The IDE channels on my motherboard were shot, so the computer store installed a Promise controller (along with a new power supply). Performance is great once I’m in Windows, but it takes about three times as long to load.
Is this normal? Is there anything I can do to optimize this?
(If anyone says “Yeah, huh huh, get a Mac” I’ll ram you so hard your cherry will come out your throat. Never cross a queen, children.)
Performance in Windows won’t be affected all that much, assuming you have sufficient RAM for the majority of the O/S to load there. Without knowing the absolute specifics of the Promise controller, I will make the following educated guess:
The controller on the system board was Ultra-DMA/66 or somesuch with a direct pipe to the processor, and the Promise is NOT Ultra-DMA OR is at the least slower because it is in a PCI slot without a direct pipe to the CPU.
You may well had had to change your BIOS settings to boot from SCSI for your system to recognise the controller card.
I had an ABIT dual IDE card and this was the same.
You will also get an addition to your boot sequence, where it has to load the IDE card BIOS before it can ‘see’ the hard drive.
Apart from boot up everything will operate the same though.
Sigh … page long response wiped by power outage -
Short mode
1: Promise ATA/100 finicky, slow loading PITA I go rid of. New Promise ATA/133 TX2 very good and loads fast.
2: Defeat onboard ATA controllers in BIOS
3: Defeat onboard ATA controllers in windows system hardware list
4: Check you have latest BIOS for you ATA card. See
http://www.promise.com/support/support_eng.asp
5: Check you have latest BIOS for your MB
6: Card being in PCI slot vs being onboard will make no difference to real world throughput requirements once drives spin up as data bandwidth pipeline in either case is bigger than drives will ever need for normal use.
7: A 10-30 second delay is within acceptable range for the system to poll available controllers then default to PCI ATA card. If delay is a minute or two the system is likely slightly confused and having to parse out where the boot drive is located. That’s why you need to defeat the non-operational controllers on as many levels as possible.
Also… make sure you have an ATA 10/133 compatible cable attaching the hard drive(s) to the ATA card. The newer cables are stiffer and you can see more slender wires in the cable vs the older 40 wire types.
Thanks, astro. It’s a Promise ATA/100 TX2. We needed a cheap solution and fast, because without it, we were without a PC period. Checked the device manager versus the Promise web site, and the drivers and BIOS versions are the most recent. I’ll check for a newer BIOS for the mobo, however.
Like I said, Windows runs great once I’m in it, which I suppose is relative to the way it was running with screwy IDE on the mobo. It just takes Windows quite a while to load. No biggie, since we leave the PC on all the time when we’re home.
If Windows were acting sluggish once loaded, then I’d be concerned.
Also, with the initial drive detection sequence on startup, it detects the CD-ROM as primary master (WTF?), then goes to the “new” screen and detects the hard drive. I’d open up the box to find out what exactly is up with the CD-ROM, but my other half would kill me. (He gets nervous and upset whenever I have to open up the box.) But the CD works great, so…
I don’t know what they did at the computer store, but the system’s working, so I’m not going to complain too much.
Can you clarify:
Thanks.
In the Motherboard BIOS setup section which you can normally get to by holding down or tapping a specific set of keys(s) at boot, there is often a sub-section where you can defeat or deactivate certain onboard MB components like the serial and parallel ports, USB and drive controllers (among other things) if there are conflicts or you are using outboard controllers (your situation) . Even if you do not de-activate these onboard components the outboard PCI component will work in most cases, but there will often be a signficant delay because the BIOS (and then windows) has to figure out that there are no available boot drives on the onboard controller, and it will sometimes poll these component hardware address locations repeatedly before it gives up and looks elsewhere for an alternative controller, which can take time.
Depending on the verison of windows there is usually a windows list of detected hardware in the “system>hardware>device manager” applet under “control panel”. Windows can/will sometimes look to these non-active controller addresses even if the OS is being booted off of an outboard controller. To minimize windows confusion as to where the main controller addresses and drives are located you should use the “properties” options for these components to deactivate them so widows will not poll that location.
As a side note your Promise controller should have two IDE slots allowing for two drives each (using a standard 2 header IDE cable) or 4 drives total. If you are using one of the two onboard controller channels for controlling the CD you will need to keep that onboard controiller address active or lose the CD.
Try plugging the CD into the second channel of the Promise outboard controller and defeating all onboard controllers, and see if that speeds things up as this will eliminate some of the the boot sequence confusion your system is having to figure out at boot up.