• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

WD 80gb 8mb cache or Maxtor 80gb 2mb cache?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

nymph

Member
Joined
Nov 17, 2002
Location
Orlando
I have been using an 80gb, 7200rpm, ultra ata 133, 2mb cache maxtor for a while.

I just purchased a WD 80gb, 7200rpm, ultra ata 100, 8mb cache because the price was right for me.

I wanted to decide if I should put the WD on my gaming/benchmarking rig, so I made a mirror copy of my complete Maxtor hard drive to my new WD hard drive. Then I ran a couple of simple bencharks on both.

The first test on each set is the Maxtor and the 2nd test is the WD.

1_2mb.jpg

1_8mb.jpg

2_2mb.jpg

2_8mb.jpg

3_2mb.jpg

3_8mb.jpg



I ran these tests a few times and pretty much came up with the same numbers, except for the cpu usage. The cpu usage for the WD got as high as 24%!! And it did this consistantly. While the maxtor stayed pretty much the same. Even though in pcmark 2002 the 8mb cache edged out the 2mb cache by a few points the other results may have more of an influence on me. Especially the cpu usage and read burst speed. Also not to mention that out of an 80gig hard drive I have 78 Gigs available from maxtor and only 74gigs available from WD.

In what way would the 8mb cache really outshine this maxtor? Is it supposed to be in gaming? I played a few of my games and noticed no difference? Would it be in copying large files? If someone knows of another free utility to use for benchmarking I would appreciate a link. Thanks.


nYmph
 
Last edited:
The larger cache will help on retrieving many medium size files that are frequently used. It may help database applications, but the seek time and sustained transfer rate are really the ones that help databases. As far as gaming is concerned, most games load everything into ram and then run from ram. The only impact would be on how fast the game loads. But, again, it is mostly dependent on seek time and sustained transfer rate.

All you get from a faster hard drive is bragging rights!
 
These drives demonstrate the different performance factors well. They have nearly identical areal data density, as reflected by the nearly identical sustained data transfer rates (STR). This particular Maxtor has better seek performance due to a faster mechanical design. The WD has better performance in many situations though due to the large buffer as well as WD's more advanced buffer management techniques. In the end the drives will act more alike than they will different.

The cpu utilization numbers point to either a mucked up windows or bunk IDE drivers. It is not possible for different hard drives to result in different cpu utilization, this is a function of the IDE controller and its drivers. The slight differences you see in the reading are just an illustration of the imprecision of measuring cpu utilzation in the manner that HDTach does. Nothing to act upon there, other than to install windows fresh and experement to find the best IDE driver for your board's controller. CPU utiliztion should be less than 5% with either drive.
 
I added a new test I found with SANDRA. I have other partitions, but I used only my main OS partition for this test. They seem close again, but the maxtor seems to be able to read more but then again has a 2ms more access time....

As far as the IDE drivers go, I will try and renisnstal them first without having to reinstall windows. I actually renistalled windows about 3 weeks ago... What cpu usage do you get (if you run HD Tach)?

Which drive would you yourself choose for a system mainly for gaming/benchmarking/occasional video encoding?

Thanks..


nYmph
 
Although it is comparing apples and oranges due to the different IDE controller and drivers, my machine reads 3-4% on HDTach's cpu utilization test. I don't have the Nforce2 experience to tell you what drivers are best, but the high cpu utilization numbers point to a cluttered windows and/or driver optimization issue. If your windows is that recent it is markedly less likely that the windows install itself is flawed, but don't rule out the possiblility until you solve the problem. Given that you have two drives I'd be tempted to lay down a fresh windows on one to rule out the possiblity before I decided to stay with the original install.

As far as which drive I would use, for me it boils down to the noise. If the WD is a good example it will be basically silent in operation, where Maxtors rarely are. The performance is going to be so close it matters more what task you care to optimize than it does the overall superiority of one drive or the other. My suggestion would be to ignore benchmarks (outside of diagnosing the cpu utilization issue), use both drives, and stay with whichever you find more pleasurable in daily use.
 
One variable which may be entering into the results are the cluster sizes. Some versions of Ghost default to a 512K cluster size and this can negatively affect drive performance and CPU utilization. Ghost does many things well, but can be less than optimal. if you are not careful with the settings.
 
Do you mean Norton ghost? I used drive copy 4.0. I did however check the cluster size since you mentioned it and it is 4kb.

Larva, I had my cousin and uncle both test their hard drives with HD tach. My cousin has the exact same WD hard drive and MOBO as me and he also got high cpu usage (24%).

My uncle has the 120gb version of this drive and gets 16-20% also. He has an Nf7-S. I think this usage is normal for this type of setup we have (nforce drivers and HDD).

Just out of curiosity I rolled back my IDE drivers to the microsoft ones and my CPU usage junped to 68.5%!!!

I think I am going to keep the 8mb cache in my main rig for gaming, etc and use my 2mb cache for another computer. Thanks for the replies...

nYmph
 
I simply don't know what is normal for a NF2 on this test. I don't use AMD stuff in general because of niggling refinement issues such as this. I have built many AXP's, but none since the advent of NF2. It may well be normal, or it may be a sign of an actual problem. Perhaps you might inquire in the more heavily trafficked AMD CPU section for input from users that have the experience to answer your question definitively. Myself it seems extreme, but sadly at times I find the basic building blocks of the AMD-supporting chipsets to come up wanting in measures such as this.
 
I think I will check the other forum to see what other peoples results are with AMD systems like you suggest.... Before I do though, I thought I would add that I tested my other system which is a WD 40gb 8mb cache with a Asus A7v333 (via kt333) and my cpu usage was 24.9%. You may be right in what you mentioned about this perhaps being normal for AMD systems....


nYmph
 
Yep, the Nforce2 is a paramount of refinement compared to Via chipsets, but still lacking compared to the generally unmatched Intel efforts.
 
Back