are probably close by then. And Athlon 64 is coming out around the same time too.
PC diehards not happy? Puh-leze. So what if Macs become faster than PCs? I don’t buy SGIs or Sun workstations – even though they are faster. Price is an important factor, because for an SGI, I could get a lab full of PCs, maybe cluster them up and get the performance I hardly need.
Tony: The PPC was clearly faster than the Intel x86 from ~94-2000 or so. Just because Intel pulled ahead doesn’t mean that IBM can’t catch back up, or even surpass Intel.
Does IBM have the capability to overshoot Intel and maintain their lead? Sure they do. Do they have the money? Sure they do. Do they have a reason? Sure they DON’T. Come on, Intel sells way more processors than IBM or Motorola for PPCs – they have more will to overtake and maintain their lead. IBM is limited to a small market – mainly Macs. Is the amount of Macs sold justify having a large r&D bill?
Tony: IBM isn’t a small company, and if they feel that Linux/PPC is a great future for them (as opposed to Linux/x86) then they would be fools to not push this architecture to its max possible performance.
IBM using Linux is unlikely anytime soon. Especially when they have beef up, at least in Malaysia, advertisements of their PCs, and trying to regain back their brand status for PCs. Linux itself isn’t all that viable in the forseable future, especially when it isn’t for a small niche market (e.g. SGI) or not on x86.
In other words, it would be a loooonnnnggggg time till IBM take any interest regarding Linux as a desktop.
Tony: Its the mid 90s all over, except they are backing an OS that is gaining not falling apart (Linux versus OS2).
OS/2 was gaining a lot. IIRC, it was the best selling software for one time. IBM f*cked it up by bitching with the OEMs. The difference with now is that IBM can’t piss the OEMs off because of the OS…. but processors is an entirely different story altogether.
Tony: and tell MS where to go.
Perhaps they should focus more on their more immediate enemies like BEA and EMC, before picking on yet another one.
Marc (crazy): 3) Why I don’t use Windows? Cause its buggy… even XP. And its slower in many ways than my Slackware Linux.
And Mac OS X is soooooooooooooooooo much more stable and faster?
Little Grrr: Amiga is what the world needs right now.
One where there is a modern *stable* release out maybe?
OS 4 was suppose to be released like, I dunno, 2 years ago?
That software/hardware integration is a major competitive advantage for Apple. See the following articles:
http://seattletimes.nwsource.com/html/businesstechnology/134689749_…
http://www.theinquirer.net/?article=9334
The only way Apple benefits from linux is being able to run more apps.
Chreo, tell me when Photoshop, Illustrator, Flash, and Office are on the linux platform and then you can ramble about how linux will take over someday. Oh, and show me marketing material that tells average-joe that linux even exists. Oh there isn’t any? That’s what I thought.
Linux to me, and I’m sure to a lot is a hobbyist OS. It will never be mainstream because it is too dis-organized. It is an OCD person’s nightmare. If you want linux to be mainstream you have to adopt new ideas, and drop a lot of the original ideas.
Until then, people are going to buy machine’s they see on TV. Apple and Windows PC’s. And that’s fine cause I don’t want my mother calling me with “what’s this dependency thing?, What’s this root thing?, what’s XFree86? What’s a kernel?”
Linux needed a re-thinking and Apple did that. Anything anyone does to make linux user-friendly and mainstream will be copying Apple’s ideas. Just look at that article recently about what linux needs to do. How many times was OS X mentioned?
me: All I know is that I don’t see or hear people talking about dual P4 systems. Some dual Athlons are out there. But overall, dual processors just don’t seem to be a viable thing right now on the desktop x86 systems and desktop wintel apps
chreo: Perhaps that is because it’s not needed as a single CPU does the job fine. The second CPU in the PowerMacs helps Apples marketing department more than it is useful for Mac users.
me: OK, so we have admitted that for whatever reasons, the P4 is for all practical purposes a single processor system. That means it’s going to have to go up against dual 970 systems. That’s the reality pal.
me: What Opteron can do does not necessarily equal what these new 970 powermacs can do. This is what Macbidulle is saying, not me. They are saying that with the bus improvements and other system-type improvements, the new mac 970 systems are getting 90 percent increase in performance out of the second process. Maybe this is true, maybe it’s false. We’ll know very soon
chreao: No the Opteron is no 970 (thank god). But the inprovments in respect to the bus is similar and in the case of the memory system is more advanced than the 970. Remember that IBMs engineer have in the past been about brute force and not grace. Motorola designs with grace (the AltiVec unit is a work of art) but that did not help against the competition.
me: Are you saying that apple can’t possibly have designed a motherboard that can take advantage of the dual cpu better than opteron? These systems are not even out yet, so we don’t really know. Although we do have possibly true benchmarks that back up Macbidulle’s claim of 90 percent increase with the second processor.
