Apples are Slow

[Mav451]

To answer someone's earlier question, my PB has 768 MB RAM. My brother's computer has 512 MB (DDR 2700, maybe 2100 I believe), and our old Dell has 256 MB. But here's the thing. I build what I thought would be a screaming machine: 3.0 GHz P4 Prescott HT, 512 MB PC3200 RAM, Asus P4P800 motherboard, GeForceFX 5950 Ultra, SATA HD, overclocked... I thought it would be pretty fast, but it gave me many problems that vastly outweighed the speed. The games didn't run as fast as they should have; after turning off overclocking, things were negligibly more stable. I got so sick of the computer that I just sold the parts on eBay one day. I'd have to say that the Mac runs faster than that machine.

Now, I'm wondering if the operating system is to blame. I've been thinking about how OS X differs from Windows, and I've come to a few conclusions. XP seems to require slighly less or about the same amount of RAM as OS X; 512 seems to be good for XP, while both 640 and 768 have worked well for me in OS X (I haven't tried 512). But about the OS itself: OS X seems to "cover up" the underlying processes, kernel, etc. better than Windows XP does. To explain what I mean, I'll use an analogy: the OSs used in cell phones, game consoles, and pocket/palm PDAs are extremely functional, yet display only pertinent information; navigation, use, and speed are the most important aspects. Now, OS X seems to me to be much like those OSs, except that the user can find great functionality that may not be immediatly obvious. XP, on the other hand, seems to sacrifice these good aspects for immediate functionality; that is, advanced features seem more accessible, and the OS seems to make less of an attempt to cover up its inards.

It's difficult to explain. Does anyone else know what I'm talking about / feel the same way?

Mike LaRiviere

Mike: Enthusiasts (including myself) mostly still recommend the 875/P4 800fsb combo to the prescott. One, the Prescott is actually slower at the same clockspeeds (3.0C vs. 3.0E, 2.8C vs. 2.8E is even more apparent). Two, the Prescott runs ALOT hotter. Intel has a throttling (clocking down) mechanism when it overheats--maybe one reason you feel XP is slow.

Third, again, megahertz myth + Intel. I'd take my AMD setup anyday. My Athlon XP-Mobile running @ 2.53Ghz easily outguns a 2.53 Pentium 4, much less even a 3.0 or 3.2Ghz P4. That is the TRUE benefit of AMD. Unlocked multipliers

Its unfortunate you had a bad experience with your PC. I think that whenever you first build, you should consider the mature platform and not the cutting edge (and then not expect bugs--there will be many). This is why I still recommend nForce2 to anyone beginning to build b/c it has had an entire year to fully mature. I've built 3 boxes with the Asus A7N8X (2600+, 2500+ to 3200+, and my setup of course) and they have nary a problem.

*on XP handling RAM--yeah, I think since it can run with only 256MB, having 512MB or 1024 doesn't help AS MUCH overall compared to OSX. Of course, having 1024MB for Photoshop or Doom 3, those are obvious, but these are software-SPECIFIC situations (not overall OS responsiveness).

One thing I know is that I should get more RAM--my iBook absolutely chokes if it runs QT/Camino + iTunes at the same time.

 

[Timelessblur]

I will not get over how Apple refuse to consider looking at AMD chips for comparing to.

 

[MikeLaRiviere]

Mav, you make some good points. It wasn't until after I'd purchased the Prescott that I read about how it was hotter and slower than the previous design. Further, I should have read more about AMD chips. For some reason, although I knew about the "megahertz myth" to a degree, I still went with the higher clock speed. Interestingly, the first homebuilt I made worked pretty well - it was a 2.4 without HT, GeForceFX 5600 I believe, lesser motherboard but similar overall. This one, however, gave me a lot of problems.

As to my brother's computer, I'm going to have to reinstall the OS, because I know that it should be running a lot faster than it currently does. Don't get me wrong - I do like Windows and the x86 architecture, and for gaming it can't be beat. But some combination of the Mac OS and the PowerPC architecture make the computer run faster than the Windows computers I use.

Mike LaRiviere

 

[MacsRgr8]

I agree for the most part, except that OSX can be less "snappy" as Windows, which is true. For example, when resizing a window in OSX, it has to catch up to the curser, but on Windows, it stays on it. Now, this barely makes a difference of half a second < snip > .

This is unfortunately what many people see as "perceived speed". And they tend to find it hard to forget.

If the ignorant buyer walks up to a brand new iMac (or even a Dual G5 for that matter!), opens Explorer (yep, ignorant ), and starts to resize the window, you'll hear him call out something like "what the heck? this is slow!" And you'll be needing a very good salesman to explain that window resizing is NOT a way to measure speed..... This salesman will be defending the Mac's speed instead of showing its beauty.
I tell you, I have heard this happen a couple of time in stores. Pretty frustrating.

