BareFeats has compiled the results of a comparative test utilizing Cinebenc, the best cross-platform benchmarking application the market has to offer.It is interesting to see how the MP 2GHz Opteron and Intel Xeon 3.06 GHz MT compared against the MP 2GHz G5. Here is also a previous benchmark session at Barefeats which was updated recently.
Cinebench Benchmarks: Intel, AMD Vs G5 2GHz MP
Submitted by Kelly McNeill 2003-10-17 Benchmarks 62 Comments
All that matters to me is that the G5 is indeed faster than the G4. I’m sure others will find this benchmark quite interesting in their decision on what new PC to purchase.
Showing all angles of the spectrum. You see an MP next to a SP next to a hyperthreading…
which then gives weight to what Apple did to tip the scales in their tests, turn off the hyperthreading.
But, overall, the G5 did weigh in very solidly indeed. While not the top-of-the-class, it was none-too-shabby against the MP Athlon.
i didn’t believe it for a second. there is no way for anyone to catch up in technology with intel, the world’s #1 computer chip maker, not even ibm. looks like g5 performs exact same as optoron tho, meaning mac needs more development work to take #2 slot away from amd
how these benchmarks will turn out when being run on panther.
” there is no way for anyone to catch up in technology with intel, the world’s #1 computer chip maker, not even ibm.”
Assuming you were quoting yourself from the past and saying that (in retrospect you were wrong…correct?
“looks like g5 performs exact same as optoron tho”
For the most part, i think its safe to saw its a tie… even though the G5 squeeked out a few extra performance points.
“meaning mac needs more development work to take #2 slot away from amd”
Are you speaking in terms of performace or market share?
I ask only because the G5 has outsold the Opteron by a large margin thus far… (if you’re speaking in terms of AMD as a whole… you’re correct, but why does it matter… if the G5 is the same if not slightly faster?)
The Xeon 3.06 GHz MT benchmark shows that if you spend enough money on PC hardware, you’ll eventually beat the Dual G5 2 Gig. Especially with overclocking, expensive heat exchangers, etc.
I’d be interested to know if any of the Wintel’s or AMD’s tested are in the dollar range of the Dual G5. In most tests where the G5 is shown to lose, it’s usually to a PC that would cost well over the price of the PowerMac.
I’m impressed at how well the two processors are running 32bit apps, but I wonder what sort of improvements we’ll see when they switch over to a 64bit OS and are the benchmarks have been converted to 64bit apps.
P4/Xeon are very much more mature than Power5 and Opteron, so it’s why I am waiting for a year or so to buy me a few new computers. IBM has spoken to us that the Power5 will have 3Ghz and etc, but I am not sure about Opteron’s planning thought. Also, I am waiting for PCI-Express to get mature.
i was not wrong! i never ever thought that g5 would beat pc, and here is benchmarks to prove it doesn’t! why are u say i am wrong?
the benchmark says optoron beat g5 by 1 point, not other way around. g5 is slightly SLOWER. by #2 slot i meant performance, as mac users claim that g5 would be majer performance boost.
u say g5 is cheaper than comparable systems, that is so not true! for the cost of a single 1.6 ghz g5 system ($2000) u can buy 4 complete 3.2GHz P4 servers from dell, or at least u could awhile ago. plz see this url:
This is kinda suspicious. For one thing there’s no mention about the specs of the units as tested which can make a HUGE difference. Two, why is there no single processor Opteron (AMD FX51?) to show us how the Opteron scales. We see that with the G5 but its scaling is rather…odd. The P4 going dual (Xeon) notices under a 2x gain in performance…while the G5 realizes a better than 2x gain. That…doesn’t happen in the real world. Something seems really off with the scaling of the benchmark.
“u say g5 is cheaper than comparable systems, that is so not true! for the cost of a single 1.6 ghz g5 system ($2000) u can buy 4 complete 3.2GHz P4 servers from dell, or at least u could awhile ago.”
Your argument is useless because you don’t factor in the additional hardware and software that the G5 gives.
If you’re going to compare against P4s… I could give an equally irrelivant price comparison and say your P4s are significantly more expensive than Apple’s eMac of which I could probably buy 2 for every one of your P4s.
Ehhh…Apple is marketing the G5 as the world’s fastest personal computer if you want the fastest computer in the world then the G5 SuperCluster at VT might be up their somewhere.
