Raid too underrated?

orangat

Golden Member
Jun 7, 2004
1,579
0
0
I know that raid0 is generally not worth the effort for typical home users and enjoy little support on these forums especially after the AT and SR articles. But has the pendulum swung too far by simply dismissing raid0?

http://tweakers.net/reviews/515/2
http://tweakers.net/reviews/515/4
Tweakers has a large pool of benches which totally contradict the AT's benchmarks on game loading. Could onboard raid controllers like the ICH5/promise vs real hardware raid be a big factor?

How much cpu utilization do current onboard raid controllers like nforce3/4/, VIA,Promise,Sil consume? Non-raid sata typically takes up 4-12% during sustained read/writes.
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
I for one have never understood the opinion of RAID arrays on the forums here. People seem to believe that RAID yields little to no benefits unless you are constantly doing large transfers. Thus, you are operating your storage system at near capacity. But people don't think twice about OCing, but how often do people run the processor(s) at near capacity? It makes no sense to me.
 

PurdueRy

Lifer
Nov 12, 2004
13,837
4
0
Originally posted by: Fullmetal Chocobo
I for one have never understood the opinion of RAID arrays on the forums here. People seem to believe that RAID yields little to no benefits unless you are constantly doing large transfers. Thus, you are operating your storage system at near capacity. But people don't think twice about OCing, but how often do people run the processor(s) at near capacity? It makes no sense to me.

Pretty much any game uses the full power of your CPU...
 

Ike0069

Diamond Member
Apr 28, 2003
4,276
2
76
I don't even bother with RAID simply becuase the price and risk involved does not makes sense to me. This is from a gamer only pont of view however.

First I must buy a second HD that otherwise I do not need. Then I double my chance of losing a main HD. Sure I get a little bit of speed increase, but I'd much rather put the extra money into a VC upgrade.
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Originally posted by: PurdueRy
Originally posted by: Fullmetal Chocobo
I for one have never understood the opinion of RAID arrays on the forums here. People seem to believe that RAID yields little to no benefits unless you are constantly doing large transfers. Thus, you are operating your storage system at near capacity. But people don't think twice about OCing, but how often do people run the processor(s) at near capacity? It makes no sense to me.

Pretty much any game uses the full power of your CPU...

When I go from BOINC (hi InlineFour) to Guild Wars, my CPU usage goes from 100% to ~38% average.
 

PurdueRy

Lifer
Nov 12, 2004
13,837
4
0
Originally posted by: Fullmetal Chocobo
Originally posted by: PurdueRy
Originally posted by: Fullmetal Chocobo
I for one have never understood the opinion of RAID arrays on the forums here. People seem to believe that RAID yields little to no benefits unless you are constantly doing large transfers. Thus, you are operating your storage system at near capacity. But people don't think twice about OCing, but how often do people run the processor(s) at near capacity? It makes no sense to me.

Pretty much any game uses the full power of your CPU...

When I go from BOINC (hi InlineFour) to Guild Wars, my CPU usage goes from 100% to ~38% average.

depends if your game is CPU limited or GPU limited
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Originally posted by: orangat
Any comments on the tweakers benchmarks?

Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...
 

JEDIYoda

Lifer
Jul 13, 2005
33,986
3,321
126
Originally posted by: Ike0069
I don't even bother with RAID simply becuase the price and risk involved does not makes sense to me. This is from a gamer only pont of view however.

First I must buy a second HD that otherwise I do not need. Then I double my chance of losing a main HD. Sure I get a little bit of speed increase, but I'd much rather put the extra money into a VC upgrade.

For gaming I use raid0. But as all good raid users do...I also don`t keep anything important on my gaming rigg!

This is from a gamers perspective!!
 

orangat

Golden Member
Jun 7, 2004
1,579
0
0
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Any comments on the tweakers benchmarks?

Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...

What raid controller are you using?. Writes on raid5 should be slower than raid0.
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Originally posted by: orangat
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Any comments on the tweakers benchmarks?

Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...

What raid controller are you using?. Writes on raid5 should be slower than raid0.

Broadcom BC4852 with 6 Hitachi 250gb SATA (RAID 5) and two Hitachi 160gb (RAID 0).
 

orangat

Golden Member
Jun 7, 2004
1,579
0
0
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Any comments on the tweakers benchmarks?

Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...

What raid controller are you using?. Writes on raid5 should be slower than raid0.

Broadcom BC4852 with 6 Hitachi 250gb SATA (RAID 5) and two Hitachi 160gb (RAID 0).

You have 6 disks on raid5 but only 2 for raid0, its hard to compare both using your setup.
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Originally posted by: orangat
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Originally posted by: Fullmetal Chocobo
Originally posted by: orangat
Any comments on the tweakers benchmarks?

Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...

What raid controller are you using?. Writes on raid5 should be slower than raid0.

Broadcom BC4852 with 6 Hitachi 250gb SATA (RAID 5) and two Hitachi 160gb (RAID 0).

You have 6 disks on raid5 but only 2 for raid0, its hard to compare both using your setup.

:confused:
Won't load here at work. Very much looking forward to seeing them when I get home. From the benchmarks I've done at home, my RAID 5 performance was on par with a RAID 0 array of the same number of disks (6) on the same controller, so I'll be sticking with RAID 5...

The 6 / 2 is what I'm running now. I did testing with the 6 drives in RAID 0, and then converted the array to RAID 5. Yay online RAID migration.
 

orangat

Golden Member
Jun 7, 2004
1,579
0
0
Gotit.

Anyway no-one seems to notice tweakers benchmarks differ wildly from AT's? The content creation benchmark jumped from 13% -> 50% after switching to another model of pci raid controller.