Hi guys)

Lurking around in forums I encountered a couple of remarks that 32 or even better 64KB cluster size is a must for modern HDD's regarding the performance-vs-features ratio. Indeed, it's speed vs cluster waste/NTFS-features related, but I would like to hear from you:Does it depend on physical parameters, OS, usage, or something? If it matters, I don't use NTFS compression or encryption features.

TY

Short answer is small clusters make more use of your hard disc space with slower access times, while large clusters utilise less of your hard disc space but offer a faster access time.

In my opinion that's one of those things you should really leave at whatever the default is. Unless you usage is very unusual, such as certain large database applications.

_________________
Data Medics - Hard Drive, SSD, and RAID Data Recovery Service Company

It may sound contrarian, but I wonder if there may be circumstances where larger cluster sizes would be better for small files. For example, if you have thousands of small files in an NTFS volume, would it be better to accommodate the entire file within the MFT? AIUI the default MFT record size is 1024 bytes, but it could be increased by adjusting the "Clusters Per MFT Record" parameter in the boot sector.

_________________
A backup a day keeps DR away.

Hi NRA,

As you can see here all depends what do you want to use ARRAY.

can you say us what is the usage of your Array?

It's not about RAID/SATA/IDE, or something; it's just as a general recommendation to sacrifice small files cluster waste and minor features (e.g. NTFS compression/encryption) for the sake of general performance increase.

Should you check a couple of forums, you could find similar statements--although no proofs, no OS/FS-wise, nor regarding the purpose (boot/OS, temp, filestorage, etc), whatever.

Actually, I did a small test using a clean 500GB SATA-2 HDD formatted as default NTFS and 64KB cluster, and installed (without full formatting) a clean Win8.1 x64 with mobo box drivers, and it does show that copying from Partition 1 to the same type (be it default 4KB or maxed 64KB cluster) Partition 2, say, the 4GB non-fragmented file is averaged to 79-37MB/s vs 104-63MB/s, which makes me wonder a little; even WMWare shows some 15MB/s difference in cluster sizes.

That's why I would like to learn how it effects/affects the HDD performance in near future and whether it's any worthy in general for a common user--and why exactly yes or no. Also I would like how reasonable it is for RAID/SAS, if you know or could assume/suppose.


TY