Click to See Complete Forum and Search --> : ATA-100 onboard slower than add-on card?
bwkaz
03-13-2001, 08:26 AM
Is onboard ATA-100 (on the motherboard, that is -- a la the i815 chipset or similar) significantly slower than an ATA-100 add-on card (like a Promise or similar)? My roommate swears the onboard controllers are slower, but neither he nor I have ever seen any benchmark results to prove or disprove it.
I'm thinking, hey, they're both on the 33MHz PCI bus, they both have an advertised 100MB/sec burst rate, so how can one be faster? Maybe the sustained rate is faster, but.....
Might it have something to do with the onboard BIOS on the PCI cards?
Basically, does anyone have benchmarks comparing ATA-100 add-on cards to on-board solutions (with the same drive, obviously)? Or does anyone know where I could find them?
I'm not planning on upgrading anything, I'm just wondering.
Thanks to all,
Bryan
bwkaz
03-14-2001, 05:53 AM
^^ bump ^^
Anyone?
Bryan
AuraEdge
03-14-2001, 12:38 PM
I think they only thing that can seperated the two are drivers. The Onboard ports are essentially on a virtual PCI bus, which gives it the same bandwidth as a card you would plug into a PCI slot.
Peter M
03-14-2001, 03:21 PM
In case of the Intel 815, VIA's DDR chipsets and all SiS chipsets, the integrated IDE controllers perform better (or at least hog the system less) - because they're on a proprietary high performance bus, not on the 33 MHz PCI bus.
For VIA's older chipsets, where the 686B or 8231 south bridge sits on the PCI bus just like anyone else, performance should be about the same, minimal differences possible, but not worth mentioning.
Regards, Peter
bwkaz
03-14-2001, 04:05 PM
Well, Peter, that's interesting.
I figured they'd be for all practical purposes the same, I definitely did not expect the onboard setup to perform better.
Makes sense though.
Bryan
[This message has been edited by bwkaz (edited 03-14-2001).]
bwkaz
03-16-2001, 08:50 AM
Confirmed. Sandra's "drive index"es:
Add-on Promise card (used my roommate's because his hard drive isn't here yet -- he had ATA-66 on his motherboard), with only my 7200rpm ATA-100 drive as the IDE1 master: 18065.
On-the-motherboard ATA-100 controller (with my 7200rpm ATA-100 HD as the master, and an ATA-33 hard drive as the slave): 18605.
So the onboard does actually outperform the Promise card, by about 3%.
Edit: I ran each test the first thing after I booted Windows. I also ran a test of the onboard ATA controller after I had left Windows running all night, and it gave me somewhere around 16800 or so.
Bryan
[This message has been edited by bwkaz (edited 03-16-2001).]
Wilan Wong
03-18-2001, 02:45 PM
Just wondering if ATA-100 helps the score any bit?? Because I own a WD 20Gb 7200RPM Drive, that is running on ATA-66 and I usually get around 22000 running on latest version of SiSoft Sandra.
bwkaz
03-18-2001, 03:23 PM
Wow, that is quite a bit better. How much cache does the disk have? Mine (I think) has 2MB.
Also:
P3 800 (133x6) at 800 (so not O/Ced)
256MB PC133 RAM
256K L2 cache, 32K L1
Maxtor HD, I don't remember the part number
Edit: Yep, 2MB 100MHz cache, it's a Maxtor 52049H4.
Bryan
[This message has been edited by bwkaz (edited 03-18-2001).]
SysOpt.com
Copyright Internet.com Inc. All Rights Reserved.