Post your HD Tach scores here!! *Update* Awesome results with 16k block size!! :)

NFS4

No Lifer
Oct 9, 1999
72,636
47
91
Duron 800 o/c to 945MHz (7 x 135) on Abit KT7A-RAID

(2) Western Digital Expert 27.3GB ATA66/7200RPM in RAID-0, 64k stripe size

Using HD Tach 2.61

Random Access Time 13.1 m/s
Read Burst 60.3 mbps

Read Speed
Max 37234 kps
Min 2622 kps
Avg 28861 kps

CPU utilization 10.9%
 

rickn

Diamond Member
Oct 15, 1999
7,064
0
0
My single IBM 15GB 75GXP P3 650e @ 961 148mhz FSB

Random Access Time 10.5m/s
Read Burst 60.1 mbps

Read Speed
Max 37915 kps
Min 23313 kps
Avg 32839 kps

CPU utilization 5.2%



 

Budman

Lifer
Oct 9, 1999
10,980
0
0
Here is it.

Random Access Time 15.2 m/s
Read Burst 75.6 mbps

Read Speed
Max 50488 kps
Min 18731 kps
Avg 36568 kps

CPU utilization 14.7%




There must be something wrong with your setup nfs4,also Win2k gets much lower score than win98.


I have 2 5400 Samsung drives on a KT7-raid using a 16k stipe,I find the 16k stripe much faster than a 64k but at the expense of more cpu usage.
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91
Hmmmmm, I'm running Win2k Pro...and this is after I did a clean install yesterday.

Maybe I'll reformat and try the 16k stripe size.
 

Budman

Lifer
Oct 9, 1999
10,980
0
0
Yes after you use the 16k stipe format using this way format x: /:32 this creates a 16k block size equal to the stripe size.
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91


<< Yes after you use the 16k stipe format using this way format x: /:32 this creates a 16k block size equal to the stripe size. >>


Will Partition Magic 5.01 do this? That's what I used to make this partition in the first place.
 

LickEmSmack

Senior member
Jul 4, 2000
389
0
0
Just benched it... And as I suspected, the scores are way too low for this drive... this has been bugging me for a while now...

Random Access Time: 17.1 ms
R.B.S.: 38.4 mbps
CPU utiliz: 13.5%

Rig: TBird 1GHz @1.225 GHz on Abit KT7A
Single 75GXP on the ATA100 channel
installed the Abit drivers for the ATA100 controller not the MS hotfix... seems ot work &quot;fine&quot; tho. Any tips on how to tweak?
 

ucdnam

Golden Member
Jan 28, 2000
1,059
0
0
1st Fujutsi 18.2 gig 10K RPM SCSI [MAG3182L]
9.8 ms Random access time
57.7 mbps Read Burst speed
30823 kps max, 8941 kps min, 26032 kps average Read Speed

2nd Fujutsi 18.2 gig 10K RPM SCSI [MAG3182L]
9.0 ms Random access time
56.4 mbps Read Burst speed
30835 kps max, 19015 kps min, 26241 kps average Read Speed

3rd Maxtor 53073U6 30 gig 7200 rpm IDE
14.0 ms Random access time
49.4 mbps Read Burst speed
31728 kps max, 15198 kps min, 24144 kps average Read Speed

My Rig
CUSL2 w/ P3852 @ 142 FSB
384 MB Ram
Pioneer DVD-104
TDK CDRW12/10/32
Adaptec 2940 U2W
ATI Radeon ES (64 MB)
SB Live Value
Netgear FA310TX NIC

All on a nice DSL connection :)
 

Vinny N

Platinum Member
Feb 13, 2000
2,277
1
81
NFS4:

I could have sworn someone determined that HDTach doesn't correctly measure the performance of RAID arrays...

When you run HDTach what drives does it list as a choice to test?

AFAIK, it lists all the physical drives in the system, versus testing a certain partition(which in the case of RAID 0 would span the two drives).

So I think you're measuring just one of the drives rather than the RAID 0 array.