So you have the G5 up against workstation class processors and hardware. Not bad at all. Anyone who has worked with a Xeon knows how quiet they are and Apple stands a better chance at surviving than AMD.
AMD had been in the red for several quarters. Fast processors, cheap prices and cheapskate customers don’t seem to help them so far.
If you correlate the G5’s and Opteron’s performance from Cinebench and Photoshop from the PCWorld article they seem similiar in performance.
A configured Dell dual Xeon 2.66 is $3500.
Why would you make these dumb comparisons?
The Dell you are talking about is actually $2400. You have to build a complete system dude. You can’t have it come with no operating system and as little memory as possible.
u say g5 is cheaper than comparable systems, that is so not true! for the cost of a single 1.6 ghz g5 system ($2000) u can buy 4 complete 3.2GHz P4 servers from dell, or at least u could awhile ago.
you call that a server?! it’s a bloody mid level desktop. a real server, probably closer to the actual machines tested with, costs $7672 with no OS, nor any other extras… just 2 3.2Ghz Xeons, 2Gb ram, and 2x146Gb in raid 0 to compare to the dual g5, that’s just over $4000. granted, it’s not the $2000 quoted for the 1.6, but it’s more comparable. =D
Apple could have said the G5 was the fastest toaster… it doesn’t matter, because it isn’t a toaster nor a desktop!
Any dual system, with that specs (and price!!) is considered a workstation or server, no matter what apple says.
Given that, the G5 (the dual 2GHz of course, that’s the one everyone talks about) must be compared to the top offerings from AMD and Intel in workstation market, and these are dual Xeons and Dual Opterons/Athlon 64 (I haven’t completly figured out AMD names on this one).
So please, lets all stop calling the dual G5 a desktop or personal computer. Because it isn’t!
I am not making excuses for Apple but its much more realistic to find a G5 dual in someone’s home than a Xeon, Opteron or BOXX system.
You can’t even buy Xeon or Opteron systems at CompUSA!
Plus the G5 is even marketed towards home buyers. You’ve seen the stupid commercial.
Apple pushes it and labels it as a personal computer, you view it as workstation. Thats fine. Its a fast personal computer and a slow workstation. That should make everyone happy but no one ever said it was a workstation.
These are not all the bench marks on this page, because there are more in graphic app’s like photoshop and after effects that show the G5 well ahead Intel and AMD.
ANALYSIS OF CPU INTENSIVE TESTS(Barefeats)
Apple has made signficant strides with the G5 to catch up to the Windows PCs in performance. The Dual G5/2.0 was faster than any of the Windows PCs we tested in 4 out of 5 tests. It was a close second to the Dual Xeon/2.4 in that one test.
When we tested the faster G4 Power Mac 4 months ago, it only was fastest in one test. http://barefeats.com/pentium4.html
If Apple stays on target, which it looks like IBM will, Apple will have one of the fastest or the fastest system out for sometime in the added speed to the G5 and Mac.
Oh, how humourous your posts are; btw, if your sarcasm detectors broken, I was being sarcastic.
Apple to PC converts moved to PC’s because there was a HUGE difference in speed between PC’s and Macs. The G5 closes this gap by a huge margin and many of these PC -> Mac converts are converting back, not because it the fastest but because there is no longer the HUGE performance difference.
As for the G5, I tried it out at the local computer store, with the latest patch applied. There is definately a WOW! factor.
As for future development, IMHO, if Apple see that their 1.6Ghz, low end G5s aren’t being bought in large volume, they might decide to bring down the 1.8/2.0ghz pricing and push the 1.6Ghz G5 into the consumer desktop range. If that is the case then their iMac/eMac range will be pushed back into being a competitive alternative to Wintel.
….PC -> Mac converts are converting back, not because it the fastest but because there is no longer the HUGE performance difference.
PC <- Mac converts are converting back, not because it the fastest but because there is no longer the HUGE performance difference.
Why are the g5s bundled together in the middle of the spectrum? because they are in order of single processor clock speed? huh? Thats a really trivial dimension to be ordering by.
Someone posts to an article saying a 3.2 Ghz server costs 500, but links to a site showing us a 2.8GHz for 500. Anyway, that is very barebones. Nothing on it. Windows XP Pro alone will set you back 200. about 100 I would imagine through OEM. DVD writer. A bit more. A good graphics card. More still. Heck, I could spec it right up there, or even higher pricewise than the Mac.