me: Yes, the G4 has altivec, but it’s bus-starved. Macbidulle is reporting that Altivec performance is jumping 80 percent on the new bus. Also SSE2 does not seem to me to be taken advantage of to the same extent as Altivec on the Mac. The G4 are so bad that the mac platform really had to suck every last ounce of juice out of Altivec just to stay remotely credible compared to x86. The same imperative obviously has not existed on the x86 side.
chreo: Um this is one of those things that indicates that the “Benchmarks” are false. My guess is that they just took the current system bus speed and found that if you slow down the bus of the 970 (900) by 80% you get the speed of the G4 (166)!
me: it could be false, it could be true. you can’t prove they are false, can you?
me: As I have explained before, I do not see how the emergence of Linux hurts apple. It helps apple because: (1) the perception will be that there’s multiple valid choices in a platform, not just one (windows); (2) linux promotes open standards that apple can adopt, so the “incompatibility” issue always presented by windows will become less and less of a factor; (3) linux apps can be modified slightly to run on OS X. Sure, at some point in time Linux could threaten apple in its markets (consumer, creative pro, edu). But that is a long way off (except edu, that could happen).
chreo: “About no 1, this is where Apple will never win because those looking elsewhere from MS will not want to go to a new proprietary (even more so than MS as Apple have complete control over hardware) platform. Linux will effectively be the “only” other choice (sad but true). Remember that Apple is still dependent upon MS apps.”
me: I think there’s lots of people who don’t have the hangup with a proprietary system that you do. I don’t. My mom doesn’t. This issue is not a big animator for everyone. Plus, many people can actually see the advantages to one company making the whole machine. For one thing, as most people can understand, it’s easier to have one source to go to for tech support. As far as dependency on MS apps, this is pretty much down to Office. If linux gets big, open office document standars will probably take over.
chreo: “About no 2, incompatibility have not been a factor that have stopped Apple in the past as MS have ported the #1 important app, Office, to the Mac. Windows Media is also unimportant as MP3 (sadly, as Ogg rules
) the most used media format.”
me: It’s a matter of perception. Most people don’t even know that Office exists for the Mac and/or that the document formats are 100 percent the same.
chreo: About no 3, If you want to run Linux apps why walk the extra mile (and double the cash) when Linux is there. As the OpenSource alternatives becomes viable alternatives (and they in many cases are already) the need for proprietary systems will decrease rapidly. Just because Linux isn’t quite there yet for Average-Joe doesn’t mean it never will be. Everything points at Linux becoming more user friendly (hint to Mr Torwalds: “leave SysV go BSD”
)
me: I hope Linux does become as easy to use as Apple. But I don’t think this will happen soon. And it won’t happen until people bundle like apple – provide a package of hardware, apps and online service that all work together seamlessly without requiring a “construction kit” approach from the user.
chreo: There are also many projects currently underway in the creative/pro space that soon will bear fruit. None of this bodes well for Apple in the long run unless Apple wants to stay as the BO of computers.
me: This all seems a way off but who knows. As I have said before, I have no problem with a world where linux is as good as the mac and apple doesn’t exist anymore. Then, in effect, I will have the quality apple experience without the downside of closed hardware. I just don’t think it’s going to be so easy to pull this off and quick.
{ I think this is the essence of this issue. What PC diehards find so infuriating or disconcerting (to the point many of them just bury their head in the sand) is that they really are advantages to making both, you can innovate and implement new things in ways that companies making just hardware or software simply can’t. But then you are “locked in” So the PC people react by just denying the advantages, pretending there’s no downside at all to the open hardware model. It’s just not the case, but some people just can’t accept it. }
The only advantage is to the computer company, They own you. There are no significant technical advantages to integrating hardware and software.
{ Are you saying there’s no way that Apple’s dual 970 systems will be faster when released in the next 60 days than the fastest, then available, desktop x86 systems? Because you can’t possibly know that. We don’t know yet. A reasonable person has to admit it’s at least possible. Me, it looks like that’s where we are headed, but nobody knows yet. }
See thats the thing, I dont think Apple will release a 970 in 60 days. It is a very untested platform and I think you will be very disappointed when they dont. It is possible that the chip will be faster when it is released but will it be faster than the Opteron, I doubt it. Even if it is faster guess what, it wont be that way for long. Technology changes so much in a year the 970 will be obsolete. It takes me 20 hrs per week just to keep up with what everybody is doing. Apple, Microsoft, Sun and some Linux companies. Everything changes everyday. Everybody has built up the hype and hope Apple will use the 970 that I think disappointment will be running rampant.
{ Apple sells packages of hardware and software. Sure, you “buy” a computer. That’s what you lug home. But inside is the software. The software is really the computer. People overrate the importance of hardware. The software is what we interact with. The hardware is just the fuel. The software is the machine. }
So what you are saying is that if I buy a Mac, it will turn me into a professional graphic artist? I will be able to pump out DTP projects that will be 100% better than everybody else?
