Slashdot Mirror
Seven Mobile ATA Hard Drives Compared
AnInkle writes "Though hard drives are allegedly the fastest advancing high-tech product, most laptop manufacturers persist in saving a buck by outfitting their units with a low-end, low-cache, low-capacity, low-spindle-speed HDD. The Tech Report takes a different angle from other mobile hard drive reviews by including one of those maligned 4,200 RPM, 2MB cache models in their roundup of 2.5" hard drives, which includes 'a 160 GB perpendicular monster and a couple of 7,200-RPM speed demons.' The results are clear that most of us would see a tremendous boost in performance by upgrading this one component."
tremendous boost in performance by upgrading this one component
If you think THATs suprising, imagine my face when I found out that FLAMMABLE and INFLAMMABLE mean the SAME THING.
"When life gives you lemons, don't make lemonade. Make life take the lemons back!" -- Cave Johnson
However, one day the included 6Gig harddisk with a really low speed (Must have been a 4200RPM, but could be less) and I bought a new 5400RPM 80Gig harddisk . That was pretty much the upgrade that gave me most speed. That, and I could finally install more than one OS and keep the machine usable ;-)
Fast harddisks do matter.... Even if I tought that it was one of the least important things in the overall speed of the machine.
Ahhh...the great dumpster continuum. Many a free computer will be found there. -- sowth (748135)
While for instance Moore's law still holds the hard drives are not developing as fast. Ca. 10% performance improvement per year in the previous years is a good estimate. That means the hard drives are actually the slowest advancing components. In modern higher end PC they are the slowest link and in many applications the most horrible bottle neck.
..), the fastest things are expensive or on development cycle anyways... The state of hard drives and their performance is simply put pathetic and will be at least for the next a few years.
There isn't a good solution available either. RAIDs can get expensive, flash and similars can be fast but there are problems with interfaces (quality, selections,
I see in their published specs that the 7200 RPM drives run at least 0.4 to 0.6 Watts higher. This may not seem like much, but right now my laptop is sucking about 17 watts of power, and that means about 2.4 to 3.5 percent higher power consumption.
Still not much, but a factor to consider.
XML is like violence. If it doesn't solve the problem, use more.
If you had enough experience with computers, or especially if you lost one or more hard drives with precious data, that hard drive stack on the first page will shorten your life with a couple of years.
It's enough for someone to fire up a match for me to panick and go check if it's not the HDD. Works every time.
most laptop manufacturers persist in saving a buck by outfitting their units with a low-end, low-cache, low-capacity, low-spindle-speed HDD.
Well, they're saving an average of 111 bucks in these examples. The "low-end" model is about 35% of the cost of the others (on average).
Now look at the performance differences. WorldBench is clocking the more expensive drives as only 30% faster (on average) than the "low-end" drive.
My own conclusion: yes, you're getting a performance boost if you pay more... But it's definately not a 1 to 1 ratio. In fact, for the money, the "low-end" drive is the best solution. So... Why do "most laptop manufacturers persist in saving a buck (or 111 bucks)? Because it's a better choice for the average consumer! Believe me... If Company A started selling only expensive drives, their market would go niche (like Alienware), and most people would purchase a "lower-end" machine.
When I fire up my laptop, its because I'm not at a outlet, or am moving somewhere. The harddrive is the largest battery pull on my old rig. I'm not looking for fast but smart. They should have large caches and 4200 rpm or maybe even lower. An old toshiba from the 80's I've had, used to run for a few hours with power management. I haven't seen that again from these new 'mobile super computers' that is flooding stores. SATA is one of the smart ideas, but they'll all be chasing the benchmark crown for speed.
I think I just cashed out all my cool points.
I have personally witnessed many mobile PCs speed up and get quieter after having the HDD swapped for a higher spec unit. In addition to the increased performance in read/write speed, many of the newer drives offer another benefit to mobile users: Lower power consumption. Upgrading an older drive to a new faster drive could increase battery life.
This space intentionally left blank.
