Benchmarking hard drives

[LiamC]

O.K, I'm sure it has been asked before but...

I want to bench some drives and I want to know what knowledgeable people think are good bench's.

I currently have access to Winbench 99, and have *no* intention of using SANDRA. Unfortunately, the iPEAK stuff that SR uses is out of my league..

What else is worth considering and what are the known bugs? I am sure that I have read that HDTach has a few issues.

I will be using W2K SP2 as the target platform and limiting the testing to IDE 7200 rpm drives in stand alone mode (ie *not* RAID).

Thanks

 

[Mercutio]

Isn't ATTO available for free?

 

[Tea]

Winbench is discredited.

Ipeak still has some serious issues to address - don't be misled by SR's hysterical articles proclaiming how perfect it is - if it was half as proven as they say, they wouldn't have to make nearly so much fuss about it.

Sandra and HDTach ... well, I'd rather use Winbench.

It's a minefield out there LiamC.

-----------------------------------

Pending some really practical and reasonably well-tested tool, one that offers a better balance between ease of use and test validity, I end up coming back to simple, low-level measures.

Winbench is a rather cumbersome but I don't think anyone has faulted its access time and DTR measurements.

Vseek is a really useful tool. Tiny, runs off floppy, and it gives you a better idea of the actual seek performance than any other tool I've ever seen. (Except Ipeak, which has a similar module included in it.)

The move to open source benchmarks is promising. RWT has the bare beginnings of a test suite happening, I think, though I don't believe it extends to hard drives yet.

 

[ddrueding]

Resurecting this...specifically I'm looking for a benchmark that can handle RAID arrays. If it is able to out poor-performing SSDs, that would be a bonus.

 

[CougTek]

IOMeter not enough for you?

 

[ddrueding]

IOMeter not enough for you?

It is just awfully complex. IIRC, I've only got it to work once.

 

[Santilli]

This isn't really rocket science.

HDTACH, the free version, will pretty much be good enough. ATTO works, as well.

 

[LunarMist]

Both of those tools are useful to confirm low level drive performance, but are not very useful as benchmarks.

 

[ddrueding]

In case anyone cares:

A pair of 120GB OCZ Vertex in RAID0 on the onboard controller of the Gigabyte X58 Extreme board, default settings, Windows 7 RC.

 

[Santilli]

I used to use that with my macs to set a minimum block size. I do it less with Windows, don't know why. Do you do the same?

 

[ddrueding]

I do it to check for bottlenecks on the system. If it passes a smell test, it's fine with me.

 

[sechs]

I still use h2benchw when I want to compare two drives. But that's a pretty unusual thing.