Will size of hd make any difference?

  • Thread starter Thread starter bobbo
  • Start date Start date
B

bobbo

New member
I'm thinking of building a computer. I've been looking at these two drives.

IBM 40GB 7200rpm deskstar 60gxp ultra ata/100 8.5 average seek time

and

IBM 60GB 7200rpm deskstar 60gxp ultra ata/100 8.5 average seek time

The 60GB one is about $20 more than the 40GB one. Would I be better off getting the 40gb one for recording, and playing back a lot of tracks? Or would you think either one would be fine when it comes to speed?
 
All other things being equal, a bigger drive will always be faster than a smaller one.

Plus, given the superb price, the choice becomes even easier. I'd go with the 60GB.
 
my preference is a 20-30GB drive.....I also prefer Maxtor Diamond Max.......

I like the 20-30 range because it is more than enough to keep quite a few projects going at the same time....when i finish a project i move the files to either another drive or burn them to CD....so with only a few GB on the drive at one time, a whole lot more is wasted space....you can keep all those old files on there, but they only tamper with the performance...its more junk that the hard drive has to read through to get to what it needs......instead of a 40-60 GB, get a 20 and a 40...one for working files and one for backup......
 
wow Gidge, we posted at the same exact time. It was freaky.
Anyway, thanks for the reply, I'll have to think about this stuff, so I thought I'd start asking a few questions.
Oh, here's another question I think. What do you think would be the best setup for ide connections to a motherboard if I had:
2 hard drives (one 40 and one 20GB)
1 cdrw
1 cd-rom (or dvd/rom thing)

Would it be best to have all 4 drives on separate ide's if my motherboard supported that? Or does it really matter if one or two of them are slaved to another one.
Also, I notice some dvd/cdrom drives are about 20 or $30 more than a low cost cd-rom drive, what can I do with a dvd drive except watch a movie on the computer screen? I don't think I would want to do that, but is there any other reason anyone knows that would be of use to a dvd drive for the computer?
thanks
 
Whoa....

Greetings

Can someone elaborate on that comment..

"All other things being equal, a bigger drive will always be faster than a smaller one"

i can't quite get my head around that

info is faster read on the outside of a disk right?
so if you have a 40gb or a 80 gb there is no difference based on size necessarily? except the seek times based on the manufacturers.

That doesn't make sense to me...someone explain
SirRiff
 
Actually I believe that the contrary is true, when you are speaking along a singular line of drives. But I'll be damned if I can find that info.

But if by "bigger" he's saying higher Max. areal density (Gbits/sq. inch) then yes, bigger is faster, assuming the same rpm.

Queue
 
SirRiff:

Gidge's comment makes sense. If you don't plan to use much of the hard drive, then having a huge one would waste space. I personally wouldn't go with that, because my current projects take up roughly 10GB, and I'm not worried about the one or two extra simultaneous tracks I might gain from using the first 10% of the drive rather than the first 25%. You know?

It depends. Basically, we're both right. :) I'm a fan of IBM drives myself, but Maxtor also has a great reputation. Lately, Seagate and Western Digital seem to be the manufacturers who are taking a beating.

Let me see if I can do a better job of explaining myself. Seek time, cache, and interface are examples of stats that hard drive manufacturers love to mention, because they look good. But the fact is, they have very little effect on a hard drive's performance.

By the way, anybody who wants to get REALLY in depth with hard drive technology, read the hard drive guide at http://www.storagereview.com. It's several hundred pages, but you'll be a hard drive expert when you're through with it. :)

Cache memory on a hard drive rarely comes into play. Most hard drives have 2MB of cache. If you have a 40GB hard drive, that is equal to .000005% of your total space. Given that, how often do you think that cache memory is hit?

Pretty rarely.

Access time, also not very important. Especially when you're dealing with audio. In that case, it means nothing because all you need to care about is sustained transfer rates.

The interface (such as ATA66 or ATA100) is also not terribly important, but there are some cases where it comes into play.

