Maxtor D740X SANDRA benchmarks

[Dr Mordrid]

Just installed a 2 drive array on one of my other systems and thought I'd share its scores given that some will wonder how various drives will do.

Here's the setup.

ECS K7S5A mainboard (yup, another one....)
AthlonXP 2000+
256 megs DDR256 RAM
Windows2000 SP-2
Fasttrak100
2 Maxtor D740X 60 gig drives (120g array)
2 drive RAID0 (what most folks will set up)
64k stripe block, set up for performance

SANDRA 2002 scores;

Composite score: 32344 (meaningless for video)

Sequential write: 59 mbytes/s (important for caputres)

Sequential read: 48 mbytes/s (important for playback)

Seek: 8ms

Not bad for a 2 drive array....

Dr. Mordrid

[Sciascia]

2 WD Ultra 10 at RAID 0 on a FT 100

Composite: 25161

Sequential write: 38MB/s

Sequential Read: 37MB/s

Seek: 11ms

Something wrong here, of course my IRQ's are as follows:

Everything that uses any real bandwidth: IRQ3

WinXP and ACPI

I need to scrounge up another 80 pin IDSE cable to test the difference in master and slave connectors. Not real optimistic here though.

[Dr Mordrid]

Mine is running ACPI as well, so that *shouldn't* be a factor.

On IRQ-11 I have;

Matrox G400DH
Linksys LNE100TX NIC
Turtle Beach Santa Cruz
Fasttrak100
IEEE-1394
USB Host Controller

There have been reports that XP has slower disk performance, so maybe that's it.

Also damning for XP was a report in InfoWorld summarized on Tech-Report.com;

http://www.tech-report.com/news_reply.x/3076/

Dr. Mordrid

[davpen]

I don't know how much faith I would put in that comparison of XP and W2K. Some of it seems down right strange.

"Our tests on a dual-CPU system indicate that both Windows XP and Windows 2000 run better on an SMP (symmetric multiprocessing) configuration with relatively slow CPUs than on a single-CPU system with a screamingly fast processor. As we added more and more load, the benefits of a dual-processor configuration became more apparent. Both OSes (using Office 2000 and optimized UIs) handled the heaviest workload (scenario 3) nearly 40 percent faster on the SMP client machine [a dual 1GHz PIII] than on the [1.5GHz] single-CPU Pentium 4."

Hardly fair and realistic to compare a PIII to a P4 like this as even a single 1Ghz P3 can outperform a 1.5Ghz P4 on some tests. And who knows what other differences there are between the systems being compared as they are never described in the original article. It sounds like they have fallen for the Intel propaganda that the faster clock of the P4 actually means it is faster and that suggests they are relatively inexperienced with PC platforms. That and the lack of disclosure about the test platforms has to cast some doubt on their results.

Not that I am not trying to defend XP. My main system uses W2K, which is the way it is going to stay for the forseeable future. I have tried XP but have not found anything in it that is of any real use to me or that I haven't already covered with better third party utilitys. And it has some things that are down right annoying (You no longer have to browse the Internet to get pop up advertising and SPAM for useless products as Microsoft has built it in to the OS, now that's progress!) That's not to say XP is all bad. It does have some useful features and extras depending on what you are going to use it for.


David.

[dchip]

Sandra

I feel sort of dumb asking because I should have gotten this diagnostic tool a long long time ago, but, where can I purchase a download of SiSoft Sandra?
I found this old link http://www.sisoftware.demon.co.uk/sandra/ in MURC General-Hardware but is that the best place to get this tool? Thanks.

[Greebe]

I personally don't trust Sandra for benching HD's even tho I do have it.

IMHO HDTach is much better

[The Scientist]

yep, that's the place to go.

AFAIK only the professional version needs to be purchased,
the standard version is a free download. Of course the professional version gives you a few more information and
benchmarking modules but all of the drive / cpu & memory
benchmarking modules are present and fully functional in the
standard edition.

Simon

[dchip]

Thanks.

[Sciascia]

Changed the cables, updates the FT100 BIOS and drivers, no improvement, about 40mb/s now. HDTach shows my system drive being faster than the RAID drive. Ultra 100 drives all around, 2 should be faster than one.

By the way what allocation unit size should I be using for the RAID drive when formatting? I don't know what the defalt is, but I did set the FT100 for A/V editing which is 64k I think. Not sure that would make a difference in speed, but just wondering what it should be.

[Indiana]

I thought that the Maxtors are faster than the IBM?
My array with two IBM 40GB GXP75 IC35L040 gets this score (HPT370, Win2k):



The radom read/write numbers seem to be a bit low to me, I seem to remember having better results before my Win2k reinstall.
There was a commandline argument for better HD-throughput in Win2k, disbling random OS accesses to the drives, however I can't remember it. Doc, I think you were the one, that suggested this several months ago, could you please rephrase this commandline?

[Technoid]

diskperf -N
Disables all performance counters

[Sciascia]

A little BIOS tweaking got me up to al ittle more reasonable numbers, but still not where it should be.

Sequential read: 51 mb/s

Sequemtial write: 40 mb/s

One thing that I did notice was that your system is file server optimized. That may be the last key, though I assume that you are running Win2k or 98SE. I'll have to dig around and see if there is a way to change that in WinXP.

I've also noticed that the Promise utility does not seem to be working correctly. There is no difference at all between SMART being enabled or disabled, although the write buffer enabled or disabled does seem to have an effect.