Slashdot Mirror
Apple's G5 Speeds Challenged
An anonymous reader was the first of a seemingly infinite stream of people to submit a URL to an argument that makes the case that the G5 isn't quite what Apple wants you to think of it. The evidence? Apple's own press material. Worth a read.
They're giving us a desktop UNIX running on 64-bit hardware, what else can you ask for? sheesh
Ultimately, it will be interesting to see the real world performance of the G5. I own a 2100+ amd athlon, but I don't feel much of a speed increase from my old 1 Ghz. As usual, a processor is only one part of a computer's performance, and the 1Ghz bus that the G5 will use will greatly contribute to the percieved speed of the system. Also, the interaction with OS X will be important. I use a G4 powerbook running jaguar, but occassionally there are slow downs - not sure why.
"When interpreting these FP results, let us keep in mind that most people use Integer (not FP) most of the time. Therefore, integer results (SPECint) are much more important than floating-point results (SPECfp). In other words, most people should ignore floating-point results because they do not use floating-point anyway (or not much)."
Well that may be true in Windows, but from what I recall the Cocoa API and PDF display model relies on floating point exclusively for screen coordinates. IOW, floating point may not be so important on Windows but it is on the Mac.
I also recall the implementors of Mathematica complaining about the integer centric nature of the Windows API.
I watched the video. (http://stream.apple.akadns.net/ - requires QuickTime). Now, I'm sure there's many ways you could tweak the benchmarks and so forth but the Photoshop and Mathematica benchmarks rocked. The G5 was 2x faster than the Xeon.
I used to get involved doing benchmarking back in the good old days of Whetstone when I worked on supercomputers. Every manufacturer had a different nasty tweak to the compilers that were pulled out only when it was time to do benchmarks for a customer. The mantra then as now was: the best benchmark is the app you want to run (since most buyers of supercomputers write their own apps, porting them for a benchmark was a possibility).
The G5's may not be the hottest thing on the planet but they're close enough to get Apple back in the ball game. Nice systems architecture, nice case and the claim is they're quiet as well. Oh, and don't forget you can put in 8GB of RAM. Now even OS X doesn't need to swap :-)
The 1GHz backplane is the real news. No processor benchmark test really takes into account the total real speed of the system when running applications.
The fast backplane will speed up IO, which is a common bottleneck. 1GHz for a PC backplane is huge. The only machine I had seen a 1GHz backplane in so far is a HP-UX server. It cost wayyy more than $2000 or even $3000.
I really believe that with this new chip alliance with IBM Apple will finally be able to put that "the OS is really cool, but PCs are always faster" stuff behind them.
Yesterday was a good day for apple.
It seems to me that if somebody wanted to use an inferior product, the first thing they'd do is develop a thick skin and at a minimum ignore the criticism being lobbed at their platform of choice. That, or choose to adopt something that seems to work better for the majority so that they don't have to feel left out all the time; obviously when you get to the point of chewing out people who are trying to show you why your choice is flawed it's become a popularity contest for you already (competing, not computing).
Try not. Do or do not, there is no try.
-- Dr. Spock, stardate 2822-3.
Absolutely correct. I was somewhat bemused by all the hoopla yesterday about the G5 and it's 'speed'. I need to know how it will run programs that I will use. I don't run benchmark software very often. =)
I'm not a graphic artist, so Photoshop is unimportant to me. I don't render video, or manipulate sound, so that's not for me. I actually mostly use my home comp for games, the internet, watching movies and listening to music. Maybe it was optimistic of me to think that I was going to find a Mac that would fit my needs, but with all the hype about the G5, I thought I would finally have some reason to be interested in Macs. Does anybody have any numbers for any other programs other than Photoshop? At least some fps in Quake 3? (I don't play it, but it's a good game benchmark)
[SIG] It's like putting a moose in the blender -- a recipe for disaster!
