Confirm Barracudas IV low RAID performance?

[<Grisu4>]

Hi,

I already have a Seagate Barracuda IV (60) und and PCI Raid Contoller card (HPT370). I think about building a raid 0 for more performance.

Some people say that the barracudas have very low raid o performance ...

Can you confirm this?

[Henry Rollins II]

No, but I can deny it. If the prive performs good as it is, I will do fine in a RAID setup too.

regards,
henry

[Jon]

Actually, I've seen many reports from different sites that vouch for poor performance when using those drives in a RAID array.

The irony of it is, is that they have a higher throughput than the controller chip of the current RAID cards can handle, basically flooding them and causing them to lag.

I'm not sure if it's a problem with HPT-based controllers, but I've seen many cases of this happening with Promise cards.

Says alot about their individual performance though. They are terrific drives.

[Henry Rollins II]

Originally posted by Jon
Actually, I've seen many reports from different sites that vouch for poor performance when using those drives in a RAID array.

The irony of it is, is that they have a higher throughput than the controller chip of the current RAID cards can handle, basically flooding them and causing them to lag.

I'm not sure if it's a problem with HPT-based controllers, but I've seen many cases of this happening with Promise cards.

Says alot about their individual performance though. They are terrific drives.

Due to a _higher_ output? How can this be? I was under the impression that all IDE-interfaces were backwards compatible(ie A slower drive on a faster interface and vice versa), even those on the raid controllers. An ATA-133 disc works with a ATA-100 interface. There must be some kind of standard of how much data a specific interface can handle, otherwise it would be impossible to buy a new harddrive for an older motherboard.
If the interface gets flooded it must be due to the harddrive that exceeds the given standards. Think about it - RAID is a storage solution that increase performance and(in the PC case: or) security for systems that handle enormous amounts of data. Wouldn´t you think it would be weird for high-end storage controller cards to be _flooded_?!?

I don´t think it sounds right - doesn´t Seagate offer some fix for this?

regards,
Henry

[Ice_Gargoylle]

do pci cards work well? doesnt the pci interface slower than a onboard controller? a pci is what 33mhz?

[Jon]

Makes no sense to me either. I'm just repeating what I've seen on at least 3 other sites. Users are just complaining about a bad lag when it comes to heavy disk access when using those drives in a RAID 0 array.

Haven't seen any news on a fix either. Most have just been telling them to trade them off for different drives. I'd like to know more than I do on it though...just sounds weird.

[Henry Rollins II]

Originally posted by Ice_Gargoylle
do pci cards work well? doesnt the pci interface slower than a onboard controller? a pci is what 33mhz?

I tried finding out what speed the integrated IDE controller of the AMD-768 chip(I guess it is about the same for other chipsets) works at, but I can´t find it. My guess that it actually IS a 33 Mhz PCI controller, integrated in the chip. Harddrives have really slow data transfer rates, compared to internal computer components. I don´t think anything faster than PCI is needed.

regards,
Henry

[Jon]

Onboard controllers and PCI controllers work at 33MHz as they are both operating on the PCI bus.

The only exception is the 66MHz/64-bit slots that high-powered workstations and server boards have. They require specially keyed PCI devices that operate at different voltages as well.

The newer MPX boards that are coming out by AMD have these 64MHz/64-bit slots that double the available bandwidth of the more common PCI slots. The Promise TX2/TX4 series of controllers are the most cost-effective solutions for IDE RAID that take advantage of the PCI66MHz throughput. The downside is that the hard drives still are limited to their respective transfer rates meaning no actual performance increase.

The MPX chipset operates at a 66MHz frequency which is an improvement over the MP chipset of 33MHz due to the newer PCI bus. However, if you use those slots with PCI33MHz devices, it will lower that bus to 33MHz overall, decreasing performance. Does not matter if it's 32-bit or 64-bit though...it is able to switch between the two depending on the device.

[<Grisu4>]

So what do you suggest?

Buying another one and give it a try, or selling the one I have ..... actually - I like the drive - it's really fast and desperately silent.

[Jon]

Hard to say. I'd hate to tell you to sell it because I know they're really good drives.

I would go ahead and get another and stay away from the Promise controllers that seem to be causing the most trouble with those drives and see what happens.

[<Grisu4>]

I have a Highpoint, HPT370 PCI Controller Card to which the first Barracuda is already connected and performs perfectly .... and - I have to tell it again - she's so silent .... ;-)