New Intel CoreDuo Mac Mini is much slower than expected...

[TheMac19]

I just, finally, talked my parents into getting a Mac. As I'll be around for a couple weeks, I'm helping them set it up and learn all the "nice-ities" it has to offer.

Gotta say, I'm a bit surprised and fairly bummed at the performance. I'm coming from a 1.5Ghz G4 powerbook, and the mini feels noticeably slower! And that's just running bare basic stuff - dashboard with stock widgets, iTunes, Safari, etc. All universal Apps, and it really feels considerably slower than I was used to with my old powerbook, and I ran all sorts of nonsense on my powerbook!

It's the 1.66 Ghz CoreDuo model with 512mb of RAM, and although I know it could use more ram, (and my powerbook had more) surely that's not the problem for it running so slowly when it's not working hard on anything.

For instance, the widgets rarely fly in smoothly, or make a nice ripple effect when opening. (Maybe that's just the integrated graphics???) But it's not just them. Even just opening safari can take quite a few seconds, same with mail.app. Several beachballs around, and a few times I've had it give up the ghost and just plain refuse to open apps until after a restart. (They'd just bounce in the dock indefinitely until I force-quitted them and restarted, although that's only happened a few times, and I've attributed it to fast user switching.)

I've repaired permissions, and tried the maintenance stuff, and the disk verifies fine, so I'm not real sure what to do. I guess nothing is really wrong, but I sure was expecting better performance, based on what I've heard from people and their macbooks. Sometimes the computer feels better than others, so I'd feel silly bringing it to the genius bar. Are there any processor tests or benchmarks I could run for free to see if this thing is performing as it should?

 

[unfaded]

More. RAM.

Yes, Apple gives you 512, and Apple should be giving you a workable amount, but 512 really doesn't cut it. Rosetta is very, VERY memory hungry as well, so any when you're running ANYTHING PPC, even if it's just in the background, you're losing a lot of memory right there.

 

[bousozoku]

It's not surprising. With only 512 MB, the PowerBook doesn't run all that quickly either. Visual effects in Finder and Dashboard are going to run faster and/or smoother with the PowerBook due to the Radeon Mobility 9700 graphics hardware. That's just the native stuff. As has been said, anything touching Rosetta needs a lot more RAM.

 

[gekko513]

All the examples you mention are not processor intensive.

Dashboard depends on RAM and on the GPU. Starting apps depends on the hard drive, RAM and I/O in general. Also you mention Fast User switching which needs even more RAM to run smoothly.

The processor in the Mac mini is very fast, but the lag and lack of responsiveness that you're experiencing is RAM and I/O dependant. If you want to test the speed of the processor, you'll have to try things like converting or rendering video or audio.

To improve on the lag issues, get more RAM.

 

[IJ Reilly]

More. RAM.

Not according to this thread. These bench tests on an architecturally very similar MacBook actually show some loss of performance with 1 gig and 1.5 gigs over the stock 512mb. So RAM is not a cure for all performance ills.

Beside, the OP hasn't provided any detailed explanation for why he thinks the Mini is "slow," except for how it "feels." And yet, everybody is in a hurry to spend his money...

 

[gekko513]

Not according to this thread. These bench tests on an architecturally very similar MacBook actually show some loss of performance with 1 gig and 1.5 gigs over the stock 512mb. So RAM is not a cure for all performance ills.

Beside, the OP hasn't provided any detailed explanation for why he thinks the Mini is "slow," except for how it "feels." And yet, everybody is in a hurry to spend his money...

He did give several examples, and they're all RAM and I/O related. The benchmarks measure performance of a type that the OP hasn't expressed any worry about, so more RAM is definitely the solution here.

Edit: And besides, as far as I can see from the benchmarks in the first post you linked to, there was noticable performance increase up to 1.5GB.

 

[unfaded]

Not according to this thread. These bench tests on an architecturally very similar MacBook actually show some loss of performance with 1 gig and 1.5 gigs over the stock 512mb. So RAM is not a cure for all performance ills.

Beside, the OP hasn't provided any detailed explanation for why he thinks the Mini is "slow," except for how it "feels." And yet, everybody is in a hurry to spend his money...

Yes well if he had said "man, my Cinebench scores totally suck" I would have told him to decrease his RAM.

 

[IJ Reilly]

Yes well if he had said "man, my Cinebench scores totally suck" I would have told him to decrease his RAM.