I'll give you some examples...

For argument's sake, let's just assume that the new Western Digital 120GB 7200RPM drive with 8MB cache is the fastest IDE hard drive in the world.

...Ah, okay. And looking over the benchmarks, it apparently is. At the very beginning of the media, it delivers a peak performance of "only" 49.3 MB/second (read, not write). This doesn't even approach the bandwidth offered by ATA66, let alone ATA100.

Now IF you happen to get a cache hit, then it will transfer at the full speed of the interface. So you'll get all eight of those megabytes to your CPU almost instantly. ;-)

It can write a maximum of 24.2 MB/sec.

So why even bother with ATA100? Well, if you have two of those puppies on the same IDE controller, that's when the extra bandwidth really comes in handy.

Consider RAID. IDE RAID controllers use standard ATA100 interfaces. Two Seagate ST320430A drives in a RAID 0 configuration (using a Promise controller card) deliver a maximum read performance of a whopping 71.3 MB/sec.

But in determining the speed of the drive itself, the interface means absolutely nothing. So it's an important thing to understand.

What DOES determine the speed of a drive? Two areas determine the speed, more than anything else: Rotation speed and areal density.

Everybody knows what rotation speed means, so I'll just skip that.

Areal density. (I'm not about to talk down to you, just read on... I'm not trying to insult your intelligence.) Why do you suppose that 4-5 years ago we were all thinking our new 5GB hard drives were awesome, but nowadays high capacity hard drives are 120GB or more - yet, they haven't grown in size at all? The platters are the same size as before, but new technology, such as using different coatings for the platters, has allowed manufacturers to stuff more and more data into ever shrinking spaces. I didn't explain this concept well enough when I simply stated that bigger hard drives are faster - bigger hard drives are faster IF THEY USE THE SAME NUMBER OF PLATTERS. That's where the "all things being equal" remark comes in. Why is this? Consider a 20GB hard drive that spins at 7200RPM, and has two platters. We'll just say that only one side of each platter is used. Therefore there are two platters with 10GB of data storage on each. Now, imagine that there's another drive - 40GB, 7200RPM, two platters. That's 20GB per platter.

Therefore, in the same amount of time, twice as much data will pass under the head of the 40GB drive, than will pass under the head of the 20GB drive. Does this translate to a drive that's twice as fast? Unfortunately, no. Things such as error correction come into play, cutting the speed of the drive down. But, the 40GB drive WILL be faster.

So, yes. All other things being equal, a bigger drive will be faster than a smaller one.

Did I do an okay job of explaining that?
 
By the way...

I don't have benchmarks for the 60GXP at my fingertips, but here's the performance of the IBM 75GXP.

As with any hard drive, it is fastest at the beginning of the media, and slowly dips until it gets to the end.

Read Performance:
MAXIMUM: About 37MB/sec
MINIMUM: About 17MB/sec

Write Performance:
MAXIMUM: About 18-20MB/sec
MINIMUM: About 8MB/sec
 
Its funny...

Greetings,

Its funny but even after a detailed explanation I still haven't grasped it. Actually, I get everything until the last paragraph

Even if there twice as much data under the head, both drives could still transfer the same amount, say 20 megs/s. if you have a 11 meg file on each hard drive (Assuming its in one big chunk) both would transfer it in one long read, and it would be the same? No?

I could see if one had to move more then another.... like the people who partition their drives on the outer edge so all that audio data gets the best transfer rate.

But if the REAL bottleneck is not the capacity of the channel (ATA10 or such) what is it? Each individual hard drives transfer rates right?

Wait...brainstorm (in mid-post)

Is the transfer rate capacity based on actually quantity of DATA transfer? Or the "flow" of data off the disk?

Cause if it was "flow" (I don't know what to call it) then I could see that double density=double flow or something like that...

Hmmm...
Basically it’s the "flow rate" of data off the disk that increases with higher density not the transfer capacity of data through the hard drive?