And, ofcourse, games.. games.. games.

 

[MrSugar]

This is unfortunately what many people see as "perceived speed". And they tend to find it hard to forget.

If the ignorant buyer walks up to a brand new iMac (or even a Dual G5 for that matter!), opens Explorer (yep, ignorant ), and starts to resize the window, you'll hear him call out something like "what the heck? this is slow!" And you'll be needing a very good salesman to explain that window resizing is NOT a way to measure speed..... This salesman will be defending the Mac's speed instead of showing its beauty.
I tell you, I have heard this happen a couple of time in stores. Pretty frustrating.

And, ofcourse, games.. games.. games.

So why isn't is as snappy? I notice with my G5 it's very very quick but it isn't as snappy as my old 2700+. Mac is worth it, but I would be curious to know why OSX can't be as quick as windows in certain things...

 

[edesignuk]

My 1.33 GHz G4 PowerBook is faster than most, if not all, PCs I've used. My brother's 2.4 GHz P4, for instance, is excrutiatingly slow - and it's got 512 MB RAM.

That 2.4GHz P4 must be so full of sh*t running in the background (spyware, adware). A 2.4GHz P4 pisses all over a 1.33GHz G4 in raw speed. Of course, that's no good if you like OS X.

 

[MacsRgr8]

So why isn't is as snappy? I notice with my G5 it's very very quick but it isn't as snappy as my old 2700+. Mac is worth it, but I would be curious to know why OSX can't be as quick as windows in certain things...

Heavy eye-candy: Anti-aliassing, on screen PDF etc.
Real time indexing.... it all takes time.

Beauty comes with a penalty: speed. Well... "snappiness".
Faster searching comes with a penalty.. Less "snappy" opening of new (especially remote) directories...

 

[MikeLaRiviere]

EDesignUK, of course there's no spyware/malware running in the background. I don't allow any computer in my house to have that kind of junk, and I install updates and run numerous checks on an almost weekly basis. Believe me, I lock down these computers like a fortress. And when something comes up, I run a number of all-out checks.

I do believe there is something wrong with the registry. But believe me, the Mac is faster than his PC. With regard to slow window resizing, I think those of us who know Macs concur that a) it is poorly-written code and b) it is a task assigned to the CPU which should be assigned to the GPU. This is the only aspect of the GUI that lags compared to Windows. If Windows did any of the bells-and-whistles things that OS X does, it would quickly choke systems with the same clock speed.

Another comparison: my recent 800 MHz iBook was much faster than my 900 MHz Duron Compaq laptop from a few years ago. The Compaq could barely run XP with its "pretty" settings turned on. A system comparable to the 800 MHz iBook would be the year-old Toshiba Satellite with a 2 GHz Celeron.

Mike LaRiviere

 

[friarbayliff]

True, you have to remember that the responsiveness of the GUI is superficial and doesn't determine the speed of the computer. OSX has far more going on in terms of eye-candy and refinement than windows does - therefore, PCs tend to have percieved speed, but not real speed

 

[Bruce Lee, PhD]

So why isn't is as snappy? I notice with my G5 it's very very quick but it isn't as snappy as my old 2700+. Mac is worth it, but I would be curious to know why OSX can't be as quick as windows in certain things...

i think some of it is the eye candy. but it was my understanding that the pokey GUI is also in some sense the dark side of a nice multitasking OS. i'd defer to someone who knows more, but i think it boils down to the fact that since BSD has good time-slicing (which wasn't designed with GUI in mind) it's difficult to monopolize the CPU enough to get real-time responsiveness in the GUI. thus you end up with wierd hacks like increasing the keyboard repeat rate improving GUI snappiness, since this forces the machine to spend more time in the relevant tasks.

do any kernel folks know the details or truth of this?

 

[Mav451]

I don't think Apple will be changing that aspect at all...perhaps its more of a unique GUI element, if you will.

With Windows you usually are running the classic interface (even more "snappy" if you are running something old like a Pentium II, in comparison to the bubbly "Fisher Price" theme); OSX on the other hand comes with a beautiful GUI. Of course, I still don't use my iBook for much longer than once or twice a week, so maybe the same theming "bug" i had on my PC isn't hitting me yet on the Mac.

 

[MikeLaRiviere]

What's this keyboard repeat rate hack? I had heard of it, but I didn't realize that it affected anything other than the typing rate, i.e., holding down a key and seeing how fast it would repeat. Some explanation? This sounds cool.

Mike LaRiviere

 

[Mav451]

True, you have to remember that the responsiveness of the GUI is superficial and doesn't determine the speed of the computer. OSX has far more going on in terms of eye-candy and refinement than windows does - therefore, PCs tend to have percieved speed, but not real speed

Well speed can be interpreted in a whole lot of ways:

1)Opening of applications?
2)The speed of the application when it is actually running?
3)GUI responsiveness (minimzing, resizing, moving)?