Misleading Prices
Both Apple and Dell are guilty of using misleading
prices. For example, Apple gives the price of the
low-end G5 as "$1999", and the high-end G5 as
"$2999". In other words, they have subtracted
$1 from a $3000 computer to make it seem cheaper,
which is absolutely ridiculous. This
demonstrates that both Apple and Dell are willing
to mislead people when stating their prices.
Geez, that's so tricky of them, I fell for it!!
Fools, this is not something that Dell or Apple invented, this has been a standard marketing trick for at least 50 years if not more...and EVERYONE does it!
"Some things have to be believed to be seen." - Ralph Hodgson
I'm leaning towards a similar conclusion (``My next boxen will be PowerMac G5'') for when it's finally time to retire my NeXT Cube.
The new case design addresses most of my complaints about the old G3/4 design (funky round handles and irregular surfaces make stacking / arranging things around those problematic, noisy (but grant it is quieter than my NeXT Cube) drive panel access---I guess the SuperDrive has no buttons on the face plate beyond eject?)
and Panther finally brings most of the missing features from NeXTstep (Faxing, PostScript support, speed) and Mac OS 9 (Labels, apparently working QuickDraw/GX like font support).
I'd give my interest in Hell though for a way to change the monolithic, immovable main menu to a movable vertical menu a la NeXTstep (w/ top-level Print and Services!), esp. w/ tear-off sub-menus, and really wish that there was a language option which would give one concise NeXT-style menu shortcut descriptions....
William
Sphinx of black quartz, judge my vow.
And ironically, the problem is that they didn't benchmark differently enough: Apple used GCC to compile SPEC on the P4 and Xeon, as well as on the G5.
While this eliminates one variable from the comparison, it also eliminates a hefty percentage from the SPEC numbers one can get with Intel's compiler.
Not really ...
Hyperthreading does not make you have 2 processors. In a "high level" nutshell, it allows queueing of thread processes to utilize periods of inactivity in the processor to execute them. In other words, Intel took a processor optimization, and slapped a "cool" name to it. Its like calling mmap() "HyperCache". Maybe on RISC chips with the reduced instruction set we should tag that as "Hyperpumped Instruction Path"
In running a single threaded benchmark, it has very little bearing on the final outcome.
So no, hyperthreading is not like having 2 processors. More like 1 processor optimized for threads.
-- Knowing too much can get you killed, but knowing who knows too much can make you rich.
Database applications are the biggies, followed by cad/cgi. These also happen to be the applications which essentially pay the bills to may companies, so signifigant gains in processing can greatly impact profits.
You think that I'm crazy, you should see this guy!
Database applications dont have the database running on the client machine. They have it running on the oracle cluster or mainframe in the back room. The client side wouldn't need 2G of memory. And nobody in their right mind is going to run their DB server off of a client box.
I would say CAD only pays the bills at an engineering or architecture firm, and I think the best CAD packages are currently for PC. While the new apple box certainly opens the door up to porting to Apple, the lag time before Intel comes out with 64 bit proccessors wont be long enough for significant entrenchmant.
Actually, the practice originally comes from a tactic that store owners used to keep their cashiers honest.
If the product is $10, then they could just make change for $20 out of their pocket, i.e., hand the customer back a $10 bill and stuff the $20 in their pocket.
Of course cashiers would conveniently "forget" to stuff that $20 back into the drawer.
But if the product is $9.95, then they have to open up the drawer to get a nickel out.
When you add up that most customers would be like at least 2-3 items, products priced at $9.95 and $19.95 would cause the cashiers to *have* to make change out of the drawer, thus keeping them honest.
Little known fact, but it's true.
My journal has hot
Apple turned off hyperthreading in the Dell precision machines, and disabled SSE2. These are modifications you're gonna notice using photoshop, so those benchmarks say nothing.
---
"The chances of a demonic possession spreading are remote -- relax."
> A G5 is faster than the fastest Intel box with Linux. Read the benchmark whitepaper. It describes the testing methodology in precise detail. In a side-by-side, controlled test, the single-processor G5 was 10% slower on integer performance but 20% faster on floating point performance than the Pentium 4 with Linux.