Notebook hard drived (2.5 inch for factor) cost about three times as much as desktop drives, because they almost never seem to go on sale. I just had to buy a replacement for my notebook, and an 80GB/5400rpm/8M 2.5 inch drive cost me about $100. I can frequently get a 250GB/7200rpm/16M 3.5 inch drive on sale for a half to a third that price these days. With a $/GB ratio that high, cost is a significant factor in adopting higher capacitry drives on mobile platforms. A secondary issue is heat - some notebooks aren't designed to handle the additional heat load from the higher-end 7200rpm (or even 5400rpm) drives, shortening the drive life significantly.
Can You Say Linux? I Knew That You Could.
Sun opteron servers are shipping with Segate Savvio 2.5" 10K serial attach SCSI drives. Astonishingly expensive, and they suck about 8 watts each. Burn your thighs today . . .
S eagate+ST973401SS&btnG=Google+Search
http://www.google.com/search?ie=UTF-8&oe=UTF-8&q=
Give a man a fish and you have fed him for today. Teach a man to fish, and he'll say "WHERE'S MY FISH, YOU IDIOT?"
I want a raid in a harddrive form factor. So I can just plug it in like a new hardrive but if one disk fails it can still live.
Since when was 10-30% overall from worst to best performer regarded as tremendous? The impact of the disk subsystem is around 30% on daily tasks, and around 70-90% on disk-intensive tasks. So we're looking at a ' tremendous performance increase' of around 10% to 25% in the best case, only achievable by owning the worst performer, and thence upgrading to the best available technology.
From the article photos it seems they were primarily interested in comparing their width, height and depth, as well as trying to build various models using them as freakishly expensive LEGO blocks.
What I wonder is if after the experience, at least one of them is behaving odd or making funny noises on startup.
While it is nice to have fancy shmancy specd laptops to tote around, you can only put faster (read: more power / heat) devices in a laptop to a certain extent. There is a curve that follows along with an opposite one, which refers to efficiency / portability and the other to power / speed.
The other end of this discussion that I've not seen discussed yet is being mobile also presents real dangers to physical disks. Perhaps having a slower spindle speed is slightly less risky for those individuals who insist on slinging around a computer while it's powered on.
Gyroscopic forces probably hit those drives harder too, with thinner platters. Anybody who has held a bicycle tire in their hands while sombody else spun it, then tried to tilt it one way or the other knows exactly what I'm talking about.
Windows has more viruses because linux has more virus coders.
If the computer runs faster you may use less power. If it takes 10 minutes to check email and slashdot with the old drive and 9 minutes with the new, you've just saved 10% time spending 3.5% higher power rate. That's about 7% less power consumed.
Man, you really need that seminar!
Boosting over all speed by up to a quarter is not a tremendous advance? Heck, I'd consider a 10% raise in my paycheck to be pretty tremendous. If my paycheck went up by a quarter, I'd blow a gasket.
I feel the same about computing. If I could boost my laptop's overall performance by 10% by making a single change, I consider that pretty impressive.
It would make a lot of sense to have 10% of your disk solid state, only spin up the real drive as necessary. I don't think multigigabyte memory will be affordable anytime real soon.
Man, you really need that seminar!
When I had a 100 MHz PC, I had a 1 GB HDD. Now I have a 400 GB HDD and I don't have a 40 GHz processor. What are you talking about?
Send email from the afterlife! Write your e-will at Dead Man's Switch.
The other issue with using ultra fast spindle speed HDs in laptops is that it makes your laptop pull to the left.
It's not offtopic, dumbass. It's orthogonal.
The next generation of disks will be fuel-injected, and not subject to this problem.
Other poster is right, the drives are too tall to fit in a laptop.
Give a man a fish and you have fed him for today. Teach a man to fish, and he'll say "WHERE'S MY FISH, YOU IDIOT?"
most laptop manufacturers persist in saving a buck by outfitting their units with a low-end, low-cache, low-capacity, low-spindle-speed HDD
That's because rational consumers 'persist' in saving a buck by buying the least expensive thing they think will fill their needs.