In that way, the first two are less dependent on your OS, and more on your hardware. I mean, 5400RPM drive vs. 7200? Or 256MB RAM compared to 1Gig of RAM? Those two factors heavily impact the first two.

 

[edesignuk]

Another comparison: my recent 800 MHz iBook was much faster than my 900 MHz Duron Compaq laptop from a few years ago. The Compaq could barely run XP with its "pretty" settings turned on. A system comparable to the 800 MHz iBook would be the year-old Toshiba Satellite with a 2 GHz Celeron.

Duron & Celeron CPU's are cheap crap (especially Celerons!!). I'd expect anything to wipe the floor with them!

 

[Mav451]

EDesignUK, of course there's no spyware/malware running in the background. I don't allow any computer in my house to have that kind of junk, and I install updates and run numerous checks on an almost weekly basis. Believe me, I lock down these computers like a fortress. And when something comes up, I run a number of all-out checks.

I do believe there is something wrong with the registry. But believe me, the Mac is faster than his PC. With regard to slow window resizing, I think those of us who know Macs concur that a) it is poorly-written code and b) it is a task assigned to the CPU which should be assigned to the GPU. This is the only aspect of the GUI that lags compared to Windows. If Windows did any of the bells-and-whistles things that OS X does, it would quickly choke systems with the same clock speed.

Another comparison: my recent 800 MHz iBook was much faster than my 900 MHz Duron Compaq laptop from a few years ago. The Compaq could barely run XP with its "pretty" settings turned on. A system comparable to the 800 MHz iBook would be the year-old Toshiba Satellite with a 2 GHz Celeron.

Mike LaRiviere

If you compare the systems, RAM vs. RAM, I think the Duron would be faster. I had a 800 Thunderbird (+/- a 950 Duron) with only 128MB of RAM and I ran Windows 98 on that for 3 years. I can guarantee you that it is easily 4-5 times more multitaskable than my 800 iBook. That means have 3-4 windows of IE open (no Firefox back then haha), Divx movie playing, CD-burning, and some (ahem) network bandwidth intensive apps haha. My iBook already chokes on running a SINGLE Divx movie running from the CD-ROM! Once I open iTunes and Camino, the grinding begins A little disappointing, but I do realize the i have a g3, not a g4

 

[MikeLaRiviere]

EDesignUK and Mav, I'll address your assertions point-by-point. EDesignUK, I agree with you that the Duron is not a good processor. My PIII 933 is much faster than the Duron 900. However, the Celeron 2.0 I used was fine for what I needed - basic applications, and it ran Counter-Strike well. They're "not good" in that they have low amonts of cache, but for basic computing, they're fine. And they are not unwarranted for use in comparison to the iBook 800 MHz.

Mav, you are using a G3, as you conceded. Using the G4, I found the iBook to be immeasurably faster than the Compaq Duron. Further, you're using Jaguar, which I found to be excrutiatingly slow compared to Panther - you might dismiss my claim by showing me benchmarks, but believe me, Panther feels much faster, as any Jaguar-to-Panther upgrader will tell you.

EDesignUK, you further said that the 2.4 P4 kills the 1.33 G4 in raw speed. Well, if you're talking about cycles per second, you're right, it's exactly 1.07 GHz faster. Factor in pipeline architecture and instructions per cycle, and the speed winner will surprise you. The AMD folks are on the right track, and the benchmarks back them up. When it's possible to accurately benchmark a PowerPC against an x86 - and each currently uses a different version of the benchmarked program - then we can compare benchmarks. But for now, we'll watch as AMD trounces comparably-spec'ed Intels.

Mike LaRiviere

 

[MrSugar]

I have to side with edesignuk, a 2.4 whipes the floor of a 1.33 ghz Powerbook. While the powerbook is in no means slow, I have used very fast wintel computers, the are quick. My G5 is quicker, but they were still very fast.

It's not worth giving up OSX for but a 2.4 P4 will most definitely run faster than a 1.33 ghz G4.

 

[MrSugar]

EDesignUK and Mav, I'll address your assertions point-by-point. EDesignUK, I agree with you that the Duron is not a good processor. My PIII 933 is much faster than the Duron 900. However, the Celeron 2.0 I used was fine for what I needed - basic applications, and it ran Counter-Strike well. They're "not good" in that they have low amonts of cache, but for basic computing, they're fine. And they are not unwarranted for use in comparison to the iBook 800 MHz.

Mav, you are using a G3, as you conceded. Using the G4, I found the iBook to be immeasurably faster than the Compaq Duron. Further, you're using Jaguar, which I found to be excrutiatingly slow compared to Panther - you might dismiss my claim by showing me benchmarks, but believe me, Panther feels much faster, as any Jaguar-to-Panther upgrader will tell you.

