I don't want to reflexively say "No problems here!" as though that has anything to do with the problems others had, but yeah, in my caase I really didn't have problems.
I also had a few things start working that had been broken, probably because Core 2 Duos and the associated chipsets and motherboards are so new, notably the clock speed controls.
My network ports got flipped around (eth1 and eth0 got mapped onto different hardware).
IMO, you shouldn't have to submit a bug to be able to complain. Writing a good bug report is a fair amount of work, and if you're expected to do it whenever the OS whenever the OS has issues, then that OS is suddenly a lot of extra work to use.
It sounds like you're giving a very lengthy explanation of why one should agree with Apple's lineup decisions, which doesn't really contradict the post you replied to, which said something very similar with a much more negative spin.
But, you've also said stuff that's not true. Most importantly, it's not a matter of not wanting to risk bleeding edge stuff. That's crap, and if it were true they would not, for example, have jumped on the Core Duo bandwagon so early. Also, the resolutions we're talking about now have been economical for several years, the lower resolution of Apple laptops has been a persistent problem for several years now.
I also don't think it has anything to do with keeping the order page simple. They've got selectable memory, hard drives, optical drives, various addons, etc. Would it really complicate things that much to give a few different options on a major component?
The only arguments that hold water are that it makes Apple's inventory easier, and that it's a good DPI to use for some applications.
However, I don't care that it makes Apple's inventory easier. That's not my problem. My problem is getting a computer that meets my needs.
Nor do I care that it's a good DPI for someone that's not me. I need high resolution to fit lots of code, documentation, and terminal windows on the screen. Many Apple fans assume that choosing the best compromise is the right answer, but that's wrong. Allowing the user to choose their favorite is the answer.
The article might call it unlikely, but I actually get pretty good results with a Mach 3. I never really had a problem with cuts, but I deffinitely like the results better WRT how my face feels after I shave.
It's great for performance, because having files that exist all over the disk is a lot better than having individual files get fragmented. They're called Extents, and MacOS uses them.
"So, I wouldn't discount the 5400RPM drive yet, especially taking into account the lower heat/power of a slower rotational speed."
Well, it's appropriate in at least some cases due to the higher capacity, but the point is that Apple isn't allowing the user to make their own decision based on their needs.
Even if sustained transfer speeds are higher, that's not the only metric by which hard drive performance is measured (to expand on your mhz myth thing). For example, a 10k rpm Raptor's throughput is roughly equal to a 7200.10 Barracuda (uses perpendicular recording). So, can we arrive at a conclusion at which is better with that information?
No, because it depends on what you want to do with it. The Raptor has a significantly seek time, and the Barracuda has more capacity; the two are mutually exclusive. Without knowing what it's going to be used for, it's impossible to say which is better.
Different priorities result in different, mutually exclusive, hardware decisions. The only correct decision is to offer both so users can pick the one they need, and Apple hasn't done that.
"I agree. In fact, let's forget about the technology and focus on the actual information density, since that's what counts. For the same total capacity, a laptop drive has higher density than a desktop drive, so it should achieve equal performance with slower RPM"
This is incorrect because:
a) It ignores seeky workloads, where rotational latency is important. b) It ignores the smaller diameter of the media, meaning the same RPM results in lower transfer speeds. Laptop drives don't achieve comparable transfer rates to desktop drives (which are over 70 mb/s sustained now). c) A higher RPM drive with the same density increases would also increase in performance, retaining its lead. Are you assuming 7200 rpm laptop drives don't get perpendicular recording technology?
The technology has improved, but the same design tradeoffs still apply. Someone willing to sacrifice power and capacity for RPMs will get better performance. Apple should expose this decision to the user, since some will want one and some will want the other. Just because 5400 rpm laptop drives can match yesterday's 7200 rpm laptop drives by some metrics doesn't allow you to say "Well, technology is over. This is the point I was waiting for and improvements beyond this are useless."
Hardware is always too slow, and improvements are always welcome.
"Of course, you have to remember that larger disks have higher linear velocities for the same RPM? So why don't the pissing contestants use 5.25'' drives? I think it has something to do with seeking across cylinders, which is another point where laptop drives rule."
