Hardware vs. Software Raid Speed

Hardware vs. Software Raid Speed

[Marc Perkel]

Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0 
striping on the motherboard. Just wondering if hardware raid (SATA2) is 
going to be faster that software raid and why?

Re: Hardware vs. Software Raid Speed

[Andre Tomt]

Marc Perkel wrote:
> Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0 
> striping on the motherboard. Just wondering if hardware raid (SATA2) 
> is going to be faster that software raid and why?

Beeing a consumer type board (AM2), the "raid on the motherboard" is in 
99.999% of the cases just software raid implemented in their Windows 
drivers, a bootup setup screen plus some BIOS magic to get the OS booting.

Re: Hardware vs. Software Raid Speed

[Alistair John Strachan]

On Wednesday 23 August 2006 08:25, Andre Tomt wrote:
> Marc Perkel wrote:
> > Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0
> > striping on the motherboard. Just wondering if hardware raid (SATA2)
> > is going to be faster that software raid and why?
>
> Beeing a consumer type board (AM2), the "raid on the motherboard" is in
> 99.999% of the cases just software raid implemented in their Windows
> drivers, a bootup setup screen plus some BIOS magic to get the OS booting.

MD has so many benefits anyway that it doesn't make sense to use anything but 
the finest hardware RAID.

For starters, with Linux MD, you can RAID partitions on a disk independently, 
even with different RAID types, and you can port your array to another 
machine without any reconfiguration, on completely different hardware. You 
can also RAID two different technologies, for example PATA and SATA, hot-add 
spares, run with a deliberately failed array, build incomplete arrays (nice 
if you're just about to start a RAID5 but don't have enough discs), the list 
goes on..

-- 
Cheers,
Alistair.

Final year Computer Science undergraduate.
1F2 55 South Clerk Street, Edinburgh, UK.

Re: Hardware vs. Software Raid Speed

[Tejun Heo]

Andre Tomt wrote:
> Marc Perkel wrote:
>> Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0 
>> striping on the motherboard. Just wondering if hardware raid (SATA2) 

SATA2 has nothing to do with hardware RAID.

>> is going to be faster that software raid and why?
> 
> Beeing a consumer type board (AM2), the "raid on the motherboard" is in 
> 99.999% of the cases just software raid implemented in their Windows 
> drivers, a bootup setup screen plus some BIOS magic to get the OS booting.

And, yeah, they're all software RAID.  Also, there isn't much to be 
gained from making RAID0/1 hardware.  The software overhead isn't that 
big.  For RAID5, having XOR done in hardware helps.

-- 
tejun

Re: Hardware vs. Software Raid Speed

[Marc Perkel]

Tejun Heo wrote:
> Andre Tomt wrote:
>> Marc Perkel wrote:
>>> Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0 
>>> striping on the motherboard. Just wondering if hardware raid (SATA2) 
>
> SATA2 has nothing to do with hardware RAID.
>
>>> is going to be faster that software raid and why?
>>
>> Beeing a consumer type board (AM2), the "raid on the motherboard" is 
>> in 99.999% of the cases just software raid implemented in their 
>> Windows drivers, a bootup setup screen plus some BIOS magic to get 
>> the OS booting.
>
> And, yeah, they're all software RAID.  Also, there isn't much to be 
> gained from making RAID0/1 hardware.  The software overhead isn't that 
> big.  For RAID5, having XOR done in hardware helps.
>

Thanks - I suspected that Raid 0 didn't gain anything in hardware unless 
they provided additional buffering or something but I just thought I'd 
ask in case there was something I was overlooking.

Re: Hardware vs. Software Raid Speed

[Mario 'BitKoenig' Holbe]

Tejun Heo <htejun@gmail.com> wrote:
> And, yeah, they're all software RAID.  Also, there isn't much to be 
> gained from making RAID0/1 hardware.  The software overhead isn't that 

The CPU cycles in fact don't usually matter. The I/O overhead (on the
PCI bus) due to multiple transfers of the (more or less) same data is
typically more interesting.


regards
   Mario
-- 
The secret that the NSA could read the Iranian secrets was more
important than any specific Iranian secrets that the NSA could
read.                           -- Bruce Schneier

Re: Hardware vs. Software Raid Speed

[Joel Jaeggli]

Marc Perkel wrote:
> 
> 
> Tejun Heo wrote:
>> Andre Tomt wrote:
>>> Marc Perkel wrote:
>>>> Running Linux on an AMD AM2 nVidia chip ser that supports Raid 0
>>>> striping on the motherboard. Just wondering if hardware raid (SATA2) 
>>
>> SATA2 has nothing to do with hardware RAID.
>>
>>>> is going to be faster that software raid and why?
>>>
>>> Beeing a consumer type board (AM2), the "raid on the motherboard" is
>>> in 99.999% of the cases just software raid implemented in their
>>> Windows drivers, a bootup setup screen plus some BIOS magic to get
>>> the OS booting.
>>
>> And, yeah, they're all software RAID.  Also, there isn't much to be
>> gained from making RAID0/1 hardware.  The software overhead isn't that
>> big.  For RAID5, having XOR done in hardware helps.
>>
> 
> Thanks - I suspected that Raid 0 didn't gain anything in hardware unless
> they provided additional buffering or something but I just thought I'd
> ask in case there was something I was overlooking.

A hardware raid controller can buy you a battery backed write cache, so
there are potentially some performance/safety benefits potentially.

We have a mix of software raid and 3ware based hardware raid subsystems.
For the applications we support, in general I can say that performance
was not the discriminating reason to choose one over the other.

> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

-- 
------------------------------------------------------------------------
Joel Jaeggli             Unix Consulting              joelja@uoregon.edu
GPG Key Fingerprint:   5C6E 0104 BAF0 40B0 5BD3 C38B F000 35AB B67F 56B2