Funny how someone always finds something to complain about. I assumed it was ordered in two sections, PCs in descending clock speed, and then Apples in descending clock speeds. Would you have preferred random ordering, so it would look more confusing?
Who said anything about offending you? I thought it was funny. Hence the mention of the word “humor” in my post. I know- so many words in it, it probably slipped by. 😉
To Meynard. I was replying to Alex Mouton.
I think it would be best if it simply didn’t show the status of the moderation (pending and already reviewed). It could tally up the votes, and move the most voted for “abuses” to the top of the review list. And after it is reviewed, future reports could just be silently ignored (if it wasn’t moderated down of course).
Do ya ever wonder if they post articles like this one just to get everyone fighting? An article like this is guaranteed to get the thread bloody. There’s zealots of all kinds hanging out in this forum, and serious OS comparison conversations are rare anyway, most of them descending into sniping and name calling, mixed in with some occasional modding and flaming by the Mods themselves. For a n00b, there is much to be learned in some of these conversations, but most of the time it’s just like throwing rats, cats and dogs in the same pen and wondering what will happen. I’m not sure if it’s for the sake of learning about computers or if it’s for entertainment value that I keep coming back. Before I came here, I’d have never thought techies could be so religious about their OSes…
1) The results aren’t really surprising. 3D rendering tends to be a strong-suit of the P4. 3D rendering contains the type of code that really lets the SSE2 unit shine, and doesn’t hit the branch penalty’s of the P4 too badly. As a result, the 50% clock speed advantage kicks in.
2) The G5 is pretty competitively priced, though it could be better. If you spec out an Alienware content creation (pro-level) system and a G5 to similar specs (dual 2.8 GHz P4 on the Alienware, dual 2.0 GHz G5 on the Mac, 1GB RAM, 200GB HDD on the Alienware, 160GB HDD on the Mac) you end up with a price of $3827 on the Alienware and a price of $3550 on the Mac. Taking into account the $450 price premium for the pro-level FireGL X-2 in the Alienware over the Radeon 9800 in the G5, you end up with the Alienware being about $150 cheaper, not too much for such expensive machines. The Alienware has better specs, and a better warrenty, but OS X might be worth a few hundred dollars to some people.
3) The Alienware example isn’t the best one for many PC users. I have a feeling that many of the PC users on this board wouldn’t buy such an overpriced system. They’d build the machine themselves (and save several hundred in the process), ditch useless amenities like GigE (who has GigE networks at home?), and come in close to a thousand dollars under the price of the G5. This is where the “Apple is expensive” comments come from.
4) Panther would make no difference in these benchmarks. 3D rendering is entirely CPU-bound, and performance should be almost the same for pretty much any OS.
that the beta of Cinebench unsed in this test is not build with a G5 optimizing compiler. It think it would be better to wait three months until these compilers are out and the whole test is running on Panther (which has much better math-libs than Jaguar) at that time – then the can make a real test.
Right now they doing the same Apple did with their spec-tests – they used a compiler that is not best optimized for one patform in the test (gcc on x86).
So – wait an see…
Yes, this is just the kind of article to get everyone riled up like monkeys on speed. But, I think the discussion has been fairly tame.. Perhaps a growing sign of maturity? Or maybe the trolls are all just asleep…
>>ditch useless amenities like GigE (who has GigE networks at home?)
I do! 🙂 I do alot of audio work and tossing around monster files on a switched gigabit network at home is nice. Giga NICs are cheap now and my switch was less then you might think. I got my CAT 6 from a local telecom place for 11 cents per foot.
Concerning these benchmarks, CooCoo is right- Cine is heavily P4 optimized, else in pure FP the Athlon/Opteron would own its ass. I wonder what difference a G5 optimization would make? It might tip the scales…but benchmarks aside, I’m extremely happy with my G5 – and thats all that really matters, right? Oh, and in performance I’d think the bandwidth of the G5 is perhaps more significant then the raw CPU speed alone. The G5 *is* workstation classed, it has balanced i/o and huge busses – in Mathmatica I hear its faster then unix workstations and far less money (so says Theo Grey, co-founder of Wolfram). In scientific apps like BLAST the G5, and the Opteron I’m sure, own the P4.
Oh, and in performance I’d think the bandwidth of the G5 is perhaps more significant then the raw CPU speed alone.