This is incorrect because it assumes laptop drives are optimized to spend the same amount of energy seeking as their desktop counterparts, which is incorrect. Something like, say, a 2.5" 10k RPM SAS drive has a good seek time, but that's a completely different animal than a laptop drive.
5400 laptop drive seek times are aroudn 12-13 ms, 7200 rpm laptop drive seek times are around 10-11 ms, and 7200 rpm desktop drives are around 8-10 ms. That's deffinitely a reason for wanting a a 7200 rpm laptop drive.
"The issue is heat. The MBPs get quite hot, and a 7200RPM drive generates a fair bit more heat than a 5400RPM one. I opted for the 4200RPM in my PowerBook to try to keep heat down."
If the issue is heat, then the issue is Apple's design. Impairing performance to make the machine a fraction of an inch thinner doesn't make sense if the machine is supposed to be a performance-oriented one.
"For anyone doing video editing, I would recommend the LaCie triple interface drive range; 7200RPM drives on a FireWire 800 chain. If you really want performance, buy two and put scratch files on one and source files on another. You've be surprised at the difference it makes."
In fact I use a similar strategy on my desktop, and the difference is huge. The problem is that since these are laptops, external scratch drives aren't always an option. It would substantially reduce the battery life, and I doubt it could power multiple external desktop drives from battery.
"it's not something worth worrying over. the newer drives have much higher data density so they can do sustained read/writes as fast or faster than a 7200rpm drive. the rpm of a drive isn't an ultimate measure of how fast the drive will work on your computer."
And are we to assume that perpendicular recording will make its way into 10k/15k rpm SAS drives, 7200 rpm SATA desktop drives, and 5400 rpm notebook drives, but be left out of 7200 rpm notebook drives?
Besides, this ignores the reduced seeking latency of 7200 rpm drives.
"Seek times over the surface of the disk won't, but the average seek distance will dimminish and the likelihood that data will be found within the track or cylinder will increase. These effects reduce average seek times when you compare, say, the first 100GB of a 160GB drive to the entire 100GB of the 7200rpm option. That is the proper way to look at it since the OS, applications, and user data don't get bigger when the drive gets larger."
You've got it backwards. Your argument only holds if the filesystem allocates space at the start of the drive and works its way to the end without leaving gaps when it fact it is optimized to do precisely the opposite. Files are dispersed across the whole drive in order to give them room to grow without fragmenting them.
The only way your I/O can be guaranteed to remain within the first X percent of the disk is if you reduce the size of your partition accordingly, and this has other performance problems, since files will begin to fragment if the partition gets too full and this will happen sooner.
MacOS contains optimizations to put certain parts of the OS closer to the beginning of the drive, but these also benefit from higher RPMs, since the rotational latency is reduced.
"Somewhere else, somebody mentioned that the 160GB drives use a different (more dense) magnetic encoding, such that their read capacity might actually be higher than before (although the seek time would remain the same). Might upgrading to the larger drive make a compromise?"
This depends entirely on how seeky your workload is. And, I haven't been keeping up with laptop hard drives, but the 7200s either have already or will soon get higher densities.
"Wrong. The iMac uses the Core 2 Duo. Core 2 IS Conroe. The newest revision of the iMac is no more using laptop hardware."
This is incorrect. Both Intel's newest laptop chips and desktop chips are marketted as "Core 2 Duo". The iMac uses the laptop version, as evidenced by the 667 mhz bus. The "Conroe" desktop version of the chip has a 1066 mhz bus.
Had I not already posted in this discussion, I would mod you +1 Gracious.:)
"I don't know why, but its a bit sad - especially on the Pro models - not because its critical in itself, but because its so much more of a pain to change the HDD on the Pro than it is on the MacBook. And there is no reason for Apple not to offer this - after all, its a PRO laptop."
Indeed. This is especially true when you consider how many people rely on Macs for creative (sound/graphics/video) work that deals with huge files, and in many cases seeky workloads that stand to benefit the most from faster drives. It's difficult for me to understand the reasoning behind this decision.