Similar terms if know, but if think I see a qualitative difference.

THEN its not the size of the hard drive that gives an increase in transfer rates, but the density, which in modern times is strongly correlated with size since hard drive specifications must remain standard!

Do I have it or no?
PS- if it is no, I will give up all forms of music and technology.

SirRiff
 
SirRiff,

Even if there twice as much data under the head, both drives could still transfer the same amount, say 20 megs/s. if you have a 11 meg file on each hard drive (Assuming its in one big chunk) both would transfer it in one long read, and it would be the same? No?

I'm not sure that I fully understand what you're trying to ask. Let's just say the file is bigger than 11 megs, because fast drives will appear to transfer a file that small almost instantly, anyway... say you've got an 100 meg file. And you want to make a copy of it, and write that copy to a new directory on the same drive. Now, going back to our hypothetical 40GB drive, that drive actually needs half as much physical area on the disk platter to store the data (vs. the 20GB drive), and again half as much physical area to write a new copy of the file. So it only stands to reason that the 40GB drive is going to do it faster, since both spin at the same speed. Right?

I'll move on... your other questions might help clarify this a bit more.

But if the REAL bottleneck is not the capacity of the channel (ATA10 or such) what is it? Each individual hard drives transfer rates right?

That's absolutely right. Motherboards are extremely fast nowadays - you can really push data across them at a high rate. And with that, theoretical IDE transfer rates have also increased. I think that ATA133 will begin to pop up next year.

Hard drives have gotten faster too, but not quite as dramatically as the other key elements of your system, such as RAM, the system bus, and the CPU. In fact, you'd be surprised how little hard drive technology has changed since the '80s. Mainly, the platters spin faster, and areal density has increased.

The hard disk, in fact, is usually the bottleneck for your entire system. That's why I have to laugh a bit when I see OEM computer systems that have your usual 1.5GHz processor... 512MB RAM... and a 5400RPM hard drive. That's a waste of power - because the slowest components of your system will always be the mechanical parts, rather than the electronic parts. Which makes perfect sense, right?

Is the transfer rate capacity based on actually quantity of DATA transfer? Or the "flow" of data off the disk?

This part, I don't fully understand... maybe I wasn't too clear when I mentioned transfer rate. This is actually the rate at which the hard drive is able to push information to and from the system bus (and the CPU, I'm assuming). That's not including the rate at which the information flows through the drive itself, because just like the system bus, that part's fast.

See, the actual act of taking a magnetic head and reading information off of a platter is physical. That's why it's so slow compared to everything else - once the data has been read from the platter, and it's transfered through the drive's electronics, through the data cable, to the motherboard, and to the CPU, that's all electronic. It's a huge difference.

That's why solid state drives - "hard drives" that are actually made up of lots of chips (flash RAM, I'm thinking, but I'm not sure) are sometimes found in server applications. With no mechanical parts that need to move around, the speed and availability of the drive becomes incredibly fast.

THEN its not the size of the hard drive that gives an increase in transfer rates, but the density, which in modern times is strongly correlated with size since hard drive specifications must remain standard!

Bingo! You're exactly right. It isn't just the fact that the drive is bigger (in terms of data area), it's that the drive fits more data into the same space. Exactly.

And good job, picking up on the fact that standard specifications are important. Mac users seem a bit more flexible, but us IBM-Compatible users are HIGHLY resistant to change. So many of the things we use in our computers today are not really DIFFERENT from what was in our computers 15 years ago. They're just a lot BETTER.

When hard drive technology goes through a truly revolutionary change, then you'll see computer speed increase dramatically. I guess the reason why hard drives of today aren't all that different from hard drives of yesteryear is, they're just so WELL made. Hard drives are so much bigger, faster, and more reliable, while being so much less expensive than they used to be, there isn't a lot of reason to try anything really different yet.

And the last thing I wanted to touch on was the main reason why a hard disk with twice as much areal density won't necessarily be twice as fast. It's because of the technology of drive heads. They've gotten better over the years, just like everything else - but they're not capable of reading data flawlessly with just one pass.