This might be the case, but the G5’s I/O architecture isn’t *that* special. Marketing BS on Apple’s webpage aside, you’ve got the following salient points: (note, the Opteron chipset is the nForce3 and the P4 chipset is the Canterwood)
1) ~7.1 GB/sec (read the ArsTechnica article on the PPC-970 to see why it isn’t 8GB/sec) of memory bandwidth. Slightly faster than the 6.4GB/sec of the Opteron and P4.
2) Point-to-Point CPU bus: This is essentially a 2nd gen FSB setup. The P4 is 1st gen, in that it has a shared frontside bus. So two processors share the same bandwidth. This isn’t as big a deal as you’d think at first, because both share the same memory bandwidth anyway. This only helps if (for example) one CPU is accessing the AGP bus while the other is accessing the memory bus at the same time. The Opteron sits in a Hypertransport mesh, in a pseudo-NUMA architecture. This qualifies as 3rd-gen. Each CPU has a hypertransport connection to a chunk of local memory, along with a hypertransport connection to up to two other CPUs. This results in very fast CPU -> CPU communication, and more importantly, allows *each* CPU to get a full 6.4 GB/sec of memory bandwidth.
3) Point-to-Point I/O links: This one is a bit of a dark-horse, because Apple doesn’t say *which* links are PtP. They just say there are PtP links for “major subsystems.” In practice, this is a draw. The Opteron has seperate PtP links for CPU, memory, AGP, and GigE and a very high bandwidth shared link for everything else. The P4 has the same, except a shared CPU link. It would be very difficult to saturate the shared link with just hard-drive, USB, Firewire, and and PCI traffic, given that in the nForce3, it is a Hypertransport link running at 3.6 GB/sec.
4) PCI-X bus: This one would be useful for high-end workstations, but not most desktops. PCI has a max throughput of 133 MB/sec, while PCI-X has a max throughput of 1064 MB/sec. On most machines, the only thing you have on the PCI bus (these days) is your audio-card. Everything else (Firewire, ATA, etc) is inside the chipset. So unless you have high-end SCSI card with a RAID array that can transfer more than 100MB/sec, or need to add a second GigE card, or need to use some super-high-end PCI-X add-in card, the feature is largely useless. My guess is that its in there to support an upcoming XServe.
In other words as a new benchmark pops up almost every day and in each of them while many are even conducted by mac users x86 comes out on top we can chalk the G5 “fastest desktop of all time” up as FUD in the same category as the Megahertz myth. I wonder why Mac users keep falling for that BS.
CINEBENCH 2003 is the free benchmarking tool for Windows and Mac OS based on the powerful 3D software CINEMA 4D R8. The tool is set to deliver accurate benchmarks by testing not only a computer’s raw processing speed but also all other areas that affect system performance such as OpenGL, multithreading, multiprocessors and Intel’s new HT Technology.
Considering the Opteron and the G5 has a single core compaired to the HT Xeons. Essentially a 4 way Xeon box against dual Opteron and G5 boxes. I think both the G5 and the Opteron did amazingly well p4 optimized or not.
Rayner, are you STILL claiming that Apple only offers 90 warrenty?
I posted on here a few days back the actual text under the apple care chooser. it said it had a 90 phone support with 1 year warrenty on parts and laybor. you can buy a 3 year in home warrenty on parts and laybor for 260 bucks or something close to that.
same goes for Dell, 1 year warrenty, but you can upgrade to 3 year with in home visits for about the smae price (mabye 10 dollors diffrent)
Apple has an industry standard warrenty on their systems, so unless the Alienware PC offers something like 2 year warrenty included when you buy the computer, I Don’t see how you can say it has a better warrenty.
“In other words as a new benchmark pops up almost every day and in each of them while many are even conducted by mac users x86 comes out on top we can chalk the G5 “fastest desktop of all time” up as FUD in the same category as the Megahertz myth. I wonder why Mac users keep falling for that BS.”
Do some research before you spread lies. The G5 dual 2GHZ is faster in some areas and slower in others unless you want to think of Cinebench as the ultimate benchmark. In this particular test it lost out to workstation class systems and beat P4. In a PCWorld article it beat all the Althon64 systems and performed as well as an Opteron in applications that it had optimizations for. No one is falling for BS your just spreading lies.
macster: don’t get your panties in a twist, although you seemingly did and they cut of the oxygen supply to your brain.