We were discussing the performance benefits of increasing RAM, not decreasing it. Anyway, you're invited to submit objective testing that shows a different result. I'd also like to see the objective testing that shows a G4 1.5 beating a Core Duo. If we're going to rely on "feelings" instead of actual measurements of things than can be actually be measured, then we're talking about witchcraft, not computers. No thanks.

 

[gekko513]

IJ Reilly, whether you like it or not, the observation of a beachball is a valid measurement and not "witchcraft", as is the observation of number of page ins and page outs in activity monitor.

I don't even need a benchmark to tell me that with more RAM I can have more apps open with more content and switch between them without beachballs than I can with less RAM.

 

[timswim78]

I agree with the OP. From my experiences, the new Intel Mini's are horribly slow with 512MB of RAM. Perhaps a RAM increase would help, but I never gave it a chance.

I sold the Mini (luckily I only lost $10), and I bought a used dual G4 PowerMac. So far, the PowerMac has done exceedingly well at multitasking, whereas the Mini suffered horribly whenever I tried to run more than one app at a time. I intend to use the G4 until the Intel Mac's are a little more mature.

 

[gloss]

The problems he's mentioning are definitely caused by lack of sufficient RAM. I'd get another 512 or (if you have the monies) 1.5gigs and see how things work then.

 

[Pressure]

Not according to this thread. These bench tests on an architecturally very similar MacBook actually show some loss of performance with 1 gig and 1.5 gigs over the stock 512mb. So RAM is not a cure for all performance ills.

Beside, the OP hasn't provided any detailed explanation for why he thinks the Mini is "slow," except for how it "feels." And yet, everybody is in a hurry to spend his money...

Remember that the Mac Mini relies on the integrated graphic chip for graphics (which uses system memory) and the available bandwidth is effectively doubled when running with paired memory sticks (dual channel).

It has little to no effect to the processor, whether or not you run in dual- or single-channel mode.

 

[QCassidy352]

RAM is not a cure for all performance ills.

When you have 512 and you're using Rosetta, actually, yeah, it is. When I first got my macbook, with stock RAM, it felt slower than my ibook G4 with 1.5 GB RAM. I went up to 1 GB, and the macbook now feels immeasurably faster than the ibook.

This is easy to test. Go to activity monitor, look at memory usage, and see how many page outs you're getting. When I had 512, i was over 10,000 page outs during a full day of use. Since going to 1 GB, I get zero. Menu meters shows the same thing. I would frequently be using every meg of RAM I had before, and now I usually have a few hundred free.

512 RAM is not enough for OS X. I mean, it'll run, but not like it should. This is especially true for intel macs that need Rosetta. I don't think just throwing money at a problem is always the way to go, but this is a no-brainer. Bring that mini up to 1 GB and it will feel like a different machine.

 

[netdog]

Upgrade the RAM. It isn't really even 512MB as the shared graphics processor is stealing resources.

You need two matched sticks to have the graphics run dual channel. You want at least 2x512MB sticks for 1GB, and preferably 2x1GB sticks for 2GB. Then you will really enjoy your Mini. It really is a RAM problem.

 

[IJ Reilly]

Remember that the Mac Mini relies on the integrated graphic chip for graphics (which uses system memory) and the available bandwidth is effectively doubled when running with paired memory sticks (dual channel).

It has little to no effect to the processor, whether or not you run in dual- or single-channel mode.

The MacBook and the Mini are virtually identical in every specification. I know the benchmarks don't report the results of adding RAM that many people expected, but that doesn't make them wrong.

 

[IJ Reilly]

When you have 512 and you're using Rosetta, actually, yeah, it is. When I first got my macbook, with stock RAM, it felt slower than my ibook G4 with 1.5 GB RAM. I went up to 1 GB, and the macbook now feels immeasurably faster than the ibook.

This is easy to test. Go to activity monitor, look at memory usage, and see how many page outs you're getting. When I had 512, i was over 10,000 page outs during a full day of use. Since going to 1 GB, I get zero. Menu meters shows the same thing. I would frequently be using every meg of RAM I had before, and now I usually have a few hundred free.

512 RAM is not enough for OS X. I mean, it'll run, but not like it should. This is especially true for intel macs that need Rosetta. I don't think just throwing money at a problem is always the way to go, but this is a no-brainer. Bring that mini up to 1 GB and it will feel like a different machine.

