HDD Benchmark clarification question

[buzzzsaw]

Dear Sir or Madam,
I am currently researching mechanical hard drives with the intent of purchasing the top performer in order to improve my overall benchmark scores. I always use your site as the base for my research whenever I am in the market for a new desktop PC component. This time I have stumbled upon an anomaly that I hope you can clarify for me. The scores and rankings in your "All 2009 3.5” Desktop Hard Drive Charts", PC Mark Vantage scores , do not match up at all with the scores and rankings for HDD's on the "PassMark Software Hard Drive Benchmarks" website (http://www.harddrivebenchmark.net/hdd_list.php ). For example they list the Samsung HD502HJ as the top mechanical while I find it at #18 in your chart. Unfortunately I neglected to review your HDD charts first and have just purchased 3 HD502HJ drives based on PassMarks rankings. They haven't shipped yet and I can cancel my order, I hope, if I am way off base with this purchase. I am completely confused now, panicing, and hope that you will be able to help me by clarifying and identifying which mechanical Hard Drives are the best performers and how I should reconcile the information in both sets of benchmarks.
Thank you.

 

[sub mesa]

in order to improve my overall benchmark scores

So you don't care about actual performance, you just want to see high benchmark numbers? Is this a research project only - or are you looking for a HDD to use for your own?

As for benchmark deviations; it all depends on the situation in which the benchmark was conducted. About 80% of all reviews online are faulty; they are improperly conducted or their numbers are off. The test procedure or specific issues may give a distorted image when you compare results based on faulty data.

As for mechanical disks; they don't differ at all that much in terms of performance. The fastest HDD is the new Velociraptor, 2,5" 10k rpm. It's still hundreds to thousands of times slower than an SSD regarding random I/O.

 

[buzzzsaw]

Thank you for the information sub mesa.
I am upgrading my system and would like to have as close to the top performing mechanical hard drives as I can afford. About every 12 or 18 months over the last 10 years my operating system hard drive has failed. To provide some stability I am now running a Raid-1 configuration but am about to switch to a Raid 0+1. Also, whenever I did a benchmark on my system I noticed that my scores were being dragged down by my hard drives. I cannot afford the current prices on SSDs and since I am about to switch to a Raid 0+1 I would like to get the top performing mechanical. The chart on PassMarks HDD benchmark site shows the Samsung HD502HJ as the top performer, beating even the 10K rpm drives. But the benchmark data on Tom's Hardware is not in agreement with PassMark. I simply want to reconcile both sets of information to ensure I make the correct purchase decision. Thank you for any additional information you can provide me.

 

[sub mesa]

Frankly, i would advise to save your money until you can afford an SSD, if you care about performance. HDDs are only good at storing large files and reading/writing them sequentially.

Also, you should not mix different type of drives in the same array, if you care about performance. It's not that it doesn't work, but as the drives will have different firmware they behave differently to I/O which may/will affect performance.

Alternatively, you may focus on more RAM (8GB/16GB) which will make disk I/O less of a problem by caching files. Investing in multiple HDDs to make them faster only yields very low rewards compared to the performance increases of Intel SSDs.

 

[eloric]

I did the same thing as you - the Samsung Spinpoints were ranked in the 60s at the time I purchased them. Got them home, installed in RAID 0 array, then uploaded the results back to passmark. It appears that they jumped in ranking overall to somewhere in the 30's!!!! Guess a whole bunch of us must have done the same thing to skew the statistics.
I am not diasappointed at all. For the price point, they are one of the best deals out there. Mine are quiet and cool, and I recommend them to my friends.

 

[buzzzsaw]

Thanks for the info eloric. That makes me feel a little better. When I discovered PassMarks site I got carried away and neglected to check back on Tom's Hardware. When I saw the ratings for the HD502HJ on PassMark I decided to go with it. I ordered three, they are shipping today. There is a large difference between Tom's and PassMark's data. I really would like to find out if PassMarks info is correct. If it is, I would like to keep using it.

 

[geofelt]

@buzzsaw:

If your objective is to win a benchmark competition, then by all means buy the best benchmarking product.

If, on the other hand, you want to improve some performance aspect of your PC, then realize that benchmark numbers are only a small part of the considerations.

A hard drive performs faster from the outer cylinders of the drive. Is that where your access patterns are located?
Are your I/O operations mostly random or sequential?
Do you do mostly reads or writes?
Is your usual I/O a large block or a small one?
Do you have many interfering tasks?
Do you do one I/O at a time, or do you build up a large queue?
Does your pattern change?
Will it in the future?
How full is your drive?
Do you do lots of deletes and rewrites?

-----Nobody knows--------

If you want a more satisfying OS experience consider a good SSD You can get an Intel X25-V for $125 today, and prices are coming down.

 

[buzzzsaw]

Best answer selected by BuzzzSaw.