When they lacked 7200 rpm drives in PowerBooks it was one thing, but to remove them after they've been added, after they've acknowledge the need makes no sense.
"You can get a 7200 rpm drive, its not a standard but its an option. But you have to drop down to 100GB. I think this reflects the manufacturers.
Apple no longer offers any 7200 rpm drive in the 15" MacBook Pros, at any capacity. It's not standard, and it's not offered as an option. The only place it's still available is in the 17" model.
Anyone who doesn't believe me is invited to check the Apple store.
iMacs use the laptop version of the chip, and Mac Pros use the server version. This leaves a pretty big gap for people in the market for something in the middle.
Unfortunately, they eliminated the option for a 7200 rpm drive. It's a significant performance hit, if you're doing something that's I/O bound on the hard drive.
The primary reason was the hardware. I don't mean this in the sense that I have a particular affinity for homebuilds, or that there aren't any other reasons. The cost was simply prohibitive with Apple, and this was big enough to cut short any consideration I might have otherwise given to the platform, regardless of merit.
I needed a workstation, but I have no use for a quad-core machine, so a Core 2 Duo or Athlon64 could easily meet my needs. I also needed a large RAID array and a scratch disk, as well as other things like multiple ethernet ports, PCI/PCI-E slots, and so forth. With Apple hardware, the only way to get what I want is to spend large amounts of money on stuff that won't benefit me (like that extra Xeon). When I tried to price out a Mac Pro to meet the same requirements it couldn't be done without more than doubling the price. Even if I were willing to go around upgrading the thing with cheaper 3rd party hard drives, RAM, etc, that stuff wouldn't be covered by Apple's warranty, and that's a big downside for me. Even then, it would still cost thousands more, and it wouldn't even be that much easier than a homebuild when all was said and done.
A secondary reason was that I've had an iBook up until recently, and getting the various *nix software I need was significantly more annoying there. A good distro's package manager will have many times the selection of the Mac alternatives such as Fink and Darwin Ports, and the time I spent compiling the missing stuff by hand on MacOS was significant. This easily overwhelms any savings of effort that I might have gotten from MacOS initially, and that's not even that much with easy distros like Ubuntu. I'm not a rabid freedom fighter, I just know empirically it's a lot more trouble for me to use MacOS.
Another way this advantage applies is that the software I need comes almost entirely from one place. With MacOS, it was a mix of Fink, Darwin Ports, stuff I've compiled myself, various.sit and.dmg files downloaded from various websites (Apple, VersionTracker, etc), and so on. With Ubuntu, it's all available from a single interface. One front end handles all the installations, removals, and updates. Even proprietary things like video card drivers and Sun's Java are handled this way. This cuts way down on the time it takes me to get a system set up with all the various apps I need. Downloading something from VersionTracker isn't difficult, but doing that over and over again for dozens of different things takes a significant chunk of time. With Ubuntu, I've found that I don't need to do it all at once, because clicking a checkbox and clicking "apply" in Synaptic takes seconds -- installing an app is barely more difficult than lanching it, and making a list of things I need would be more trouble than installing them when I notice they're missing.
I've seen what Macs have to offer, and I don't think I'd be interested even if it didn't cost so much more to meet my needs.
Remember how long it took to get rid of NT4/98? Lots of people are still using 2k, and XP has been out longer than other desktop releases. XP is going to be around for a long time.
If the move to Vista is stretched out over a number of years, much of the cost will be absorbed by normal new hardware spending, and I don't see XP becomming rare until the next decade.
Slashdot Mirror
You can bet your ass I'd be bitching if this had been a colo box that was now unreachable.
I don't want to reflexively say "No problems here!" as though that has anything to do with the problems others had, but yeah, in my caase I really didn't have problems.
I also had a few things start working that had been broken, probably because Core 2 Duos and the associated chipsets and motherboards are so new, notably the clock speed controls.
My network ports got flipped around (eth1 and eth0 got mapped onto different hardware).
