Do DMA66 drives read slowly?

1 reply [Last post]
ajaques
Offline
Joined: Jun 20 1999

Hello all,
I use Fast's "Fastcap" capture program, which has a built in speed
test (called "Benchmark"). Using Maxtor "DMA33" drives I normally get
results of Write: 12MB/s and Read 11MB/s.
A month or two back I bought a 10GB Maxtor drive and used Maxtor's
software to "tell" it that it was a DMA33 drive. Write speed was
reported as 14MB/s, but Read was down to 9MB/s. I have since had the
chance to do FastCap benchmark tests on two other Maxtor drives; one
16GB and one 20GB, and they are both also reported as Write 14MB/s,
and Read 9MB/s. The same applies to FastCap tests of the drives on
two other computers.
Actually that speed is fast enough so I am not that bothered, but:

Does anyone know if "DMA66" drives are supposed to be slower on
read than true DMA33 drives?

If so is this generally true or just applicable to Maxtor drives?

If it is not supposed to be like this then can anyone suggest a
reason or a cure?

Thank you,
Tony...


duncan hancox
Offline
Joined: Aug 15 2000

the tests of drive speed used for video suitability rely on testing the sustained data transfer rate of the disk.

this is the rate that consecutive data can be read or written from the drives magnetic media.

since the media reading is the slowest point, the computer to hard disk interface has little or no effect on the measured data rate.

it doesnt matter if the computer can talk to the hard disk buffer at a 33megs/sec or 100megs /sec or even 1000megs/sec, the hard drive buffer can only talk to the magnetic media at a rate dependent on the rotational speed of the media and the density of data on the magnetic surface.

so, the computer to disk interface only becomes a problem if you are using a very slow interface that uses lots of computer power (such as the old PIO modes) or the data access is more random than video streaming, where the data buffer and cacheing functions in combination with the computer/disk interface can speed up disk access.

the bottom line is that the numbers measured are a function of the disk sustained media data rate (rotational disk and data density being the most important aspects)

there are many ways of measuring this, and no 2 disk test programs will come up with the same results

as long as the sustained data rate is fast
enough not to cause dropped frames then the numbers dont really matter at all!