Most people buying PCs have absolutely no idea how to compare one computer to another. Even most Jeff K's understand nothing beyond screen dimensions and clock speed (and I've worked with enough IT people toto understand that Jeff K is the rule, not the exception). Of course, even the bottom of the line $650 Dell XPS comes with a 7200 RPM 8MB Cache HD, so I'm not sure what kind of poor sucker is still getting the 4200 RPM dog described in the article.
Swapping is the bottleneck. So remove the bottleneck. No seriously. The harddisk activity you are most likely to notice is memory swapping. Swapping can be disabled. Of course you run out of memory if you do that, so add more memory. I find that with 2GB no application ever complains of having not enough memory despite there being exactly 0MB of swap space. I run some pretty memory intensive stuff too. It turns out most of this stuff is designed to run well on systems with only 512-1024 MB (particularly games rarely use more, even if it is avaialable). That extra GB is cheaper than a new harddrive and if 2 is not enough make it 3 or 4. It's not like win32 processes can address more than 2GB anyway!
At least under windows, memory swapping is implemented very stupidly. Basically the system will spend (your) time swapping even when there's plenty of memory available. I've observed it swapping applications to disk with over 75% memory available. This causes all sorts of noticable delays when you try to actually use your system (e.g. switching from application A to application B). With 2GB available, windows should run out of excuses to swap but it will still swap.
Disabling swap space effectively stops this behavior. Especially on slow harddisks this means a huge performance improvement. Depending on your software you can do with much less memory. I've disabled swap space on machines with only 512MB which you are unlikely to exceed running just office type applications. In all cases that I did this the result was an immediate, noticable performance increase.
In case you do run out of memory, you get an out of memory error. I find that closing applications usually is a good solution. Much better than windows continuously wasting my time with unnecessary UI blocking harddisk activity. Anyway, given the low cost of memory, I'm very intolerant towards having my time wasted due to the fact that there's not enough.
Jilles
I hold that floopy drives are the slowest advancing technology. they started out at 360 KB some time in the eighties, and are currently at 1.44 MB.
[\dry wit]
Are they really? I'm on an iBook G4 w/768Mb of memory. I hardly ever shut the machine off; it does sleep mode perfectly, so there's never really a need to reboot. When I'm working, it's on only a few applications at a time; the second or two it requires to load a seldom-used application is so rare that I honestly can't see how a faster hard drive would provide me much benefit at all.
In all honesty, the slowest thing about my computer is me. Even if an app were to load instantaneously, my brain is still gonna spend a few seconds getting its shit together to actually use the application, let alone do anything truly useful with it.
--
Don't like it? Respond with words, not karma.
One thing's for sure. It ain't Lithium-Ion battery life.
Any technology distinguishable from magic is insufficiently advanced.
Quite frankly, I feel that 30% understates how slow some laptops feel because of their slow hard drives. I don't really consider myself to be terribly impatient, but a fast hard drive is the best reason I can think of to use a desktop whenever possible. However much as these drives bother me, they do have thier merits. The 4200 RPM drives consume noticeably less power, and are much cheaper, though capacity still has more sway on the bottom line.
There's a nice middle ground for laptops, I think. When my 4200 RPM drive in my laptop died 6 months back, I replaced with a 40 GB 5400 RPM drive with a 16 MB cache. It does reduce my battery power slightly, but the faster spindle and enormous cache make it worth the loss for me. At the time, it was one of the cheapest drives listed on Newegg at $70 shipped. Considering that even the cheapest 20 GB 4200 RPM/2 MB cache drives are $60 shipped, I'd call what I got a good deal.
"I do a grep for shit, bollocks, and tits before checking in code. I'm professional..." -RECURSIVE_META_JOKE, reddit.com
I never thought of it that way before, but consider:
I can name a couple of PC brands like Alienware and Voodoo, but I can't really think of anyone who is "known" for making similarly high-end laptops.
I realize that's because most people need them for mundane, business-related tasks, but with so many (myself included) using laptops at work as desktop replacements, you'd think the specs would matter more.
Our company just rolled out "new" laptops for the staff: IBM/Lenovo T-41 ThinkPads, which were first introduced almost two years ago! That's "good enough" for a laptop around here, and I'm sure that's the consensus at the majority of offices.
