Benchmark ATA against Promise RAID 0?

  • Thread starter Thread starter Noozer
  • Start date Start date
N

Noozer

Hey all!

Right now I have a single ATA drive and a pair of RAID drives. Both have the
same image on them. I figure it's a good time to try and benchmark the IDE
interface against the Promise onboard RAID on my Asus mainboard.

Is there a simple benchmarking program I can run to see what kind of
throughput I'm getting on these drives?

Thanks!
 
Noozer said:
Hey all!

Right now I have a single ATA drive and a pair of RAID drives. Both have
the same image on them. I figure it's a good time to try and benchmark the
IDE interface against the Promise onboard RAID on my Asus mainboard.

Is there a simple benchmarking program I can run to see what kind of
throughput I'm getting on these drives?
Click to expand...
 
Noozer said:
Hey all!

Right now I have a single ATA drive and a pair of RAID drives. Both
have the same image on them. I figure it's a good time to try and
benchmark the IDE interface against the Promise onboard RAID on my
Asus mainboard.

Is there a simple benchmarking program I can run to see what kind of
throughput I'm getting on these drives?
Click to expand...

HD Tach is pretty popular for general HD benchmarks.
http://www.simplisoftware.com/Public/index.php?request=HdTach
 
Noozer said:
Hey all!

Right now I have a single ATA drive and a pair of RAID drives. Both have the
same image on them. I figure it's a good time to try and benchmark the IDE
interface against the Promise onboard RAID on my Asus mainboard.

Is there a simple benchmarking program I can run to see what kind of
throughput I'm getting on these drives?
Click to expand...

I've installed Sandra 2004.10.9.89 and did some bencmarking.
- Windows XP SP1a & all updates
- Asus P4C800-E mainboard
- 2.6Ghz P4 @ 800Mhz & single stick of Kingston PC3500 (no overclocking)
- 60gig Maxtor D740X-6L (whiny bearing)
- 40gig Seagates are ST340016A Barracuda IV (quiet)
- All drives have same image installed and were defragged before
benchmarking.

These number don't seem to make sense... I used the File System Benchmark.
Each test was run twice with and without the Windows cache involved.

Single Maxtor 60gig PATA alone on normal IDE port:
- 18896 KB/sec & 18076 KB/sec without using Window caching
- 14536 KB/sec & 14965 KB/sec using Windows caching (Lower???)

Two Seagate 40gig Barracude PATA drives on single cable using Promise RAID0
with 128k blocks:
- 17002 KB/sec & 18023 KB/sec without using Windows caching
- 15750 KB/sec & 15917 KB/sec using Windows caching

I tried to also connect a single 80gig Maxtor drive to the Promise in IDE
mode, but it wouldn't boot with the image. (Drive does boot fine in another
PC on normal IDE controller)

So...
Why are the drives faster when NOT using the Windows HDD caching system?
Why is the single Maxtor drive faster than the RAID0 configuration?
Are this numbers good/bad/average?
RAID0 also offered a smaller block size (16k or 64k, not sure). Would this
be faster?

Thx!

P.s. In regards to cutting a short ATA cable mentioned in another post...it
was a different PC.
 
Noozer said:
I've installed Sandra 2004.10.9.89 and did some bencmarking.
- Windows XP SP1a & all updates
- Asus P4C800-E mainboard
- 2.6Ghz P4 @ 800Mhz & single stick of Kingston PC3500 (no
overclocking)
- 60gig Maxtor D740X-6L (whiny bearing)
- 40gig Seagates are ST340016A Barracuda IV (quiet)
- All drives have same image installed and were defragged before
benchmarking.

These number don't seem to make sense... I used the File System
Benchmark. Each test was run twice with and without the Windows cache
involved.

Single Maxtor 60gig PATA alone on normal IDE port:
- 18896 KB/sec & 18076 KB/sec without using Window caching
- 14536 KB/sec & 14965 KB/sec using Windows caching (Lower???)

Two Seagate 40gig Barracude PATA drives on single cable using Promise
RAID0 with 128k blocks:
- 17002 KB/sec & 18023 KB/sec without using Windows caching
- 15750 KB/sec & 15917 KB/sec using Windows caching

I tried to also connect a single 80gig Maxtor drive to the Promise in
IDE mode, but it wouldn't boot with the image. (Drive does boot fine
in another PC on normal IDE controller)

So...
Why are the drives faster when NOT using the Windows HDD caching
system? Why is the single Maxtor drive faster than the RAID0
configuration?
Are this numbers good/bad/average?
RAID0 also offered a smaller block size (16k or 64k, not sure). Would
this be faster?

Thx!

P.s. In regards to cutting a short ATA cable mentioned in another
post...it was a different PC.
Click to expand...


Asus A7N8X Deluxe Rev2
AMD 2500+ (2300MHz)
WDC 80GB JB 8MB cache
512MB PC3200

Using a single ATA drive SANDRA 2004.10.9.89 File System Benchmark gave me
an average of 35363 KB/sec over three runs with write caching enabled.

You said: " Two Seagate 40gig Barracude PATA drives on single cable using
Promise
RAID0 with 128k blocks"

Unless I'm misunderstanding the way this RAID controller works, you will
need one
drive per IDE channel. Jumper each as Master and connect each drive on the
end of separate IDE cables. Then create the RAID0 array. 64KB chunk size is
optimal unless you
are moving large files the majority of the time. In that case, 128KB chunks
are better. See pages 5-25 and 5-26 of the manual.
 
S.Heenan said:
Asus A7N8X Deluxe Rev2
AMD 2500+ (2300MHz)
WDC 80GB JB 8MB cache
512MB PC3200

Using a single ATA drive SANDRA 2004.10.9.89 File System Benchmark gave me
an average of 35363 KB/sec over three runs with write caching enabled.

You said: " Two Seagate 40gig Barracude PATA drives on single cable using
Promise
RAID0 with 128k blocks"

Unless I'm misunderstanding the way this RAID controller works, you will
need one
drive per IDE channel. Jumper each as Master and connect each drive on the
end of separate IDE cables. Then create the RAID0 array. 64KB chunk size is
optimal unless you
are moving large files the majority of the time. In that case, 128KB chunks
are better. See pages 5-25 and 5-26 of the manual.
Click to expand...

It's an ASUS P4C800E mainboard. Only one PATA raid connector on the
mainboard. It's really designed for SATA RAID, but does support PATA
connector.

I might just reimage the drive using the smaller chunks to see if that helps
at all.
 
Back
Top