EDesignUK, you further said that the 2.4 P4 kills the 1.33 G4 in raw speed. Well, if you're talking about cycles per second, you're right, it's exactly 1.07 GHz faster. Factor in pipeline architecture and instructions per cycle, and the speed winner will surprise you. The AMD folks are on the right track, and the benchmarks back them up. When it's possible to accurately benchmark a PowerPC against an x86 - and each currently uses a different version of the benchmarked program - then we can compare benchmarks. But for now, we'll watch as AMD trounces comparably-spec'ed Intels.

Mike LaRiviere

Not to start anything, but I really would like to see any sort of benchmark out there that shows a 1.33 G4 running faster than a 2.4 P4. I really don't think they exist. Maybe in certain programs, but not overall. It was my understanding that there were benchmark programs that showed accurate levels of speed, perhaps not though.. I would be curious to know.

 

[rueyeet]

"Perceived speed"....

I'm not one of those who has dropped the cash for the extra RAM for my PowerBook, so yes, I notice that the interface is just a little bit less responsive than Win 2k Pro on the P4 I use at work. The resizing thing in particular is annoying because it's such a no-brainer thing in Windows, and yet on the Mac you've got to position the window and then size it, instead of just dragging out whichever side or corner is convenient.

But that's more than balanced out when someone sends a message on Outlook with a large attachment, and Outlook has to stop and think about it interminably before loading the message text in the preview pane. Or when I copy a table from Access and open Excel to paste it in, and Excel takes for-freaking-ever to open because the copied table is apparently hogging memory (yes, I know you can export from Access to Excel, but the results aren't formatted as well as for a copy/paste), and...well. I could go on and on with the constant little "gotchas" of where Windows and its approved apps completely negate whatever "snappiness" my completely stock Powerbook lacks.

Upshot? I don't care about a bitty little barely perceptible difference in interface speed, as long as I don't voluntarily have to spend any more time with Windows than I have to.

 

[edesignuk]

-MikeLaRiviere

I'm not an Intel fanboy, and I agree that AMD clock for clock are faster than Intel. But, that 2.4Ghz P4 should still wipe the floor with a 1.33Ghz G4 in the vast majority of tasks. Then again as I said before, it all depends on how well you get on with the OS and can navigate, this will play a big part in the speed you "feel" you a getting.

 

[TDM21]

Just to throw my 2 cents into this conversation, but would measuring a computer's "power" in flops not be a really good comparison between wintel and mac?

 

[Mav451]

Flops is as bad, if not worse than the Megahertz Myth. Whenever, I want a laugh (well geeky laugh haha), I go to this hilarious site:

http://forgetcomputers.com/~jdroz/pages/09.html

Its kinda scary how far they take it, and how much the guy actually believes it!

Apple 933 MHz G4 Power PC 13,400 (January '02)
AMD dual 246 Opteron 13,667 (Aug '03)

G4 = 2 Opterons? I don't think so.
We're talking about 167 bus vs. 1600 HyperTransport here.

 

[BornAgainMac]

Ghz

"Ghz".

As a Mac user, I hate that word. I try not to mention it when I talk to my Windows friends when going over specs on my Mac. If they ask, I show them the dock effects and they get distracted and forget they asked me.

 

[Timelessblur]

Things that are loaded full of BS and miss infomation in the computer world

1. Mhz means all
2. FSB speeds that tell you are utter crap. 800mhz is marking it has a 4x mulitpler on it and they are easily be pushed.
Even on P4 mobo if you look in the bios the FSB is 200mzh. The 800 mhz number comes from other location around there that can be messed with. (I dont remeber exaclyt where it is in the BIOS. On my AMD chip FSB is currently overclock to 220 Mhz putting my chip at 2.2Ghz. I do know there is on place in the BIOS that I can change a number that is around 800mhz and I can see how you can call it FSB but still 200 Mhz is the real FSB.

CPU clock speed is FSB time a mulitpler. My AMD 64 has a 10x mulitpler so a 200mhz FSB puts my chip at 2.2 Ghz. a lot of penitums have a 17.5x mulitpler on them so they go even higher. (hence the reason I think the 2.5Ghz G5s are overclock 2Ghz G5 because there reported FSB speeds are even higher than 800 mzh)

 

[The Cheat]

(hence the reason I think the 2.5Ghz G5s are overclock 2Ghz G5 because there reported FSB speeds are even higher than 800 mzh)

The 1.8GHz and 2GHz G5's already have FSB speeds above 800MHz. All G5's run a bus speed that is half that of the core (in other words they all have a 2x multiplier). Therefore the new 2.5GHz model has a FSB of 1.25GHz. It's no mystery.