The OP was using Intel binary applications, so Rosetta doesn't come into play. He is saying his G4 1.5 "feels" faster. How, and in what way, we don't know. This matters. Once again, I challenge anyone to produce objective bench tests that show the G4 1.5 performing better than a Core Duo.

 

[QCassidy352]

IJ, I'm just curious - what are you arguing for/trying to prove here?

The OP maybe was a little inexact in his wording, but we all (hopefully ) know what he means when he says that the mini "feels noticably slower." He means it's sluggish and the powerbook isn't; we don't need benchmarks to understand that. He even specifies tasks/apps like widgets, mail, internet, and ilife.

I, and many other posters here, have used enough different macs running OS X with different specs to understand what the OP means and to know that it's a RAM issue. Do you seriously disagree with that?

 

[macgeek2005]

I just, finally, talked my parents into getting a Mac. As I'll be around for a couple weeks, I'm helping them set it up and learn all the "nice-ities" it has to offer.

Gotta say, I'm a bit surprised and fairly bummed at the performance. I'm coming from a 1.5Ghz G4 powerbook, and the mini feels noticeably slower! And that's just running bare basic stuff - dashboard with stock widgets, iTunes, Safari, etc. All universal Apps, and it really feels considerably slower than I was used to with my old powerbook, and I ran all sorts of nonsense on my powerbook!

It's the 1.66 Ghz CoreDuo model with 512mb of RAM, and although I know it could use more ram, (and my powerbook had more) surely that's not the problem for it running so slowly when it's not working hard on anything.

For instance, the widgets rarely fly in smoothly, or make a nice ripple effect when opening. (Maybe that's just the integrated graphics???) But it's not just them. Even just opening safari can take quite a few seconds, same with mail.app. Several beachballs around, and a few times I've had it give up the ghost and just plain refuse to open apps until after a restart. (They'd just bounce in the dock indefinitely until I force-quitted them and restarted, although that's only happened a few times, and I've attributed it to fast user switching.)

I've repaired permissions, and tried the maintenance stuff, and the disk verifies fine, so I'm not real sure what to do. I guess nothing is really wrong, but I sure was expecting better performance, based on what I've heard from people and their macbooks. Sometimes the computer feels better than others, so I'd feel silly bringing it to the genius bar. Are there any processor tests or benchmarks I could run for free to see if this thing is performing as it should?

Don't complain. Apple's designers sometimes have to take a crap, they are humans after all, and they take ***** just like everyone else. They come out in the form of the Mac Mini.

People expect it to be faster than it is, and I have news for you, IT'S NOT GETTING ANY FASTER, IT'S A PIECE OF CRAP.

 

[IJ Reilly]

IJ, I'm just curious - what are you arguing for/trying to prove here?

The OP maybe was a little inexact in his wording, but we all (hopefully ) know what he means when he says that the mini "feels noticably slower." He means it's sluggish and the powerbook isn't; we don't need benchmarks to understand that. He even specifies tasks/apps like widgets, mail, internet, and ilife.

I, and many other posters here, have used enough different macs running OS X with different specs to understand what the OP means and to know that it's a RAM issue. Do you seriously disagree with that?

Yes, I suppose I'd say I seriously disagree. I have run OSX on many Macs with different specifications, enough to know that the blanket statement that 512 Mb "isn't enough for OSX" to be untrue. I am arguing against using "feel" over objective tests as a measure of performance. I am arguing against the universal panacea approach of adding more RAM whenever something performance-related, no matter how vaguely it's defined, is reported. I am against telling people that they absolutely must spend more money before I really know what is going on.

Once again, just because the bench tests don't confirm the conventional wisdom, doesn't mean that they should be ignored. They might be suggesting the need to look elsewhere for performance-related issues. Don't you think this is the more scientific approach to problem-solving?

 

[gloss]

I think it's perfectly scientific to assume that, since the machine has 512mb of RAM AND relies on said RAM for its onboard graphics, that graphics-based slowdown may well be caused by a shortage of memory.

Don't think too hard. Occam's Razor cuts cleanly for a reason.

People expect it to be faster than it is, and I have news for you, IT'S NOT GETTING ANY FASTER, IT'S A PIECE OF CRAP.

Funny that almost all the reviews mention the exceptional speed.

 

[miniConvert]

Upgrade the RAM. It isn't really even 512MB as the shared graphics processor is stealing resources.