Any technology distinguishable from magic is insufficiently advanced.
Dr. Nick: 'Inflammable' means flammable? What a country.w w.imdb.com/Quotes%3F0096697&e=9797
http://www.google.com/url?sa=U&start=2&q=http://w
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
I have two hard drives in my Thinkpad. Have to go without internal CD drive, but still interesting. Particularly with external USB CD drives (even ones that draw all the power they need via USB).
XML is like violence. If it doesn't solve the problem, use more.
The cheap flash memory (like in my 1GB thumbdrive) has ~100,000 r/w cycles. If your internet cache was there, you'd hose that memory within hours.
The cheap flash memory in thumbdrives is not the only flash memory around. 100,000 cycles nowdays for flash would be considered bad. I'm seeing 1,000,000 refered to now for decent CF.
It would make a lot of sense to have 10% of your disk solid state, only spin up the real drive as necessary. I don't think multigigabyte memory will be affordable anytime real soon.
Multi-gigabyte memory is affordable NOW. CF cards can be used as IDE hard drives and 2GB units are cheap (even 4GB looks cheap to me). I've been using CF for firewalls (OpenBSD/pf) for more than a year now without any troubles. I just switch on soft updates and use noatime.
War crimes, torture, lies, illegal spying... Would someone give Bush a blowjob, already, so he can be impeached?
There is plenty of trade offs to be made on hard-drives.
Size, weight, power consumption, price, capacity, rotation speed, platter density, and ruggedness.
For mobile harddrives, the number of platterns is verylimited by size, and platter size itself is limited.
So they never get anywhere near the capacity of its desktop equivalent. However, less and smaller platters means also less power consumption so its still another point why its so in mobile computers.
The platter density helps capacity, and bandwith and thats what harddrive manufacturers are trying to improve.
Rotational speed costs more weight, power and in some cases hurts platter density.
The improvement of bandwith, can be done either improving density or rotation speed.
Improving seek times can be done either by using more power for faster seek head, or using smaller platters. Or better electronics.
In overall, if you would allow desktop powerconsumption size and weight for your harddrive you would get desktop capacity and performance out of your laptop harddrive. However if you sacrifice some capacity and performance you can cut down both the active and idle power consumption by order of magnitude which is more important than having desktop performance and capacities for most laptop users.
Emacs is good operating system, but it has one flaw: Its text editor could be better.
In my day, a hard drive was going across the country in a car with kids in the back whining, "Are we there yet?" We didn't have it easy like you whipper-snappers with your SATA and your RAID! Hell, Raid was for killing bugs.
The Uncoveror: It's the real news.
I was referring to replacing an 80GB drive, which isn't currently feasable. If you used a few GB for the OS and frequently used files you could get away with running the harddrive very little, which makes sense.
Man, you really need that seminar!
As an anecdote, my 4200 RPM laptop drive is much faster than my 7200 RPM desktop drive, both using ATA 100 and disk-bound applications. I think one reason is that the laptop drive has more cache. But there are loads of other factors. Platter size is one interesting point where bigger doesn't always mean faster, especially with regards to seek times.
BTW, the article's point about platter mass is pretty moot IMHO, because it only affects spin-up energy expenditure. Even in a laptop-style usage where the HD turns off now and then, the effect is negligible. The important thing in my experience is the faster spin-up time of laptop drives. I also happen to like quiet computers, and it's one reason why somebody might choose lower RPM in a hard drive.
Escher was the first MC and Giger invented the HR department.
The speed increase from a low end to high end is rarely more than 30% but the price is often 2-3X higher.
By the way combining a high end CPU with a low end Diskdrive is pretty much maximum folly but this is the way they are configured. Resason: People do not know anything about the disks but "know" they "need" the faster Intel chip to make TehIntanet go faster
Help fight continental drift.
HDs were exploding in size in the late 90s becaues of the replacement of coil read-heads by thing film heads based on quantium magnetic resitance effects (sounds better than just GMR heads).