You accuse me of lying? Now lets see:
The G5 dual 2GHZ is faster in some areas and slower in others unless you want to think of Cinebench as the ultimate benchmark
Now, where did I lie? I never claimed otherwise! but what I said is that apple did lie about the “fastest desktop of all time” just as they did with the “first 64bit desktop of all time” statement. The G5 isn’t as blazing faster than anything else as Apple hyped it to be.
In this particular test it lost out to workstation class systems and beat P4.
So the G5 isn’t even a worstation class system? how pitiful.
In a PCWorld article it beat all the Althon64 systems and performed as well as an Opteron in applications that it had optimizations for.
link please. apart from that, some benchmarks the G5 wins, many it does lose. result: it is not the fastest computer ever. rinse and repeat. Thus, Apple spreads FUD.
No one is falling for BS your just spreading lies.
“You’re” not “your”. Apart from that, I still wait for you to point to any lies spread by me. All I see is false advertisement (BS) by Apple.
Thank you for your attention.
what a joke.
a real server’s scsi card alone costs $500.
the chassis usually runs a couple hundred.
wtf am i even bother replying, i build, buy and take care of servers. it’s my job.
some joker posts a dead dell deal, (every thread about apple pricing, he posts the same link) and that’s proof of the standard cost of a dual zeon server.
we need moderation or a way to plonk dumb $h*ts like that.
a nice dual zeon, dual opteron or g5 is going to costs thousands.
Hal, there is this thing called “MARKETING” that ALL companies do. When Intel stated that the P4 “makes the internet faster” no one said anything. When MS says they have released that XP is the best version of Windows they have made some would argue that. ALL companies hype their product to be the best weather its vitamins, sport supplements, cars, BMW(the ultimate driving machine), Mercedes(engineered like no other car in the world). Its called marketing. Everyone is going to hype their product in face of the competition.
University of Texas hypes their Dell computer cluster at 3Tflops as, “This remarkable resource will place The University of Texas at Austin among the top institutions worldwide in high-performance computing power.”
Meanwhilee VT deploys a G5 supercluster capable of 17.6Tflops while costing $30 million dollars less and may rank right behind The Earth Simulator.
Is UT lying? No but they did candy coat it a little like a lot of companies. Apple is no different. Intel and AMD have both done it too. Apple would be insane to call their computer, “the slowest desktop in the world!” What kind of marketing idiot would do that?
As far as Apple being a workstation or worstation as you like to call it, its not marketed as one and never has been. The G5 is a personal computer that you will find in many homes and businesses unlike the Opteron or Xeon sitting on a desktop. We’ve already deployed G5s in the company that I work at, the Xeons are in the server room not on a desktop.
G5 dusts Athlon64 in Photoshop
The Premiere results and MS Word results have been argued in slashdot as irrelevent. Look how close it comes to the Opteron. They didn’t even test the G5 with Panther which may or may not have made a difference.
G5 and Xeon
G5 VS Xeon, Athlon64 on PS7bench
barefeats has tons on Mac versus PC benchmarks. You really should do some research on this topic instead of having everything hand fed to you. You got the facts delivered to you, draw your on conclusions. If the G5 beating Xeons and Athlons64s on some tests is too slow for you then I hope you find something that works for you.
“what a joke.
a real server’s scsi card alone costs $500.”
But, but, but…the Xserve does not have any SCSI. At all. Are you trying to insinuate that it is not a real server? Prepare to be flamed! Ducks for cover
Oh gee, the clusters? you do know that the dell one is older and the price includes a crapload of other stuff in addition to merely the machines, right? If not, welcome to having fallen for a mac zealot lie… or do you call that marketing too?
I like how you fling around other FUD instead of backing up your claims at all that I’m lying. In fact, you are the one who was lying there. Now, you recount what I said myself: that the G5 is fast, but in fact does not at all beat the other players and leave them in the dust as was claimed. That claim I call a lie.
Handfeeding me the facts? haha, I am the one who stated them as they are and you called me a liar over them. Looks as if you are the hypocrite here. As for not testing with panther and whining about it: they also didn’t test on WinXP 64 (that was sarcasm).
Now repeat after me: Apples marketing is a lie, the G5 is not “the worlds fastest personal computer” neither was it “the worlds first 64bit personal computer”. You call it marketing, I call it lies and false advertisement.
Thanks for playing.
” Oh gee, the clusters? you do know that the dell one is older and the price includes a crapload of other stuff in addition to merely the machines, right? If not, welcome to having fallen for a mac zealot lie… or do you call that marketing too?”