You need two matched sticks to have the graphics run dual channel. You want at least 2x512MB sticks for 1GB, and preferably 2x1GB sticks for 2GB. Then you will really enjoy your Mini. It really is a RAM problem.

I have to agree. I ordered mine with 2GB RAM from Apple (yes, expensive, I know) and it absolutely flies. Having so much RAM also means I never need to completely quit anything I use often - things like Camino, Mail and iCal - meaning they're all instant.

There's always over 1GB of RAM in use when I check - so 512MB really isn't enough. I believe Apple should sell all their Macs with a minimum of 1GB RAM, but that, of course, has price implications and there'd be far more moans about Apple not having an inexpensive Mac than there are from people who buy low spec but then don't expect to see any beachballs.

 

[QCassidy352]

Once again, just because the bench tests don't confirm the conventional wisdom, doesn't mean that they should be ignored. They might be suggesting the need to look elsewhere for performance-related issues. Don't you think this is the more scientific approach to problem-solving?

Using artificial 3rd party benchmarks may be the more "scientific" (i.e. quantifiable) approach to problem-solving, but it's surely not a better approach. Apps like cinebench and xbench are good for exactly one thing - telling you how well a computer performs running cinebench or xbench. They mean so close to nothing in terms of real-world use that I consider them worthless. Cinebench or xbench scores simply do not accurately reflect the real-world experience of using a given computer system.

Let's face it - "feel" may be unscientific, but how a system's performance "feels" determines how happy we are with it most of the time. What good is a high cinebench score when using the machine is frustrating? And why would you care about a low cinebench score if your machine does everything you want in a timely manner? Again, I think these benchmarks don't tell us anything that matters about the performance of a computer.

And let me point you again to my first post. There's a very simple way to prove me right or wrong. The OP just needs to look in activity monitor and see his page outs. My guess is he has thousands. If I'm right, more RAM will fix it (surely you don't disagree with that?). If I'm wrong, I'll eat my hat.

I've used enough macs with insufficient RAM to know that my hat is safe.

 

[zap2]

Your crazy! I just got a Mac Mini Core Duo and its faster then my old iMac G5, and my G5 has 1Gb of RAM, and this Mini only has 512mbs! I have the 1GB Stick with me but i just have not yet put it in. With 1.25Gbs, this baby will fly!

 

[aristobrat]

Apps like cinebench and xbench are good for exactly one thing - telling you how well a computer performs running cinebench or xbench. They mean so close to nothing in terms of real-world use that I consider them worthless.

My thoughts exactly.

Although I appreciate IJ's thought that people shouldn't needlessly throw money into fixing a performance problem, I haven't seen a post yet where someone's added more memory to a 512MB machine and didn't "feel" a performance increase, be it an Apple machine, a Dell machine, an IBM ThinkPad, etc...

 

[IJ Reilly]

Using artificial 3rd party benchmarks may be the more "scientific" (i.e. quantifiable) approach to problem-solving, but it's surely not a better approach. Apps like cinebench and xbench are good for exactly one thing - telling you how well a computer performs running cinebench or xbench. They mean so close to nothing in terms of real-world use that I consider them worthless. Cinebench or xbench scores simply do not accurately reflect the real-world experience of using a given computer system.

Let's face it - "feel" may be unscientific, but how a system's performance "feels" determines how happy we are with it most of the time. What good is a high cinebench score when using the machine is frustrating? And why would you care about a low cinebench score if your machine does everything you want in a timely manner? Again, I think these benchmarks don't tell us anything that matters about the performance of a computer.

And let me point you again to my first post. There's a very simple way to prove me right or wrong. The OP just needs to look in activity monitor and see his page outs. My guess is he has thousands. If I'm right, more RAM will fix it (surely you don't disagree with that?). If I'm wrong, I'll eat my hat.

I've used enough macs with insufficient RAM to know that my hat is safe.

"Real world" is an intangible and subject to an infinity of variables. Not that bench testing is perfect by any means, but it is valuable for controlling variables, so that the results are comparable. Bench tests are designed to approach real-world use, but in a controlled fashion. They aren't totally artificial, and surely they aren't as subjective as "feel" or are they as removed from actual performance as counting page-outs.

I was as surprised as anyone to read the results of the testing. I would have expected some improvement in performance with the addition of RAM, if only marginally. The fact that this did not occur is perplexing, and I think worthy of further consideration -- not outright dismissal.