Apparently they never got so far as Chapter 1 in Hennesy & Patterson, where you learn the mantra of "make the common case fast".
Sheesh, evil *and* a jerk. -- Jade
gcc produces inferior code on both platforms. Intel's C compiler kicks the shit out of gcc, and likewise metrowerks C and IBM's C compiler kick the shit out of gcc too.
gcc's x86 backend has had a lot more work than the ppc backend.
It would be interesting to see intel's C on x86 vs IBM's C on PPC. Compare chips and compiler writers with one stone :)
On the development machines we run, we have no less then 1 GB of RAM. We are compiling various apps, and though it would work with 256 or 512 of RAM, cutting 5 minutes off the compilation time is worth it.
Also, though I don't do graphics, I have several friends that do, and they regularly work with *huge* graphics files.
I think that for most end users and casual users 256 and 512 is enough for now. It is when you start pushing the bounds that the extra memory really helps.
Random Musings
If they would have just taken Gcc "out of the box" and benchmarked what you said would have been true. But they heavily optimised gcc by adding G5-specific code (from IBM's compiler? I hope IBM hasn't stolen it from someone else ;) and specific "lax" malloc() routines etc...
I'm sure that the actions, filters, and files that apple uses for their photoshop performance displays highlight the mac's prowess as much as possible while, at the same time, try to bash the x86 machine as much as possible.
I think apple's purpose for these claims goes beyond the fact that their trying to sell machines. They're trying to exterminate all of these myths that have been going around for the longest time about their hardware/ software. 90% of the people I know that don't like macs don't like them because the ones in their middleschool/highschool were horribly upkept and would not work or crash too often to be usable. I think that a major reason why apple went with the BSD underbelly in OSX.
Also, these fucken trolls on slashdot with that story of "my 350mhz g3 is barely usable if I'm copying a file and playing an mp3." Fuck that, I had my 132mhz 7600 (604 based machine) running fucken OS 7.6 and I could download, listen to mp3s, chat, and surf the web with minimal problems. Granted, I had 256mb of RAM in there, but it was fine. Only when the applications started getting more robust that that computer began getting unusable, and by that time, I had a 450mhz G3 which is STILL in use.
Although, apple does piss me off sometimes with their claims which, although true, ARE misleading and cause these mac fanatics to make outrageous claims based on Apple's statements/ demonstrations.
Although many mac users (fanatics?) are idiots, I think that there's a much higher percentage of windows users who are, too. And the windows users are much more likely to pick a fight about it.
Platform choice is a preference! Use what you like. Use what likes you. Use what makes you whole.
...spike
Ewwwwww, coconut...
I could swear I've seen this somewhere else today... hmm... yes that would be this article on AMDzone. Take a look at the second page for the SPEC score comparison...
we are all invisible unless we choose otherwise
http://www.amdzone.com/articleview.cfm?articleid=1 296
Ug... /.ed so soon.
Since I was bored, I went to look at apple's site. Something struck me as odd. I think they are using the same chip for all three models. From a "hurry and get these out the door" view it makes sense to use the same chip, why take the extra effort to detune them, just run it at a lower FSB.
1.6GHz PowerPC G5
800MHz frontside bus
200*8=1600
1.8GHz PowerPC G5
900MHz frontside bus
225*8=1800
2GHz PowerPC G5
1GHz frontside bus
250*8=2000
Quad pump them and there you have your 800,900,1000 FSB.
I'll be willing to bet that someone figures out how to make a the 1.6 a 2.0 within two months. Then again, I've been way off before. The MB could be waaaaay different.
I hate to say it, but I told you so.
The guy may, or may not be a troll. However, the sheer amount hate mail, and the level of it, was stunning. What kind of people write stuff like that? Very few of them even attempted to address the guys points, and those that did made a hash job of it (nobody uses int math? wtf?)
Did you notice how almost all of the hatemail was addressing him in the third person?
He went onto a discussion board somewhere about the post (probably MacNN, probably one of the worst reputation Mac websites in terms of brainpower) and just cherry picked the comments he could take apart easily.