That ended quite a while ago.
We were stuck at 250GB for _years_.
Pertenticular recording could give another boost the next years, maybe an order of magnitude, yes, but the last 5 years HD were very slow in progress.
HI O WISE PRINCE. WHT TOOK U SO DAM LONG?
Give it 2-5 years, and it will happen I think.
1 80090
http://slashdot.org/comments.pl?sid=183698&cid=15
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
I tend to go for the large capacity drives. Back in 2000 I picked up a 23gb behemoth for my powerbook. At the time it made it the fastest computer in the building where I worked, with more storage space than the fileservers. Bragging rights if nothing else. I always find myself saying "THIS will be the last word in hard drive space, I will never fill this". Then a year later I'm out of space and checking out the latest tech. I doubt this will ever change, and I would not be surprised in the least to have a 1tb laptop drive sometime in the next 10 years.
Last month I upgraded the hard drive in my new powerbook, from a cramped 80gb to a cavernous 160gb. I could really care less about the hard drive speed, which was 4200rpm iirc. What matters to me is (1) how much free space will I have when the smoke clears, and (2) what's the warranty. I take it for granted that this seagate, with its comforting 5 year warranty, WILL DIE sometime before it's 5 years is up. Then I'll get a free replacement. The one that's in the laptop now is backed up weekly in preparation for the disaster that will come. Since this drive has only been out for a few months I don't expect them to have had the bugs worked out so it will probably last two years at most I figure.
I find it really bad news that the standard hard drive warranty has slipped from 3 years down to 1 year on many drives. Now some of those companies that used to ship with 3 yr standard warranty are now charging the same for their 1 year warranties, and then you can choose to buy thier "gold" (or whatever) line that has the 3 year warranty on it, for an extra charge of course. Seagate still sets the bar at 5 years for most of their drives, tho their MTBF seems to be slipping so I expect they are paying a lot for replacement drives. At five years, I figure by the time the warranty runs out I'll be shopping for a new, larger drive anyway.
I work for the Department of Redundancy Department.
It seems no one here is able to stick to one interpretation for Moores Law.
The original law related to increased transistor count for money per year. Here we have some people changing it to Mhz increase per year. Then they want to compare it to hard disk transfer speed increase per year while others are saying storage capacity (or density) per disk per year is more appropriate.
No wonder there is so little agreement.
Good post.
It doesnt make any mention of what Hard disks will do in the 2 to 5 year period.
You also chose the worst $per meg disk on the market.
Otherwise. I like your analysis.
Skimming the article one thing is very clear: the Seagate Momentus 5400.3 is by far the best drive for any use where noise and heat matter. It's results are very impressive.
Note that it is the only perpendicular drive in the round-up and that the article says it is the most expensive.
Hard disks will outpace NAND flash, both in terms capacity and $/GB. So you can choose between a 32GB NAND flash disk or a much bigger hard disk.
But there's a crossover capacity between flash and a rotating media. E.g. at the moment I'd use flash for upto 2-4GB. You can see it with MP3 players - small capacity ones tend to be flash because it's cheaper for low capacities. High cap ones tend to be hard disk based. The crossover capacity should double each 1.5 years with Moore's law, I suspect.
But I'm working on the assumption that for most users 32GB or so is enough. An OS and Office suite will fit in a fraction of that even in five years' time. So ultraportables should switch over. Unless some other, cheaper solid state memory becomes possible of course, but that would just accelerate the switch away from hard disks.
echo -e 'global _start\n _start:\n mov eax, 2\n int 80h\n jmp _start' > a.asm; nasm a.asm -f elf; ld a.o -o a;
"Hard disks will outpace NAND flash, both in terms capacity and $/GB.br>
My research shows this hasnt been the case for the last 3 years and I have reason to think the next 3 wont be any different.
I always thought the reasons flash was successful in small capacity MP3 players was that flash is much smaller than even the smallest hard disks, it uses less power thus reducing the size of the battery needed and has higher G shock tollerance. I guess another reason they are used is that hard disks lose their economic advantage when made in very small capacities, which is what you are saying.