IDE to SATA adapter

[jaclaz]

As far as I can tell, you both agree that Ultra DMA 2 has a maximum transfer rate of 33.3 MB/s, so it doesn't seem that Ultrs DMA 4 is needed for CD/DVD use.

Sure , the speed of Ultra DMA2 is a fact (as per specifications), the question is if the speed transfer of a N x speed for a DVD-ROM reader/writer is in any way faster than that (I have no doubt that the CD is much slower at *any* multiplier).

If the Wikipedia data is accurate, submix8c's Lite-On 20x should top at 27.70 MB/s, some 17% slower than what the specs (and conversely the 40 wires cable) should allow, so what is/was the reason to have also UDMA 4 (and conversely 80 wires cables)?

jaclaz

[jumper]

The Nx speeds are for physical reads/writes. The read cache / write buffer can be accessed at much higher speeds (up to the interface maximum).

[jaclaz]

The Nx speeds are for physical reads/writes. The read cache / write buffer can be accessed at much higher speeds (up to the interface maximum).

Sure, but this doesn't make the device "faster", nor by itself justifies the *need* for a faster mode .

I mean, more or less a cache or buffer behaves as a funnel, the idea is to have a continuous flow on the narrow end no matter how "intermittently" the larger end is fed, i.e. AFAIK is all about "regularity" and not about "speed", possibly it becomes relevant with non-sequential reads, though I doubt it can deliver a data transfer higher than the "label" 33.3 of the bus.

Actual tests (these are VERY old, if anyone can provide more recent ones it would be nice) on 16x drives:

http://www.tomshardware.com/reviews/10-dvd-burners,911-14.html

show an actual peak transfer rate of around 16x i.e. according to the Wikipedia article 22.16 MB/s, with the very faster one with 16.16*1.350=21,816 and 16.16*1,385=22,38 (usual mess between using 1000 and 1024), i.e. the faster of the lot exceeds the "label" specification by 0.16/16=10 %exactly (and on peaks only).

So I don't think there is a "real world" *need* for the Ultra DMA4, the reason for it must be *something else*

jaclaz

[cdob]

the reason for it must be *something else*

http://en.wikipedia.org/wiki/Parallel_ATA#Speed_of_defined_transfer_modes

its maximum theoretical transfer rate on the cable

There are two devices per cable.

Imagine read and write to the same parallel cable.

The situation is different at a serial cable: one device per cable

Hence it's maximum theoretical transfer rate on the device.

Beside user data there are commands to control the device.

Driver, cable and hardware result to user data transfer rate:

the limit can be 20x DVD speed user data at a UDMA2 PATA SATA adapter.

A optical drive offers max speed at end of media only.

I doubt a importand real world limitation at UDMA2 PATA SATA adapter.

Contrary any adapter is a additonal risk:

at a average IDE optical drive: do not buy a adapter, buy a new optical drive.

at a very good IDE optical drive: try one or more adapter(s)

And remember power supply, does current PSU offers a relating power connector?

[submix8c]

OK... It has a 2MB Buffer and apparently

UDMA 4 is supposed to allow faster burst rates for one thing, and the transfer rate tests seem to confirm this - but like I said, I'm not sure if I'm seeing any great benefits from it. On the other hand, if my writing quality is not made worse, then there's at least no harm in it.

(http://www.cdrlabs.com/forums/gsa-h22n-and-udma-t22887.html)

Can anyone make heads or tails of this? Useful?

http://www.convertcenter.com/ide-udma66

edit - No difference in transfer rate?

http://www.hardwarecentral.com/showthread.php?36818-UDMA-33-vs.-UDMA-66-drives...

/OT

Edited November 23, 2012 by submix8c