It's not like he actually *got* that hatemail. He didn't even post an email address with the article.
Isn't it funny how you can bend things to make you look favorable - just like Apple may have done?
Was not my experience, actually... With gcc-3.2.x (the 3.3 is, supposedly, even better for SSE2/MMX2) on Windows (under Cygwin) I produced an executable, that worked slightly better than that produced by Intel's compiler (a lot of double-precision math).
Both of them were about 4 times faster, than the binary produced by the Visual C compiler -- from Microsoft.
YMMV, of course...
In Soviet Washington the swamp drains you.
From what *I've* heard the reason for the weird numbers wasn't necessarially to keep the cashiers from stealing, but to keep the stores from doing their sales under the counter (so to speak) and not reporting the sales tax. To get the drawer to open to get that nickel (actually, after taxes it would be more like 83 cents or something), you have to ring it up in the cash register, where the sale is recorded.
;)
Either way, its a pretty good explanation
If I have been able to see further than others, it is because I bought a pair of binoculars.
I know people are all ga-ga over the "G5" CPU and all of it's 64-bit capabilities, but the main thing for a LOT of people that will make these new Macs a lot more attractive is the rest of the system. We finally get a decent system bus, PCI-X and Serial ATA. Imagine how much this is going to better music and video production, being able to shove data in and out of various sub-systems that much more quickly. Top that off now with the CPU, and these machines DO smoke !!
...
;-)
The other issue raised in the article is "most user" tasks depend on integer performance, rather than FP. Well, at our institution (where we do medical imaging), we like (rather NEED) that FP rating. Right now, the system of choice has been Dell PCs, that people set up on their own with RedHat or Mandrake in dual-boot. We used to buy ASL, but have had just too many component failures to keep buying them. What people do is they do their processing and analysis in Linux (FP) and reboot or VMware into Windows to do their papers and presentations. So while we are not a huge market, the reasearch community is NOT a small one, and we are less price-sensitive than home consumers or industry. We need hardware to do what we need it to do, and we don't mind forking out the extra $$ to get it.
These new machines are better for us on two fronts. We can now do our analysis on a single OS and not have to bother about 2 OSes on our systems and also, we don't have to worry about setting up Linux, and the compatibility issues with hardware, etc
So when the time comes, I'll be asking my boss for one of these
If you visit the Apple store, click on the big advertisement on the center to select your PowerMac G5. In the upper-right part of the screen it states:
"Just how fast? Get the proof here.". Following this link will take you to Apple's own site where you can read details about the benchmark.
What's missing?
The comparison between G4-optimized benchmarks and the current G5-optimized benchmarks.
Ayup
actualy, if you read the source, you would notice a few things tweaked in the linux kernel (talking about 2.4.20ish kernels)
/*
/* /* This decides where the kernel will search for a free chunk of vm
to quote linux/include/asm-i386/page.h:
* This handles the memory map.. We could make this a config
* option, but too many people screw it up, and too few need
* it.
*
* A __PAGE_OFFSET of 0xC0000000 means that the kernel has
* a virtual address space of one gigabyte, which limits the
* amount of physical memory you can use to about 950MB.
*
* If you want more physical memory than this then see the CONFIG_HIGHMEM4G
* and CONFIG_HIGHMEM64G options in the kernel configuration.
*/
This is speaking about kernel memory limit, which leaves you with up to 3 gigabytes of space for user processes. That is the default, if tweaked, you can get it up higher to 3.5gigabytes... but that limits the kernel to about 500megabytes.
There are _other_ issues, when dealing with single processes, if your code staticaly allocates memory , like...
int foo[1000][1000];
the system normaly uses brk(); to allocate the memory.. this is done from the bottom up.. but if you use mmap(); to grab memory, it comes from the top down.
in include/asm-i386/processor.h there is another parameter that tweaks the memory used for mmap();
* User space process size: 3GB (default).
*/
#define TASK_SIZE (PAGE_OFFSET)
* space during mmap's.