Read the press release:
AUSTIN, Texas–(BUSINESS WIRE)–Oct. 3, 2003– The University of Texas (UT) at Austin, the nation’s largest university, today introduced a Dell high-performance computing cluster (HPCC) at its Texas Advanced Computing Center (TACC), to help enable critical research by scientists and engineers across disciplines.
October 3rd of this year is old? That was 15 days ago so its not old, this is a NEW cluster that is deployed at UT. What is this “crapload of stuff” you are talking about?
Apple is assisting VT and the MacOSX clustering software for the G5s was written at VT.
“Handfeeding me the facts?”
I provided links for you and all to see…where are yours? No one is whining about Panther because like I said the results may or may not be faster. That goes either way.
“You call it marketing, I call it lies and false advertisement.”
I guess we are going to see a lot of class action lawsuit then? Right?
Hal, you still haven’t provided a decent and non-inflammatory rebuttal to any of the scores on the links that I put up?
How can Apple possibly market the G5 as fast if it beats Xeons and Athlon64s on apps important to Mac users like Photoshop? How can Apple do this?
hals probably running an underpowered, underrammed pc.
Thats the impression I get from the windows/itunes feedback here and on the web.
Lots of old anemic rigs out there.
And they All bash macs speed.
Hal, there is this thing called “MARKETING” that ALL companies do.
All companies do marketing. Apple lies.
When Intel stated that the P4 “makes the internet faster” no one said anything.
That’s not a lie. For certain things (VRML, etc) that nobody uses, its not a lie. Its a gross exaggeration, but not a lie.
When MS says they have released that XP is the best version of Windows they have made some would argue that.
“Best” is subjective. For the home user, XP is the best version of Windows ever made, because its the first NT-based home OS. If MS said that “XP is the fastest OS ever made” that would be a lie.
ALL companies hype their product to be the best weather its vitamins, sport supplements, cars, BMW(the ultimate driving machine), Mercedes(engineered like no other car in the world).
Again, all of those are subjective. Mercedes isn’t saying that their chasis is the stiffest of any car, which would be a lie.
As for the University of Texas, note they said that the cluster would place them among the top supercomputers. They will clearly be *among* the top supercomputers with their cluster. They didn’t say that they would be third (like the VT cluster might be) which would be a lie.
Apple is saying that the G5 is the fastest personal computer ever. That’s clearly not the case. Speed is something objective. You can measure speed. The measurements show that the G5 is *not* the fastest PC. Thus, its a lie. Its the same as when Apple said that the G3 was twice as fast as the fastest Pentium II.
I don’t know Joe, dual G5 2GHZ is too slow for hal. He says so anyway. I’m sure he is running Quad Xeon or an overclocked dual Opteron.
Also note: the world’s first 64-bit PC was arguably the DEC Alpha based PCs. Also, the $1000 Sun SPARC systems also probably qualify, because of their price point.
you are overly sensitive to things that Apple says, but call anything x86 and over exajuration, or stateing a reletive term.
please. you are excusing your favorite platform companies for the exact same crap that you bash Apple about.
if speeding up VRML counts as an over exaguration of “speeds up the internet”
then a G5 kicking the crap out of anything on the x86 market when it comes to Photoshop is an over exajeration of “the fasted personal computer inthe world”
macster (IP: —.dc.dc.cox.net) dont be so excited about 17 teraflops there are theoretical. Now cool down and think why IBM fastest eSeries are Opteron based. G5 never beat Athlon 64 FX. Read again.
or it might have something to do with IBM building servers and want a solution for EVERYONE.
Apple marketing their computers as fastest is nothing to get upset over. The thing is it bothers people who will never even consider a Mac so I find that strange.
Microsoft says XP is secure but no one calls them out on it but if Apple chooses to advertise their computers as fast people get upset! Boo Hoo!
Maybe this will make people happy when Apple comes out with a dual G5 3GHZ
“PowerMac G5 Dual 3GHZ the fastest computer on some stuff but not everything and defintely not the fastest computer ever made”
That should sell a lot of G5. We should make hal and Rayiner head of marketing at Apple. Then they can tell the truth about the G5s and how horrible they really are!
regarding the clusters everyone talks about:
about no one crtitisizing intel, microsoft, etc for their claims, I recommend you to read osnews more frequently. All the lies and exagerations were totally trashed everytime.
P.S.: sorry about the english.
macster (IP: —.res.east.verizon.net)
“Microsoft says XP is secure but no one calls them out on it but if Apple chooses to advertise their computers as fast people get upset! Boo Hoo!”
