IDE and SATA benchmark results

[overclockingoodness]

Hello,
I have my Windows XP Pro on my IDE drive and other applications and benchmarks on my SATA drive.

I was wondering, if I install the benchmarks on my SATA drive and run them - will I get different results (because the SATA drive is my secondary drive) than if I install Windows XP and the benchmarks on the SATA drive and remove the IDE drive from the system?

Thx.

 

[lockmac]

SATA drives are slightly faster than IDE drives so performance should be a little bit better using the SATA drive

 

[overclockingoodness]

Method 1: I know there's a performance difference, but I am asking if I should clean install windows XP and benchmarks on SATA and remove the IDE drive from the system

OR

Method 2: if I should keep the Windows XP installation on the IDE drive, install the benchmarks on the SATA drive and run them from the SATA drive?


Will my benchmark results be skewed if I go with Method 2?

Thx.

 

[Blain]

If you want pure benchmark results, you should load each HD up exactly the same.
Yes, that means reformatting.
If you're going to do all that...
Try different cluster sizes also. And let us know the results.

 

[overclockingoodness]

Well, I was trying to avoid all the hassle and simply run the benchmarks to get accurate results. Other people have told me that the results would not be different (less than 5% difference between the two, if even that).

My hard drives are pretty much the same:

1. Western Digital 7200 RPM 8MB Cache IDE
2. Seagate 7200.7 RPM 8MB Cache SATA

Let me know.

 

[bob4432]

Originally posted by: overclockingoodness
Well, I was trying to avoid all the hassle and simply run the benchmarks to get accurate results. Other people have told me that the results would not be different (less than 5% difference between the two, if even that).

My hard drives are pretty much the same:

1. Western Digital 7200 RPM 8MB Cache IDE
2. Seagate 7200.7 RPM 8MB Cache SATA

Let me know.

the bottleneck is not the connector, it is the actual mechanics of the drive. sata will move, in theory ~150MB/s where ata 133 should move, in theory ~133MB/s, but the hard drives you have listed will only be able to attain a sustained transfer rate of ~50MB/s with ~70-80MB/s burst. at this point in the game, unless you get the newer 74GB raptors, the drives (sata / pata) are basically the same with different connectors.

 

[overclockingoodness]

So basically I won't notice a difference if I run the benchmarks from the SATA drive while I have my OS on the IDE drive, right?

Alright, I think I am going to start benchmarking for my school project. Thanks.

 

[bob4432]

Originally posted by: overclockingoodness
So basically I won't notice a difference if I run the benchmarks from the SATA drive while I have my OS on the IDE drive, right?

Alright, I think I am going to start benchmarking for my school project. Thanks.

depends on how fragmented the drives are, but in theory the difference will be small if both drives are degragged and in good working order. just run hdtach on both and learn

 

[Blain]

Of those two drives, the Seagate WILL win. :laugh:
To benchmark correctly, you should compare apples to apples... Seagate XXXGB IDE to Seagate XXXGB SATA

To save you some time, look at the results from Seagate HDs, all 160GB, all with 8mb cache,
1 SATA with NCQ, 1 SATA without NCQ and 1 PATA

 

[overclockingoodness]

My goal is not to compare the two drives, but my goal is to compare GPU, CPU and Motherboard performance at different clock speeds and settings and my problem is that my OS is on the IDE drive while my benchmarks are on the SATA drive. So, I was hoping that I could avoid reformatting and installing everything on the SATA drive and get the same results.

Even after looking at the data, it seems to me that difference between the PATA and SATA drive is less than 5% (much less in many cases), so that shouldn't make a difference.

Am I thinking this right?

 

[Blain]

The better review sites reformat and reload, even when testing GPU's. Just use and old drive laying around to do your testing with.
Do you want to benchmark the right way, or just be another hack, playing around?

 

[overclockingoodness]

Well, it's supposed to be for a school project BUT would you really notice a difference (in all honesty) between performance if AnandTech used this technique to benchmark GPUs and CPUs.

 

[Blain]

I'm screaming at my monitor...
DO the benchmarks with a fresh reload each time! :shocked:


>> If you don't want to benchmark the correct way... just fabricate results. That way you can get back to your PS2 <<

 

[overclockingoodness]

Okay...fine...geez. You don't have to get downright rude with me just because I asked a simple question.

 

[Blain]

Originally posted by: overclockingoodness
Okay...fine...geez. You don't have to get downright rude with me just because I asked a simple question.

Maybe I didn't, but then maybe I did.

 

[bob4432]

overclockingoodness, use search. there is no reason for you to duplicate what others have done. basically unless you o/c a pretty good amount, you really won't see much difference. 10-15MHz in the FSB won't change your performance much, nor with that amount with the gpu.

 

[overclockingoodness]

Well, I was fine with the first part of the comment, BUT...

>> If you don't want to benchmark the correct way... just fabricate results. That way you can get back to your PS2 <<

...this just offended me quite a bit. You are treating me like I am just another "wanna-be enthusiast" who want to take the short route out of a school project. I, on the other hand, was thinking of adding and proving another point in my project (I can do whatever I want with the project, as long as it relates to the topic).

I think if the performance between the two similar IDE and SATA drives is similar (less than 5% difference between the two), I can point out this fact in my project. When you look at the prices, I bought the Seagate SATA for over $100.00 while my IDE drive was under $40.00 (both are 80GB), so I was hoping to add price to performance ratio in my data tables.

Anyways, thanks for clearing up the misunderstanding.

 

[overclockingoodness]

Originally posted by: bob4432
overclockingoodness, use search. there is no reason for you to duplicate what others have done. basically unless you o/c a pretty good amount, you really won't see much difference. 10-15MHz in the FSB won't change your performance much, nor with that amount with the gpu.

Yes, you are correct and I am aware of that. But before I went on with my testing, I wanted to make sure that I wasn't skewing the results by having the OS on an IDE drive and benchmarks on a SATA drive. Regardless, I think your first post to my topic proved to be quite hopeful and even after looking at the benchmark results (linked in this thread), the difference was only about 5%, so I think I am good to go.


I am not asking if I should benchmark my system after every 10-15MHz increase, but I was asking if my results would be skewed if my OS was on the IDE drive while my benchmarks were on the SATA drive.

 

[Blain]

LoL... :laugh: Mission accomplished!

My personal feeling about benchmarking is this...
* Every variable, except the test subjects should be removed. A.K.A. "Apples to Apples" testing.
* Test runs should be done at least 3 times each to get an average score for each run.
* Data should be "pure" and untainted by personal bias.

 

[overclockingoodness]

LoL... Mission accomplished!

Yep.

Every variable, except the test subjects should be removed. A.K.A. "Apples to Apples" testing.

I agree unless it gives me another point to note in my project.

Test runs should be done at least 3 times each to get an average score for each run.

I will definitely do that.

Data should be "pure" and untainted by personal bias.

Well, I don't think it would do me any good to represent the results in favor of one manufacturer. It's not like my teacher is bribing me (with an A+) to make sure that I get better results with a SATA drive over the IDE.:laugh:

 

[bob4432]

just remember that all the retail stores are dumping their pata drives because of the satas, that is why they are so cheap now. personally, for me i will always use pata over sata. pata or scsi for me