WD 1000JB & Sandra 2k2

[Mahamoti]

Just installed a shiny new Western Digital 100GB 7200RPM ATA/100 IDE HDD w/8MB buffer... HDTach puts it at 37700k/sec avg, which is on par with other benchmarks I've seen on the net. When I fire up SiSoft Sandra 2k2, though, I only score ~21000, which is 3000 less than the reference ATA/100 drive, and 7-8000 under other benchmarks I've seen. Any ideas why?

System Information:
PIII 1GHz
Intel i815E MB
256MB PC133 RAM
GeForce2 MX/400 64MB
WD 100GB SE HDD (80MB NTFS, 20MB Linux) IDE Primary Master
Seagate 30GB 7200RPM ATA/66 HDD (FAT32) IDE Primary Slave
Creative 12X DVD IDE Secondary Master
Yamaha 16/10/40 CD-R/W IDE Secndary Slave
SB Audigy
3Com 10/100 NIC
300W PS
Windows XP Pro build 2600

All drivers are up to date, including the Intel IDE drivers, and I even flashed the BIOS w/the latest version.

The Seagate drive comes in at about 19000, faster than the reference ATA/66 drive, so I know Sandra isn't configured wrong. Is NTFS really _that_ much slower? If so, that's one hell of an argument for FAT32! Or could it be my Linux partition is confusing Sandra some sort of way? Help! I need justification for buying Western Digital's top drive!

---------------
Mahamoti

 

[pillage2001]

Funny, I'm running win2k and I got like 30k with my WD1200JB if it matters. Your scores should generally be higher than that.

 

[Rand]

Sandra as a HDD benchmark is completely useless, you may as well pick numbers of the top of your head and claim them as benchmark scores.
I've seen 15K SCSI drives get their ass kicked by 7200RPM IDE drive in Sandra, while 5400RPM drives beat out 7200RPM IDE drives, and 10K drives reign supreme.... Sandra HDD benchmarks are all over the place with absolutely no semblance of accuracy when measuring real world performance under any realistic usage characteristics.
Not to mention they tend to be extremely inconsistent, and SiSoft's HDD benchmarks tend to be influenced significantly by numerous other aspects of system performance besides the HDD alone.

I would highly recommend IOMeter, WinBench99 depending upon whether you wish to measure server or desktop application performance.
For low level performance characteristics WinBench 99 is excellent, and HDTach is recpectable.

 

[Mahamoti]

Ok thanks for the comments... I'll give the other benches a try... I'll post the results when I'm done!

*keeping my fingers crossed*

 

[Mahamoti]

Well, I guess I've got more work to do. Here's why:

StorageReview.com WinBench99 results for WD 1000JB: 9.8MB/sec Business, 33.4MB/sec High-End
My Results: 7.6MB/sec Business, 20.4MB/sec High-End

WTF is wrong?!?