Hard Drives - any benchmarks actually show the built-in CACHE makes a difference?

[MadRat]

Between a 512k, 2mb, and 8mb cache I do not see any real world difference. The larger caches are supposed to be way better for transfers; the caches allow the hard drive to continue pulling data while the drive controller is waiting to use the IDE channel. But for sustained transfer speed using one device on the IDE channel does it actually benefit us to have larger caches?

 

[Derango]

Find benchmarks of the 2MB Western Digitial 100GB drive, then find the same benchmarks for the 8MB WD drive. I belive the only thing different between those drives is the cache size.

 

[Rand]

<< But for sustained transfer speed using one device on the IDE channel does it actually benefit us to have larger caches? >>



A larger buffer only has a very minmal gain in terms of sustained read/write speeds across the drive, indeed the buffer size alone seldom has any significant impact upon any low-level performance characteristics of hard drives at all.
It's the overall results and average service times for any single given operation that show the significant gains to be had.



<< I belive the only thing different between those drives is the cache size. >>


Firmware.
The formware is at the heart of the differences between the 2MB and 8MB revisions. Without the firmware being specifically optimized for the larger cache the performance differences would be virtually non-existent as is evidenced by many of Maxtor's initial 512KB/2MB HDD's that performed identically depsite the significantly different cache sizes.
Firmware optimizations can have a more significant performance impact on typical desktop applications then can any single low level performance characteristic of the drive.

StorageReview has done a couple reviews of similar drives with varying cache sizes, most recently Western Digital's 2/8MB cache variances on their 100/120GB drives.