{ This is pretty dumb. The likelihood of Apple using the 970 is about 98 percent. }
I put it at 65% but neither of us is Apple and as previously stated, the rumors about Apple using the 970 is just that rumors no one knows for sure and Apple isnt saying. It might be possible they were able to dish out a 4 ghz G4 or the long awaited G5. Rumor sites started the 970 thread, Apple didnt and neither did IBM. I am looking forward to WWDC tho just to see what Apples dishing out, hell I have to go.
{ I think Apple is a company that actually has to work to stay alive. So far I judge they deserve my money. Will they make it, I don’t know. I do question whether enough people care about the quality and the innovation. But it will have been a fun ride in any event. }
Oh Apple does have to work very hard and I do not knock them for that. People do care about quality and there are many people who believe the PC companies make very good, quality equipment.
As much as I enjoy the smooth integration of Apple’s hardware and software, I cannot justify the price differential between their hardware and x86 hardware. Regardless of whether they get PPC970 processors or not, I can build a machine for 1/3 the price, and have more machine at the end of the day, if I use x86 hardware. I recently configured an AMD machine for doing 3-d visualizations (it’s for a small company that can’t afford a 4 or 8 way system)
AMD AthlonMP 2400+ (x2)
Asus A7M266 board (i think thats the correct number)
2 GB PC2100 DDR RAM
2 x 200 GB Western Digital Caviar drives (8 Mb read cache)
128 Mb GeForce 4 Ti 4200 video
Sound Blaster Live 5.1
WinXP Pro
18″ Viewsonic LCD
and did that for $2197 and some change.
When I went to Apple and configured one of their dual 1.42 GHz G4 machines, the end price was over $6200 dollars…
I like OS X just fine, but I can’t justify that price for it.
Apple has not announced 970 (or AMD) based systems.
AMD64 (Opteron) is out there now. Via and nVidia already have Opteron chipsets ready. Suse, Redhat and Mandrake have Opteron optimised distros available now. Win XP Pro and Server have been announced for Opteron.
Keith, what does it matter when you can’t run OS X. I think I can count 5,000 posts exactly like yours. Blah blah I can build a better machine for 1/3 the price. Well that’s great, I can too. But none of these machines will ever run OS X and that’s why I have two macs and one PC. OS X is worth whatever money you’re saving by buying a PC at 1/3 the price.
Please don’t bother posting stuff like this…everyone’s read it 5,001 times.
me: { I think this is the essence of this issue. What PC diehards find so infuriating or disconcerting (to the point many of them just bury their head in the sand) is that they really are advantages to making both, you can innovate and implement new things in ways that companies making just hardware or software simply can’t. But then you are “locked in” So the PC people react by just denying the advantages, pretending there’s no downside at all to the open hardware model. It’s just not the case, but some people just can’t accept it. }
dohnert: The only advantage is to the computer company, They own you. There are no significant technical advantages to integrating hardware and software.
me: The advantages are more practical than technical. Take consumer video editing. It was just a whole lot easier for apple to put in firewire hardware, put in the OS hooks for this and create iMovie than for a group of wintel companies to do this. Anytime multiple companies are involved, there’s negotiations over who pays for what, who gets the revenues from this, etc, etc. The result is that it took three more years at least (1999 for apple versus say 2002) for consumer video editing to be tolerable on the PC platform. Still, it’s inferior to iMovie. And would consumer video editing on the PC even existed without firewire? Probably, but it would have taken longer.
As far as “owning” you, there is a major constraint on apple. They have to appeal to new users/switchers with half-way decent/tolerable prices. And since they can’t charge existing users more than new users/switchers, apple isn’t really in a position to take advantage all that much. Sure, they can charge some additional increment. But part of that goes to R&D. Apple has not been making obscene profits so there is not much empirical evidence to support the claim that apple has been able (practically speaking) to take advantage of users.
Of course none of this pressure on apple would exist without x86. As I said, x86 brings a lot to the table in the computing “ecosystem”. It’s best for all these different models to interact and compete.
dohnert: See thats the thing, I dont think Apple will release a 970 in 60 days. It is a very untested platform and I think you will be very disappointed when they dont. It is possible that the chip will be faster when it is released but will it be faster than the Opteron, I doubt it. Even if it is faster guess what, it wont be that way for long.
me: Lots of signs point to release very soon – the one-month WWDC delay, reports of motherboards coming out of Taiwan, IBM production on track. This is speculation, but it’s not unsupported. Untested? We all know apple could have samples of the 970 and you only need some of those to work up a motherboard and prototype and do the testing. This is no big deal. Plus, testing on the mac platform is quicker because of the standardized hardware.
dohnert: “People do care about quality and there are many people who believe the PC companies make very good, quality equipment.”
me: The main problem is microsoft. When you will get the business no matter how good or bad your product is, when there is no effective competitor, why should you make a better product. They have no incentive to do so. Innovation and monopolies just tend not to go together.