*/
#define TASK_UNMAPPED_BASE (TASK_SIZE / 3)
this limits brk(); to the first gig of memory.. which causes some of my users's fortran code to blow up.
thankfully glibc is smart, and will brk() from the bottom if it runs out of mmap space. so i just tuned TASK_UNMAPPED_BASE to be TASK_SIZE - 0x40000000 for my cluster nodes. now I can use up to 2gig of memory for a single fortran process.
That's a really good point. I've been thinking about it, and I really want a G5, but it's just not worth it at this price point. If it comes in at around $2000 CDN ($1300 USD as you mentioned) in something like an iMac, I'm in. I'll keep my fingers crossed.
Random is the New Order.
I went from an Atari 800 with a Atari 410 Cassette drive to a Compaq Plus with a Floppy Drive & a 10MB Hard Disk Drive.
I refused to use the hard drive for like six months because floppy disk was more than enough for me.
Of course, I was like eight at the time too and had no real idea what a system running from a hard disk would perform like...
As a rock-in-roll Physicist once said, No matter where you go, there you are.
The reason people care is when you go to the Apple site there is a big headline that the Apple G5 is the world's fastest desktop computer, when in fact any way you cut it, it is not. If any other company pulled this kind of shit it would be ridiculed in a minute, but Apple abuses the loyalty of its users. As someone who uses macs occasionaly at work and home, I like the product but hate the BS tactics of the company and stupidity of a small but vocal portion of its user base.
That's true, but for decades Apple has had the perception in some circles as a company that does things "differently." Hell, that's their whole reason for existence, right? So I think there are people who will be genuinely surprised to discover that Apple has so THOROUGHLY cheated. I mean, this makes nVidia look like saints.
No, I am not a Mac zealot who thinks that Intel or Gates, or whoever it is that day is the devil.
I think that's obvious from the fact you aren't threatening to kill the guy who wrote the article. Evidently unlike most of the people writing him email. Seriously, what's with these people coming so unglued?
-Looking for a job as a materials chemist or multivariat
gcc produces inferior code on both platforms. Intel's C compiler kicks the shit out of gcc, and likewise metrowerks C and IBM's C compiler kick the shit out of gcc too.
:)
not necessarily. we've production code that is 8x faster on x86 w/gcc than intel's icc 7.0. we're in discussion with their engineers about why. that blew my mind, though.
just a note, so you don't take it for granted
A good benchmark would be: Digidesign's Pro Tools Free
Bzzt. There's no free Pro Tools for Mac OS X.
The Logic versus Cubase test was perfectly valid. It tells us that Logic on a G5 can do much more than Cubase on a Dell. Just because the test doesn't tell you what you want to know, that doesn't mean it's not a valid or useful test.
is it's all crap. I don;t care if your Intel, AMD, or Apple. They all lie about how fast there cpu's are and benchmarks can always be easily scewed - whether its by turning off SSE2 (as in this article) or if they change the bios settings from the default on one machine (as toms hardware said AMD recommended for their 3200+ benchmarking) - its all made up.
I'm never going to ALWAYS have the latest and greatest system (My pc is a P4 1.6 and the mac i'm using to type this is a Dual G4 500) but what makes my choice for what System to use comes down to one simple thing....
WHAT THE SOFTWARE I AM USING RUNS BETTER ON.
When I want to use telnet, ssh, ftp (cuteftp, dreamweaver), or any internet related app - I find that for my setup they seem to work better on my pc for some reason, when I want to run Photoshop (although when working with files over 80megs it seems to open faster and run filters faster on my PC) or Illustrator or even Quark they seem more responsive on my Apple.
If I had to make a choice and choose only 1 system (glad I don't have too) I would probably choose a really expensive PC (like a Dual P4 3.2ghz with HT or something) ONLY because it would be cheeper for me to build (read not purchase a Dell) than the cost of a single mac.
But in real life I don't have to make a choice or rather i've made the choce to use both - and hay I can still play Diablo on both too - although if I want to play more games like Arcanum or Final Fantasy or Neverwinter Nights i'm kinda stuck only using my PC.
