Sequential Read or Burst performance for system/boot drive?

[Agent_K64]

Hello,

My PC now has two hard drives, with the following performance measured in HD Tach, and I'm wondering which drive I should use for installing Windows 7 on:

Seagate Barracuda 7200.9 250GB
--------------------------------------------
Average sequential read: 57.7 MB/s
Random access: 12.9 ms
Burst: 254.3 MB/s

WD Caviar Black 1TB SATA3..running in SATA2 mode
--------------------------------------------
Average sequential read: 112.6 MB/s
Random access: 13.1 ms
Burst: 233.6 MB/s ...also, why is this lower than the Seagate?

I assumed the WD drive would blow the Seagate out of the water, and it does for sequential read, but would this benefit performance as the OS drive, or would the Seagate be better with its higher burst speed and slightly lower random access time?

Thanks!

 

[elel]

Tough. I think that access time would be the most important metric in a boot drive, but is the small advantage in this enough to offset the disparity in read speed? I don't know.

 

[sub mesa]

Burst speed is the speed of the interface (SATA) and the DRAM cachechip - not of the mechanical harddrive itself. The number is basically useless.

If you want to test system disk performance, download HDTune Pro evaluation and run the Random Access test. Or download CrystalDiskMark and look at random 4K - or AS SSD which is very similar.

 

[Agent_K64]

Burst speed is the speed of the interface (SATA) and the DRAM cachechip - not of the mechanical harddrive itself. The number is basically useless.

If you want to test system disk performance, download HDTune Pro evaluation and run the Random Access test. Or download CrystalDiskMark and look at random 4K - or AS SSD which is very similar.

Tried both benchmark tools, got mixed and weird results:

CrystalDiskMark
--------------------
Seagate: C and D
WD: F, G and H

4K
5 tests/1000MB
C: Read: 0.426, Write: 0.691
D: Read: 0.418, Write: 0.618
F: Read: 10.72, Write: 2.636 (50MB test, W7 hidden sys part)..useless test?
G: Read: 0.567, Write: 1.323 (
H: Read: 0.701, Write: 1.361

4K QD=32
5 tests/1000MB
C: Read: 0.608, Write: 0.718
D: Read: 0.556, Write: 0.604
F: Read: 27.84, Write: 2.657 (50MB test, W7 hidden sys part)..useless test?
G: Read: 0.762, Write: 1.361
H: Read: 0.780, Write: 1.440

Looks like the WD is a bit better at reading, and much better at writing. But the odd thing is, why is the 2nd partition on the WD drive faster than the first?? D: Shouldn't it be the other way around? ...

HD Tune Pro: ST3250624AS (Seagate) Random Access

Test capacity: full

Read test

Transfer size operations / sec avg. access time avg. speed
512 bytes 73 IOPS 13 ms 0.036 MB/s
4 KB 75 IOPS 13 ms 0.297 MB/s
64 KB 67 IOPS 14 ms 4.194 MB/s
1 MB 29 IOPS 34 ms 29.228 MB/s
Random 41 IOPS 24 ms 20.827 MB/s


HD Tune Pro: WDC WD1002FAEX-00Z3A0 (WD) Random Access

Test capacity: full

Read test

Transfer size operations / sec avg. access time avg. speed
512 bytes 72 IOPS 13 ms 0.035 MB/s
4 KB 74 IOPS 13 ms 0.290 MB/s
64 KB 70 IOPS 14 ms 4.420 MB/s
1 MB 44 IOPS 22 ms 44.081 MB/s
Random 54 IOPS 18 ms 27.678 MB/s

HD Tune, however, reports the Seagate to be slightly faster on 4KB and smaller when the full capacity is tested. If I test only 40GB (with the checkbox) then the WD drive takes the lead, at 0.538 MB/s vs 0.419 MB/s.

So, a few questions...
Which benchmark should I trust more?
Why does CrystalDiskMark report the 2nd partition to be faster than the first? (First partition is about 195 GB, second is 736 GB.)
Does the partition layout affect performance? Should I change the one I have above?
I'm running in IDE mode. Is it recommended to switch to AHCI?

Thanks for reading over all this!

 

[Agent_K64]

My apologies for double posting.. it seems I'm not allowed to edit my post.

Would the 40GB test in HD Tune be more important than the full capacity one, as the OS would be stored within that first 40GB? Or is this a flawed assumption?