Power Full Cloud Hosting Zones

Home » Archives for September 2015

Ram lesson Key points

Posted by Ktech | 10:08

Ø The disk controller interface governs the maximum data transfer rate. IDE, SCSI and ESDI are roughly twice as fast as the tradition...

0 comments

Ram working Keeping it fast

Posted by Ktech | 10:06

One phenomenon that many users of high performance computer notice is the gradual deterioration in the speed of access of their di...

0 comments

Wednesday, 23 September 2015

Ram lesson Key points

Ø The disk controller interface governs the maximum data transfer rate. IDE, SCSI and ESDI are roughly twice as fast as the traditional ST-506 interface.

Ø The speed of a disk drive depends mainly on its access time. Drivers with access times of 25ms or less are generally considered fast.

Ø IDE drives are good choice for all single-user machines that need disks up to 500 Mbytes. Over 500 Mbytes, and for network file servers, the choice is between ESDI or SCSI.

Ø ESDI drives are simple and standardized. SCSI drives are more sophisticated and so potentially higher performing but they are less standard.

Ø Sector interleave is used to match the data transfer rate of the controller and processor to the maximum transfer rate of the disk. Too little sector interleave is far worse than too much.

Ø A catching disk controller provides the ultimate in disk performance and it is the natural partner for MCA or EISA bus machines.

Ø Software disk catching, or the use of a RAM disk, provides useful speed gains in disk access but uses main memory.

Disk fragmentation slows disk operations but can be avoided by using a defragmentation utility. Incipient disk errors can also cause a disk to slow down

One phenomenon that many users of high performance computer notice is the gradual deterioration in the speed of access of their disk drives over time. Many think that this is a psychological effect, other think that it is to do with the disk being too full. Neither of theses explanations is correct the slowing of disk systems with use it due to fragmentation. When you start using a hard disk there is generally plenty of unused space available in large contiguous chunks. This allow the operating systems to allocate storage to files, even large files, so that they are stored in one place of the disk. A file that is stored using sectors that are contiguous is fast to read and write.

However as the disk is used free space is used up (and also created) by the deletion of files. Slowly but surely the free space becomes broken up across the disk’s surface and eventually it isn’t possible for the operating systems to find chinks of free space large enough to store a file in one place. In this case the file will be stored using two or more chunks of free space – i.e the file is stored using non-contiguous sectors. This makes the file slower to read and write because the drive’s heads have to be moved across the surface of the disk to read the entire file . this is fragmentation and its effect can be a dramatic slow-down in the operation of a disk. You van often tell a badly fragmented disk by the sound or the head moving from track to track as a file that is scattered across the surface is retrieved. It gives the impression that the drive is working hard for simple file operations.

The solution to the problem is to reorganize the use of the disk so that fragmented files are stored using contiguous sectors and the free area of the disk is again composed of contiguous sectors. The reorganization of the existing files results in an immediate improvement in the read time and the newly contiguous area of free space means that fragmentation will not occur again when files are created. Thus defragmentation improves , both read and write times back to the original performance of the disk as first delivered. Of course , as the disk is used and files are created and deleted fragmentation will slowly develop but another session defragmentation soon cures that.