IMO, you shouldn't have to submit a bug to be able to complain. Writing a good bug report is a fair amount of work, and if you're expected to do it whenever the OS whenever the OS has issues, then that OS is suddenly a lot of extra work to use.
"/bin/sh" is different all over the place, it's just supposed to be POSIX, you shouldn't assume it's bash.
A better alternative is to use "/usr/bin/env bash". This will work even if the path to bash is not what you expect it to be.
It sounds like you're giving a very lengthy explanation of why one should agree with Apple's lineup decisions, which doesn't really contradict the post you replied to, which said something very similar with a much more negative spin.
But, you've also said stuff that's not true. Most importantly, it's not a matter of not wanting to risk bleeding edge stuff. That's crap, and if it were true they would not, for example, have jumped on the Core Duo bandwagon so early. Also, the resolutions we're talking about now have been economical for several years, the lower resolution of Apple laptops has been a persistent problem for several years now.
I also don't think it has anything to do with keeping the order page simple. They've got selectable memory, hard drives, optical drives, various addons, etc. Would it really complicate things that much to give a few different options on a major component?
The only arguments that hold water are that it makes Apple's inventory easier, and that it's a good DPI to use for some applications.
However, I don't care that it makes Apple's inventory easier. That's not my problem. My problem is getting a computer that meets my needs.
Nor do I care that it's a good DPI for someone that's not me. I need high resolution to fit lots of code, documentation, and terminal windows on the screen. Many Apple fans assume that choosing the best compromise is the right answer, but that's wrong. Allowing the user to choose their favorite is the answer.
The article might call it unlikely, but I actually get pretty good results with a Mach 3. I never really had a problem with cuts, but I deffinitely like the results better WRT how my face feels after I shave.
It's great for performance, because having files that exist all over the disk is a lot better than having individual files get fragmented. They're called Extents, and MacOS uses them.
All modern filesystems do it.
"So, I wouldn't discount the 5400RPM drive yet, especially taking into account the lower heat/power of a slower rotational speed."
Well, it's appropriate in at least some cases due to the higher capacity, but the point is that Apple isn't allowing the user to make their own decision based on their needs.
Even if sustained transfer speeds are higher, that's not the only metric by which hard drive performance is measured (to expand on your mhz myth thing). For example, a 10k rpm Raptor's throughput is roughly equal to a 7200.10 Barracuda (uses perpendicular recording). So, can we arrive at a conclusion at which is better with that information?
No, because it depends on what you want to do with it. The Raptor has a significantly seek time, and the Barracuda has more capacity; the two are mutually exclusive. Without knowing what it's going to be used for, it's impossible to say which is better.
Different priorities result in different, mutually exclusive, hardware decisions. The only correct decision is to offer both so users can pick the one they need, and Apple hasn't done that.
"I agree. In fact, let's forget about the technology and focus on the actual information density, since that's what counts. For the same total capacity, a laptop drive has higher density than a desktop drive, so it should achieve equal performance with slower RPM"
This is incorrect because:
a) It ignores seeky workloads, where rotational latency is important.
b) It ignores the smaller diameter of the media, meaning the same RPM results in lower transfer speeds. Laptop drives don't achieve comparable transfer rates to desktop drives (which are over 70 mb/s sustained now).
c) A higher RPM drive with the same density increases would also increase in performance, retaining its lead. Are you assuming 7200 rpm laptop drives don't get perpendicular recording technology?
The technology has improved, but the same design tradeoffs still apply. Someone willing to sacrifice power and capacity for RPMs will get better performance. Apple should expose this decision to the user, since some will want one and some will want the other. Just because 5400 rpm laptop drives can match yesterday's 7200 rpm laptop drives by some metrics doesn't allow you to say "Well, technology is over. This is the point I was waiting for and improvements beyond this are useless."
Hardware is always too slow, and improvements are always welcome.
"Of course, you have to remember that larger disks have higher linear velocities for the same RPM? So why don't the pissing contestants use 5.25'' drives? I think it has something to do with seeking across cylinders, which is another point where laptop drives rule."