Can you point me to the web page with adequate citation?
debman (IP: —.cable.mindspring.com) – Posted on 2003-10-18 23:07:37
or it might have something to do with IBM building servers and want a solution for EVERYONE.
processor architecture has nothing to do with system avaiability for everyone. AIX or Solaris servers work just fine in the mixed environement. No, eSeries with Opteron are very fast, holding world record in TCP-H benchmarks in the 100GB and 300GB database size category.(http://www.tpc.org/tpch/results/tpch_perf_results.asp)
that stilldoes not mean that IBM would not want a diverse product line for diverse needs and consumers opinions.
the eSeries is CHEAPER than the power5 systems. the eSeries is their internet product for small business. why would they need a power4/power5 based solution?
“Can you point me to the web page with adequate citation? ”
mp, thanks for clarification on the VT and UT cluster. I still hope the G5 puts in a good showing went it is benched.
From the PCWorld scores the Athlon64 beats the G5 in Premiere which is not even optimized for G5 unlike PS which the G5 beats all of the Athlon64 systems. The use of Premiere as a bench has been discussed and invalidated on /. as is the use of a Microsoft product as a performance benchmark on a Mac system. These apps were NEVER made to run fast on a Mac. Its no wonder that Premiere is no longer being made for Mac. Anyone who does NLE on a Mac uses FCP not because Premiere is good but because its performance just sucks.
What PCWorld basically did was pick two bad programs to bench on the Mac versus AthlonXP. Why not use Avid or Mathematica, Maya or Cinebench? Apps that are important to Mac users? A strange anomaly was that the G5 performed similiar/faster than the Opteron in Quake3.
if speeding up VRML counts as an over exaguration of “speeds up the internet”
Apple using Photoshop to say that the G5 is the fastest personal computer in the world is like Intel using the faster VRML to say that the P4 is the fastest CPU in the world. Saying that something speeds up certain apps is not an absolute. I think the commercial is as ridiculous as the next person, but you can’t really call it a lie because its so vague. Saying that the G5 is the fastest PC is a pretty direct and unambiguous statement, and if that isn’t clearly the case, its a lie.
I also think that Microsoft “we’re secure” thing is bullshit. Its a lie. However, you always hear MS say “the most secure Windows yet,” rather than making an absolute statement.
I know when Linux gets more popular, there’ll be stupid commercials about it too, and I’ll think those are bullshit as well. Its something we have to put up with because marketing people are allowed to roam free on the streets. But at least most companies try to make ambiguous bullshit. Apple just flat-out lies.
macster (IP: —.dc.dc.cox.net) see there is difference between optimization for G5 and Athlon 64. Athlon will gain much more than G5 from optimized software because as of now all apps are bearing x86 legacy (16-bit). And this is definitely not efficient.
The pages you provided say that XP is the most secure windows up to date. MS does not compare windows security to any other OS. So Microsoft claim can be true (although I dont see any defference between XP and 2000 plus they added more stuff into OS making it in fact less secure than w2k).
Opteron is not designed for games. Better is Athlon64. Opteron can be used as workstation while Athlon64 was designed with home users in mind.
Debman, for server Power4 is indefinitely better than G5. G5 is to Power4 the same way as Athlon64 is to opteron. Both G5 and Athlon64 are targeting home users. It does not mean that these CPU are slow.
Now eSeries are used to build third biggest cluster on the world in Japan. The Grid will use Globus toolkit.
The expensive is Power4 the cheap is G5, opteron, Athlon64 or Xeon.
Hal certainly isn’t the sharpest tool in the shed, is he? He offers no facts, just opinions that are supposed to carry the weight of a minor deity.
I’m absolutely convinced that Apple may actually have understated the facts when describing the G5 as the world’s fastest computer. After all, the tests were run using a Panther beta, and software that hadn’t been thoroughly optimized for the G5. Here are my facts.