As for my HDTach scores...

here ya go :)

hdtach bench
 

sad

Senior member
Jun 15, 2000
437
0
0
raid0 fireball AS

HD Tach 2.61
Random Access Time:9.1ms
Read Burst Speed:74mbps
Read Speed:
Max:75757.0kps
Min:26430.0kps
Avg:60250.4kps
hdtach.jpg
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91


<< NFS4:

I could have sworn someone determined that HDTach doesn't correctly measure the performance of RAID arrays...

When you run HDTach what drives does it list as a choice to test?

AFAIK, it lists all the physical drives in the system, versus testing a certain partition(which in the case of RAID 0 would span the two drives).

So I think you're measuring just one of the drives rather than the RAID 0 array.
>>


You may be right...it lists my WD as:

(Read-Only) &quot;Partitioned&quot; WDC273BA

Then it has a listing for:

(Read/Write) Physical Drive 0

 

ahfung

Golden Member
Oct 20, 1999
1,418
0
0
P3-827
Epox BX7+
256MB SDRAM
Windows 2000 SP1
IBM 75GXP DTLA 30GBx2 RAID0


With Highpoint HPT368

Random access: 12.5ms
Burst Read: Over 80MB/s, can't read
Read Speed (max;min;avg): 80.7MB/s;19.7MB/s;56.1MB/s
CPU utilization: 16.7%

With Promise Fastrak 100

Random access: 12.5ms
Burst Read: Over 80MB/s, can't read
Read Speed (max;min;avg): 58.0MB/s;3.1MB/s;37.6MB/s
CPU utilization: 7.3%
 

sad

Senior member
Jun 15, 2000
437
0
0
Budman, no can't do:frown:. It locks up everytime I have that checked.
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91


<< Budman, no can't do. It locks up everytime I have that checked. >>


Same here :(
 

Budman

Lifer
Oct 9, 1999
10,980
0
0
I think i remember reading somewhere that you must turn off cd auto insert notification to fix this.
 

sad

Senior member
Jun 15, 2000
437
0
0
Auto Insert notification has no effect at all. Those with Kt7a set PCI#2 Access #1 Retry in bios to disabled as recommended by the kt7faq and you'll get a boost of speed.
 

borealiss

Senior member
Jun 23, 2000
913
0
0
ibm ultrastar 9EP
max read 19k, min 10k, average 15k
cpu utilization 1.7 %
average seek 10 ms
read burst 50 megs/sec

quantum atlas III
max read 18k, min 14k, average, 13.3k
cpu utilization 1.7 %
average seek 9.1 ms
read burst 69 megs/sec ....yeah, that's right. i double checked that number too.

//edit
forgot read burst
 

przero

Platinum Member
Dec 30, 2000
2,060
0
0
2-30GB IBM 75 XGP - RAID-0

Random Access Time - 9.9ms
Read Burst Speed - >80mps

Read Speed
Max. 51482 kps
Min. 26114 kps
Avg. 41226 kps


 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91


<< Those with Kt7a set PCI#2 Access #1 Retry in bios to disabled as recommended by the kt7faq and you'll get a boost of speed. >>


But will that work since I'm on the Highpoint controller?
 

sharkeeper

Lifer
Jan 13, 2001
10,886
2
0
Please do NOT FORCE a 16 KB allocation unit size on a 30+ GB volume! This will result in far too many allocation units (clusters) and cause performance to drop. Sensitivity to defragmentation will also increase. If you want less slack, use NTFS.

Block (stripe) size and allocation unit (cluster) size are too completely different things! Don't confuse the two!!!

For a pair of 27.3 GB drives, 32KB allocation unit size is default and recommended. 16KB block size is the recommended block size for FAT32 with 32KB clusters.

If you convert your volume to NTFS (start>run>cmd: CONVERT C: /FS:NTFS) it will be using 512 BYTE (0.5K) allocation units. NTFS is more efficient with larger volumes especially if they contain of lot of files.

HD Tach in Windows 2000 tests full volume of the drive without selecting advanced size check. Only in 9x do you have to do this otherwise it only checks the first 8GB and it will also make the seeks seem a lot faster than they are.

For testing the true read and write performance of your array, I recommend using ATTO bench.

Cheers!
 

NFS4

No Lifer
Oct 9, 1999
72,636
47
91
Well, when I get a chance I'll convert the drive to NTFS and change the stripe size to 16k.