This is incorrect because it assumes laptop drives are optimized to spend the same amount of energy seeking as their desktop counterparts, which is incorrect. Something like, say, a 2.5" 10k RPM SAS drive has a good seek time, but that's a completely different animal than a laptop drive.
5400 laptop drive seek times are aroudn 12-13 ms, 7200 rpm laptop drive seek times are around 10-11 ms, and 7200 rpm desktop drives are around 8-10 ms. That's deffinitely a reason for wanting a a 7200 rpm laptop drive.
"The issue is heat. The MBPs get quite hot, and a 7200RPM drive generates a fair bit more heat than a 5400RPM one. I opted for the 4200RPM in my PowerBook to try to keep heat down."
If the issue is heat, then the issue is Apple's design. Impairing performance to make the machine a fraction of an inch thinner doesn't make sense if the machine is supposed to be a performance-oriented one.
"For anyone doing video editing, I would recommend the LaCie triple interface drive range; 7200RPM drives on a FireWire 800 chain. If you really want performance, buy two and put scratch files on one and source files on another. You've be surprised at the difference it makes."
In fact I use a similar strategy on my desktop, and the difference is huge. The problem is that since these are laptops, external scratch drives aren't always an option. It would substantially reduce the battery life, and I doubt it could power multiple external desktop drives from battery.
"it's not something worth worrying over. the newer drives have much higher data density so they can do sustained read/writes as fast or faster than a 7200rpm drive. the rpm of a drive isn't an ultimate measure of how fast the drive will work on your computer."
And are we to assume that perpendicular recording will make its way into 10k/15k rpm SAS drives, 7200 rpm SATA desktop drives, and 5400 rpm notebook drives, but be left out of 7200 rpm notebook drives?
Besides, this ignores the reduced seeking latency of 7200 rpm drives.
"Seek times over the surface of the disk won't, but the average seek distance will dimminish and the likelihood that data will be found within the track or cylinder will increase. These effects reduce average seek times when you compare, say, the first 100GB of a 160GB drive to the entire 100GB of the 7200rpm option. That is the proper way to look at it since the OS, applications, and user data don't get bigger when the drive gets larger."
You've got it backwards. Your argument only holds if the filesystem allocates space at the start of the drive and works its way to the end without leaving gaps when it fact it is optimized to do precisely the opposite. Files are dispersed across the whole drive in order to give them room to grow without fragmenting them.
The only way your I/O can be guaranteed to remain within the first X percent of the disk is if you reduce the size of your partition accordingly, and this has other performance problems, since files will begin to fragment if the partition gets too full and this will happen sooner.
MacOS contains optimizations to put certain parts of the OS closer to the beginning of the drive, but these also benefit from higher RPMs, since the rotational latency is reduced.
"Somewhere else, somebody mentioned that the 160GB drives use a different (more dense) magnetic encoding, such that their read capacity might actually be higher than before (although the seek time would remain the same). Might upgrading to the larger drive make a compromise?"
This depends entirely on how seeky your workload is. And, I haven't been keeping up with laptop hard drives, but the 7200s either have already or will soon get higher densities.
No, iMacs have the laptop Core 2 Duo as evidenced by the 667 mhz bus. The desktop version of Core 2 Duo uses a 1066 mhz bus.
"Wrong. The iMac uses the Core 2 Duo. Core 2 IS Conroe. The newest revision of the iMac is no more using laptop hardware."
This is incorrect. Both Intel's newest laptop chips and desktop chips are marketted as "Core 2 Duo". The iMac uses the laptop version, as evidenced by the 667 mhz bus. The "Conroe" desktop version of the chip has a 1066 mhz bus.
Had I not already posted in this discussion, I would mod you +1 Gracious. :)
"I don't know why, but its a bit sad - especially on the Pro models - not because its critical in itself, but because its so much more of a pain to change the HDD on the Pro than it is on the MacBook. And there is no reason for Apple not to offer this - after all, its a PRO laptop."
Indeed. This is especially true when you consider how many people rely on Macs for creative (sound/graphics/video) work that deals with huge files, and in many cases seeky workloads that stand to benefit the most from faster drives. It's difficult for me to understand the reasoning behind this decision.
