hard drives & raid - benchmark and compare!

[Guest eva2000]

more here

4x disks raid 0

 

hdtach3 32MB gave me 114-121MB/s burst and 108-109MB/s average read

 

ATTO 32MB gave me 120+ MB/s read and 112-115MB/s write

 

drivebench 128MB gave me 120+ MB/s read and 113-116MB/s write

 

winbench99 disk tests still running the transfer test but taking ages for 900+ GB raid 0 array - around 20-25mins per 200GB segment LOL - but showing a constant linear line of ~123MB/s so far up to 450GB segment

 

 

 

 

 

[Guest eva2000]

raid 5 with Raidcore BC4852

raid 5 array from 3x 250GB 7200.8 NCQ SATA + 1x 250GB 7200.8 NCQ SATA dedicated spare on Raidcore BC4852 controller hehe

 

1GB file transfer from 250GB maxtor maxline plus 2 PATA single drive to Raid5 array = 60MB/s still :cool:

 

But benchies show slower speeds but still >60MB/s.

 

 

 

 

[red930]

@eva2000

 

When running ATTO make sure you use settings exactly like I have with Direct I/O and Neither in the box below it.

 

This keeps everyone's benchmarks consistent.

[Guest eva2000]

@eva2000

 

When running ATTO make sure you use settings exactly like the I have like Direct I/O and Neither in the box below it.

 

This keeps everyone's benchmarks consistent.

yeah sorry these benchmarks are several months old... before I knew about this thread

[chennhui]

Anyone?

 

Hi,

 

I am refering to

http://www.dfi-street.com/forum/showthread...ighlight=spread regarding the spread spectrum clocking on Hitachi HD. Apparently the spread spectrum clocking option increase the burst speed on RAID 0 (with 2x and 3X HDD) for both 3.0Gbps an 1.5Gbps.

 

Do I need to enable the SATA spread spectrum inside the BIOS option, in order to benefit from the burst speed increase by the spread spectrum clocking? Does spread spectrum clocking decrease the stability?

 

Thanks.

[red930]

@chennhui

 

Changing the Spread Spectrum setting in BIOS had no effect on my benchmarks. Neither did Spread Spectrum on the drive.

 

My benchmark results at all combination of settings were within the margin of error for the benchmarks.

 

The only significant change in performance according to the benchmarks was enabling SATA II over SATA. SATA II resulted in a nearly 90% increase in performance as measured by HD Tach and ATTO.

[BlackDragon24]

Hey Roadie, and others,

 

Have a small problem here. I was able to install a RAID 0 array on my Raptors this morning after troubleshooting a couple of things. First I should say that I was trying to do a 32K stripe with 8K cluster, and am using a Windows XP-SP1 original disk. I was using the original RAID drivers that came with the MOBO.

 

1) I followed AG's RAID guide to setup my stripe and aray, set it to boot, then rebooted into partition magic 8.0 via floppy. Set cluster to 8K, reboot to windows install, hit F6, load RAID driver then storage driver. It allowed me to install all of the files, but upon first reboot, when the widnows installation screen was about to come up, I would get BSOD saying device initatition failure.

 

2) I tried to use default cluster of 4K doing quickformat....this time it allowed me to go thru the entire windows installation process (asked me for product key, etc.), but upon first reboot to get into the OS for the first time, it BSOD'd and rebooted...too quickly for me to read what was going on.

 

3) OK, Now I decide to give the full Default format a go, so I let Windows do the default 4K format (not quick this time). All is peachy. I was able to get into windows, install all of my drivers, and am able to write this to you guys now.

 

Now, this 32/4 setup is quite fast, don't get me wrong, but I am curious why I couldn;t setup anything but the default cluster and only by doing the full format (not quick) within windows setup?

 

Any thoughts? I thought the 4K default necessity was only for XP SP2?

[Sorrento]

ExRoadie, what about enabling Spread Spectrum on a motherboard that doesn't support it?

 

I was thinking about how the BIOS setting affected the results, and perhaps its just too little what it can do or maybe it doesn't work... but I think the motherboard has to support this function in order to work with it. What do you think?

[red930]

The Spread Spectrum setting in BIOS had no effect on my benchmark results. Same for the setting on the drives.

 

I guess my cables are of sufficient quality to render the setting unnecessary. I am not using DFI SATA cables. I scammed some SATA cables from the Shuttle boxes I've built over the last year or so.

ExRoadie, what about enabling Spread Spectrum on a motherboard that doesn't support it?

 

I was thinking about how the BIOS setting affected the results, and perhaps its just too little what it can do or maybe it doesn't work... but I think the motherboard has to support this function in order to work with it. What do you think?

[Sorrento]

 

Thanks.

[chennhui]

@ExRoadie, thanks for the useful information.

[Frallan]

Ohh well looks like I wont spend the money on additional HDDs to go RAID5...