New
#21
Ok, based on the mis-information that is rampantly thrown about here about the difference between defragging a ssd and not, lets look at the building blocks of ssd.
I have pulled up the specs on one of the chips used in ssd drives. This is not the only chip used, but the technology is very similar between these chips, and the functionality is typically the same, but speeds may vary.
NAND Flash Memory
MT29F2G08AABWP/MT29F2G16AABWP
MT29F4G08BABWP/MT29F4G16BABWP
MT29F8G08FABWP
• Endurance: 100,000 PROGRAM/ERASE cycles - this means the typical 2gb block can be written/erased 100,000 times.
• Page size:
x8: 2,112 bytes (2,048 + 64 bytes)
x16: 1,056 words (1,024 + 32 words)
• Block size: 64 pages (128K + 4K bytes)
• Device size: 2Gb: 2,048 blocks; 4Gb: 4,096 blocks;
8Gb: 8,192 blocks
• Read performance:
• Random read: 25μs
• Sequential read: 30ns (3V x8 only)
• Write performance:
• Page program: 300μs (TYP)
• Block erase: 2ms (TYP)
So, for those of you who 'ASSUME' that defragging will not improve performance, pay attention.
a sequential read at the chip level is nearly 1000 times faster than a random read. now most of the reads on these drives will be sequential, because the data is written that way, but the more fragmented the drive is the more often its going to do a random access vs a sequential access.
So, performance WILL be improved if the drive is extremely fragmented.
BUT, since you have a very finite limit on the number of write/erase cycles, I wouldn't defrag a ssd more than one time, or at the very most very very seldom. (like once a year maybe at most)
Based on this information, I have a ssd for my OS, and I have a IDE drive for my data. And I'll be disabling my windows swap file, or moving it to my IDE Drive.