According to Mathematica’s president, the G5 is more than twice as fast as a P-IV (I know, he didn’t talk about Xeons or AMD). Luxology is developing high-end 3D software, and couldn’t even build a demo that made the G5 break a sweat. BLAST software runs more than 5 times as fast on a Mac. Running the P-7 Bench, Apple wins again convincingly. When a Virginia University wants to build a Supercomputer, they make a cluster of 1100 G5s, approaching a theoretical performance of 17.6 Tflops (for comparison; a cluster of 1150 Xeons (2.6 Ghz) comes in around 11 TFlops). An Adobe After Effects test shows that the G5 is 2.5 times faster than a 3 GHz P-IV and almost twice as fast at dual Xeon (2.4 GHz). Acrobat runs about 1.5 times as fast on a G5. And so on…
The only tests I’ve seen from the PC world (before now) that suggest a PC workstation may be faster have been lame and frankly bogus tests like Microsoft Word scrolling (I mean, come on!) and Premiere rendering (which is so old, it couldn’t be done in OS X, it had to be done in Classic!) Oh yes, and gaming benchmarks that rely heavily on Direct X.
So Cinebench, puts out a benchmark beta, and the G5 doesn’t come out on top. It’s a beta, for crying out loud. Even Cinebench has said that they expect to get dual-Xeon sized numbers once the software is optimized for the G5; this is just a first step. So even this one victory may be short-lived.
By any measure, the G5 is a world-class computer. It’s incredibly fast, and with Panther just days away, will become faster still. AMD will probably give it a run for its money, and I really hope they do well, but over the next year or so, Intel is going be out in the cold.
You can compare Word XP to v.X. Look at the font rendering for example: For sceen fonts, Word v.X uses Quartz font rendering with anti aliasing. True Type fonts look similar to what they would if printed. Clearly more processor cycles has been used to achieve this. The Word XP just uses simple anti aliasing and screen fonts don’t look like what they should if printed. We should see a performance difference in Panther as the font engine has been overhaul. Mind you that Word v.X did is not optimised for mac.
read what I said again.
I said POWER4 and POWER5, NOT G5
the e-Series uses the Opteron. the origional poster claimed that the reason the Opteron is used is because it is more powerful than the G5, and I was making a point (that apparently was not clear enough) that they would not use a G5 is a server system, and the fact that they are not using the Opteron in their z-series just points to how IBM actually sees the Opteron, as a small business server chip and nothing more.
I agree on the premiere posting, that it wasn’t a good mac product. Espectially Adobe Premiere 6.5, it just sucked in the way of preformance on Mac OS compared to PS, Final Cut etc. And that is Adobe’s faught for making a bad OS X port. No wonder people swithced to Final Cut Pro Adobe……….it’ts because of Premiere
“see there is difference between optimization for G5 and Athlon 64. Athlon will gain much more than G5 from optimized software because as of now all apps are bearing x86 legacy (16-bit). And this is definitely not efficient.”
This is tough to say because of how well Athlon64 already performs. Will companies like Adobe make specific optimizations to Photoshop just for Athlon64 especially if it performs SSE intensive operations better than the P4 already?
Will Adobe or other software publishers make Athlon64 specific software when their software already runs fast when coded for X86?
This is tough to say. It would benefit AMD. I will tell you that Adobe and other software makers will tune Photoshop and other apps to fly on the G5. Apple has that marketing clout and an established creative user base that will demand it and buy the software. The real G5 Photoshop optimizations will be in Photoshop 8 for Mac.
“The pages you provided say that XP is the most secure windows up to date. MS does not compare windows security to any other OS. So Microsoft claim can be true (although I dont see any defference between XP and 2000 plus they added more stuff into OS making it in fact less secure than w2k).”
This is true that MS does not compare their OS to anyone else but they do say that it is the most secure Windows OS ever. This is not true. A lot of WindowsXP vulnerabilities also affect subsequent version of Windows like 2K and NT. Some go even as far back as 95 and 98. So I would say that the current version of Windows (XP) is no more secure than any other in light of the continued security issues that plague Windows.
“Opteron is not designed for games. Better is Athlon64. Opteron can be used as workstation while Athlon64 was designed with home users in mind.”
True. The Opteron was not that much slower than the G5 and was actually hobbled by its 128MB video card. The Athlon64 is better for games but pricey. I will stick to upgrading my AthlonXP rig for gaming.
There is no Power5 yet. For server opteron is better than G5. There is nothing mysterious about it. I dont remember arguing that someone is or will be using G5 for server. I stated that Power4 is better than G5 (so agreeing with you). I dont understand what you trying to contest. Beside, that was the discussion about Opteron/Athlon64/G5. So putting here present (power4) or will be (Power5) system makes no sense at all. I was pointing to TPC results to prove that at specific price range IMB chose opteron not G5 and that there are reasons for it.