Yes, but performance might be an issue (although only likely to noticeably hit you if the slave gets a lot of use - for example if your swap partition resides there, or if you try including it in some sort of mirroring arrangement) - with IDE, only the master or the slave on a single channel can be accessed by the system at any one time. Also if the drive controller on the master device fails, the slave will not be accessible either (although the data on the slave and the slave drive+controler etc itself will be intact)
Sure. I suggest making it the seconday master and putting the optical as the secondary slave. You can run as many drives, both physical and logical (and even network and virtual), as you want, though I’ve found that things start to get cumbersome when you’re up to about a dozen drive letters.
Do you have a connection for Serial ATA on your motherboard? IDE is on its way out. If you have the space in the case and a serial ATA connection, I would suggest getting a serial ATA hard drive if it’s not much more. It’ll be usable for longer and you won’t even have to disconnect the second optical drive.
Yes. Or you could get a IDE controller card and keep what you have, and still have more places to hook up drives. …Figure $35 for a 2-channel, $50 for a 4-channel card. -You may want/need a controller card anyway, if you find that two drives sharing the same cable run really, really slow–a controller card will totally cure this problem, and often, it is the only way to cure this problem.
You can get SATA controller cards too, but CD/DVD drives are not available in SATA configs, last I saw, so the SATA card would support additional hard-drives only. SATA interfaces have higher data-dransmission speeds, but so far there is no drive that can send data that fast anyway. So for now IDE works just as fast as SATA.
~
Good point. I was looking at Seagates drive specs today (they are the only one of the big three that publish real drive specifications written by engineers and not marketing morons. The Barracuda 7200rpm 400GB drive has a maximum possible sustained data transfer rate (reading from the outer cylinder, where the bits are most tightly packed, angularly) of 65 megabytes/second. Makes you wonder why the industry kept pushing interfaces faster and faster when the drives simply aren’t capable of operating any faster than the speeds provided by ATA66. Even two ultra-high performance drives can’t saturate a ATA133 bus, even if ATA drives could stack the bus. Now we have SATA150 and only one drive per channel. WTF?!
There’s one case that I can think of where the higher bus speed can help, and that’s if the application requests something that happens to be in the hard drive’s RAM cache. The typical high-performance hard drive has 8MB of cache, which takes ~120ms to pull across the bus at 66MB/sec, but only ~50ms at SATA’s 150MB/sec. The difference in instantaneous responsivity can be noticeable, but as you say it drops off when you look at sustained performance and the data’s being pulled off the platters all the time rather than being cached.
That’s exactly my point. The biggest advantage a SATA bus will ever have over an ATA66 bus is 70ms on a single transfer.
And most modern systems are dedicating a lot more than 8MB to drive caching making it extremely unlikely that data is in the drive’s cache and not in the OS’s cache.
actually, some manufacturers have been coming out with optical drives that are SATA - here is a plextor DVD burner (I know it is a burner, not just a DVD drive, but a burner is technically an optical drive as well):
I am pretty sure I have seen CD-ROM SATA drives as well, but I admit that SATA optical drives are few and far between.
Well, I must be mistaken RE: SATA CD-ROMs - the only SATA optical drive I can find is the plextor, and a quick google search leads me to believe that that plextor is the only SATA optical drive commerically available.
I’ve been led to believe that an IDE channel will operate at the speed of the slowest device plugged into it - since the vast majority of optical drives have a 33MHz interface, a hard drive on the same channel will also be forced to run at ATA33 speeds, probably a third of its nominal capacity. Digging around Google I can find lots of references to this alleged fact, but none of them what I’d describe as truly authoritative. The best I can find is this from Microsoft, which indicates that it may be true, depending on your IDE controller:
So I guess it depends whether your motherboard has a posh IDE controller on it. Finding this sort of technical detail could be a bit tricky, though…