Imagine a book. If you read it at your normal speed, you're probably not going to have to go back and reread things. But unless you're a speed reader, if you take your finger and start moving it quickly across each line, and you try to follow your finger with your eye, you may have to read over a lot of lines more than once before you pick every word up.

Same with hard drives. Sometimes a head has to pass over information more than once before it can pick all of it up. There are some really complex error correction algorhythms to try and correct this, but it isn't perfect.

Are you feeling a bit more clear on this? It sounds to me like you're getting it pretty well.
 
Wow

Greetings,


Hey Eurythmic, I think i got it now (But I won't be posting on storagereivew.com anytime soon...)

I never thought of it that way. I always knew that modern hard drives were much more dense, but i always figured maximum trasfer rates had increased giving the faster access speeds.


I was looking at those Western Digital 100BB SE and 120BB. Pretty impessive stuff...i don't think i would ever have to delete a file for the next decade if i had one of those...its funny because an 8M cache would have been a significent addition to my computers ram a year ago..


Thanks alot for the crash course in drive theory,
Its all going towards building my super mainframe in the future.

SirRiff
 
Re: By the way...

:Hey Euryth I have a question for you.What is the best thing to back your files up on each time you lay down a new track? I know its way to easy for you but if you will?
 
Re: Re: By the way...

Randall said:
:Hey Euryth I have a question for you.What is the best thing to back your files up on each time you lay down a new track? I know its way to easy for you but if you will?

Oh, don't worry about it. I love trying to help people out!

I'm not exactly sure if you mean "track" as in "finished song", or "track" as in, well, you know, individual track. As far as individual tracks go, I just keep the Cakewalk bundles on my hard drive until I'm through with them.

But I'm betting you're talking about your finished songs, so I'll tell you about my backup scheme. Sorry if this is a little more in depth than what you're looking for. :) But here's what works for me...

CDR is the only medium worth considering right now, in my opinion. Every other backup medium has one or more of these problems: Too expensive, too slow, not random access, or not enough storage space.

So here's what I do:

My songs are often pretty complex, track-wise. I've done a couple that used over 50 tracks. Stuff like that requires a lot of track bouncing and destructive effects, due to the speed of my hard drive and CPU. So when possible, I like to export as many individual tracks as possible and hold onto them, just in case I don't like an effect I added, or I want to have someone remix a song, or something like that.

For instance, in a song I recently did, I saved individual .wav files of each vocal track, each keyboard track, each guitar track, etc. So if I decide I don't like my final mix, and I added an effect destructively, I can always go back and grab one of those original dry tracks and try again.

And then, of course, I'll back up my original Cakewalk bundle. I'll try to save a new bundle every time I reach a new stage in the song's progress - maybe one when I've got any MIDI tracks converted to .wav, one when I think I have the backing track where I want it, one when I have the lead vocals down, and one when all the overdubs are done. Something like that.

I know, this sort of sounds like overkill - I'm not actually this religious on every song, but I try to save as much as possible, at as many different stages as I can. The song I just mentioned, for instance, currently takes up two gigabytes on my hard drive.

Once it's finally time to get all of this junk off of my hard drive, I follow a very simple rule: Burn two copies of every CD. Remember this saying: Data doesn't exist unless it's in more than one place. One copy goes into a CD binder near my computer, so I can pop it in anytime I need it. The other copy goes into my normal tower or shelf storage system, and I don't use it, let alone open the case, again unless my original copy goes bad. In that instance, I can break out the "do not use" copy and run off another backup.

It all sounds pretty anal, I know. But I try to carefully archive every song I write, because if I don't have some way to hear it again, I'll forget it. Plus, with stuff I'm recording, a song can take literally months to finish. I would just die if I ever lost all that work.

Hope that answers your question! If there's anything more, I'd be happy to try and help again.
 
Back
Top