Posted 10/19/2009 at 11:19:42am

Even the CLI defrag operation is not much different from the GUI driven one. The defrag is still slow (takes hours and hours) and all files especially system files are not fully defragged. Back when I had Vista, Diskeeper 2008/2009 was the defragger I used; way better than Vista's defragger - much faster and defragged all the files including the system files.

Posted 08/05/2009 at 10:10:11am

File fragmentation happens to the filesystem i.e. to the logical disk, and regardless of the actual storage hardware (SSD, HDD etc, which are 'insulated' from the filesystem), NTFS will fragment. If the fragmentation is bad, then file access will slow down; SSDs are less susceptible to file fragmentation-related slowdowns since random read time is very low. But SSDs have poor random write performance, and free space fragmentation at the filesystem level can lead to unnecessary I/Os when writing a file (fragments), and in combination with the erase-write latency of the SSD cells, this may slow down the file write. Diskeeper's SSD optimizer as mentioned in the article prevents this by running in the background (It's actually an add-on to Diskeeper 2009) and optimizing the contiguous free space at the filesystem level.

Posted 12/15/2008 at 11:00:50am

File fragmentation does not affect SSDs much because SSD random reads are very fast. Where the SSD falters is in random write times, and this is exacerbated by *free space* fragmentation, not file fragmentation. I am not sure, but I suspect this has to do partly with split (wasted) I/O requests and with the time required to erase+write data in MLC SSDs. Even the Diskeeper guys say that frequent file fragmentation is not required for an SSD, and free space fragmentation is what is necessary
http://www.diskeeperblog.com/archives/2008/12/hyperfast_is_al.html#comments

Another thing, in the article it was suggested (but not recommended) to put the paging file on a USB drive for 'performance gains'. That doesn't sound very logical...the bottleneck here will the USB bus speeds, which are lower than SATA bus speeds by an order of magnitude (480Mbps vs 1.5Gbps). So I doubt whether it will provide any benefit at all, and will most likely hamper performance.