Ave Molech Setting
Just tell me how fast Lightwave renders. That is where time actually equals money for me.
"Derp de derp."
I assume it's just the laws don't require the price to reflect sales tax. The store isn't going to put a price any higher than it has to. But in a way, that does keep the taxes lower. I think people would be much more upset if they bought 10 gallons of gas at 75 cents/gallon and had to pay $12.50 for it (gas prices in the U.S. are always shown including tax). There are occasionally other places such as movie theaters that also include tax in their price.
Of course, the real kicker is that coupons seem to be applied after the sales tax is calculated. So, you pay sales tax on the marked price of the item, not the actual price you're paying. That, or the stores are pocketing the difference...
"Save the whales, feed the hungry, free the mallocs" -- author unknown
I think it's just fine if apple uses a specialized malloc() library. It was built to take advantage of the G5 processor, and they are candid about that. What is really going on is that they are trying to even the playing field a bit, because the compilers used on the intel-type machines have been in development on those processors for a long time. They already have a lot of optimization built-in. The G5s are supported, but haven't been for long. Apple is trying to simulate the malloc() library that might be used later, once they can makemore efficient. Oh, and not to mention, this 'argument' he puts together depends on you not reading the same PDF he claims to have read. Go download it and make your own informed conclusions. From the perspective of a long time programmer and x86 user...this doesn't appear too misleading. Anymore than what intel clims HT can do (another flaw in this guy's argument).
====
Crudely Drawn Games
With Apple's price point at $3K, they're priced up there with the entry level high-end workstations. HP's Itanium 2 workstation sells for around $3.3K. Sun's base 64-bit workstation is a little under $2K. So Apple's 64-bit offerings have to be compared with the expensive boxes, not what's selling at WalMart.
Apple is probably ahead on price/performance and usability in 64-bit desktops, but they're not first.
...of that Haxial article is Adam Hinkley, the guy who made Hotline. Just in case anyone cares.
In a 64-bit address space, all your pointers take up twice the space.
This IS SIGNIFICANT in many applications, and they must be run in the 32-bit mode.
The extra address space helps you only if you're willing to spend another several hundred dollars to get over 2 gigs of RAM, or are willing to put up with a huge swap.
I hereby place the above post in the public domain.
You don't get it. The only relevance the compiler has is with how good the resulting code is. I could write a compiler that generates code for x86 line and for the PPC line and in the PPC code generation I could emit a short loop after every actual instruction which caused the code to run at 2% the CPUs actual ability. Would that make it a fair comparison? No. If I just generated bad code would that be a fair comparison? No. So the point is that you should be using the best available compiler for each architecture. Use intel's compiler for x86 chips. Use IBM's compiler (assuming that they have one) for PPC chips.
Oh for god's sake just drop it already. We all know you CANNOT compare PPC and x86 processors! With clock speed, it's like Miles and Kilometers, and every system setup is different, conditions are different, hell even temperature can affect performance. Comparing the Apple results to the Intel results... its more likely that Intel has ALSO manipulated their tests!
That would be a first. Everyone I know who write FP-heavy code other than for playing around or proof-of-concept type of apps tries to avoid GCC due to it's low performance, just like I do. On x86, Microsoft's, Intel's and Metrowerk's compilers all outperform it(Don't know about Inprise's compiler, haven't used it in a long while). On Mac's, Metrowerks(The only one I've used on Mac's) compiler outperforms GCC. On MIPS, it's MIPSPro or Metrowerks that counts, GCC is right out, the performance so abyssmal that one sometimes thinks it's interpreted code, rather than compiled.
And GCC's focus is not on getting maximum performance, but to be an Open Source compiler. Lots of target platforms, too many cooks involved in the soup, so performance will never be optimal, but you will find it on many platforms instead.
Apple blew the Dell away with Photoshop, PDF viewing, and Mathematica.
I think the point is that Apple crippled the Dell, turning off Hyperthreading for example. So you can't even trust the Photoshop benchmarks.