When they lacked 7200 rpm drives in PowerBooks it was one thing, but to remove them after they've been added, after they've acknowledge the need makes no sense.
"You can get a 7200 rpm drive, its not a standard but its an option. But you have to drop down to 100GB. I think this reflects the manufacturers.
Apple no longer offers any 7200 rpm drive in the 15" MacBook Pros, at any capacity. It's not standard, and it's not offered as an option. The only place it's still available is in the 17" model.
Anyone who doesn't believe me is invited to check the Apple store.
I'm still waiting for a Conroe-based Mac.
iMacs use the laptop version of the chip, and Mac Pros use the server version. This leaves a pretty big gap for people in the market for something in the middle.
Unfortunately, they eliminated the option for a 7200 rpm drive. It's a significant performance hit, if you're doing something that's I/O bound on the hard drive.
Opera is better out of the box, but Firefox is better once you start installing extensions. Provided you need them anyway.
The primary reason was the hardware. I don't mean this in the sense that I have a particular affinity for homebuilds, or that there aren't any other reasons. The cost was simply prohibitive with Apple, and this was big enough to cut short any consideration I might have otherwise given to the platform, regardless of merit.
.sit and .dmg files downloaded from various websites (Apple, VersionTracker, etc), and so on. With Ubuntu, it's all available from a single interface. One front end handles all the installations, removals, and updates. Even proprietary things like video card drivers and Sun's Java are handled this way. This cuts way down on the time it takes me to get a system set up with all the various apps I need. Downloading something from VersionTracker isn't difficult, but doing that over and over again for dozens of different things takes a significant chunk of time. With Ubuntu, I've found that I don't need to do it all at once, because clicking a checkbox and clicking "apply" in Synaptic takes seconds -- installing an app is barely more difficult than lanching it, and making a list of things I need would be more trouble than installing them when I notice they're missing.
I needed a workstation, but I have no use for a quad-core machine, so a Core 2 Duo or Athlon64 could easily meet my needs. I also needed a large RAID array and a scratch disk, as well as other things like multiple ethernet ports, PCI/PCI-E slots, and so forth. With Apple hardware, the only way to get what I want is to spend large amounts of money on stuff that won't benefit me (like that extra Xeon). When I tried to price out a Mac Pro to meet the same requirements it couldn't be done without more than doubling the price. Even if I were willing to go around upgrading the thing with cheaper 3rd party hard drives, RAM, etc, that stuff wouldn't be covered by Apple's warranty, and that's a big downside for me. Even then, it would still cost thousands more, and it wouldn't even be that much easier than a homebuild when all was said and done.
A secondary reason was that I've had an iBook up until recently, and getting the various *nix software I need was significantly more annoying there. A good distro's package manager will have many times the selection of the Mac alternatives such as Fink and Darwin Ports, and the time I spent compiling the missing stuff by hand on MacOS was significant. This easily overwhelms any savings of effort that I might have gotten from MacOS initially, and that's not even that much with easy distros like Ubuntu. I'm not a rabid freedom fighter, I just know empirically it's a lot more trouble for me to use MacOS.
Another way this advantage applies is that the software I need comes almost entirely from one place. With MacOS, it was a mix of Fink, Darwin Ports, stuff I've compiled myself, various
I've seen what Macs have to offer, and I don't think I'd be interested even if it didn't cost so much more to meet my needs.
"Any filesystem with serious data loss on a power failure is not acceptable, period."
I seriously doubt that I would care if my squid proxy box lost the filesystem with the cache on it.
It is entirely application-dependant.
"So by unprotected, they mean some old installation without any recent patches, not a patched machine with no firewall."
They also mentioned attacks by worms that are irrelevant if you're not running stuff like (for example) an SQL server.
Remember how long it took to get rid of NT4/98? Lots of people are still using 2k, and XP has been out longer than other desktop releases. XP is going to be around for a long time.
If the move to Vista is stretched out over a number of years, much of the cost will be absorbed by normal new hardware spending, and I don't see XP becomming rare until the next decade.