Asus crippled ICH10 performance on their X58 boards.

[CodeNinja]

well if the PCI bus isn't the reason for decreased performance, any ideas on what is?

My guess would be something in the BIOS.

Although, honestly, I can't imagine a lot of people who are dropping the big bucks on SSD arrays are crippling them by running them on the onboard soft-RAID anyway.

Be curious to see what my 1+0 WD Black array does, when I get it together later this weekend.

[TimoneX]

Information on the ICH's connection on the x58 is readily available in concise form here:

http://en.wikipedia.org/wiki/Intel_X58

[Ender17]

My guess would be something in the BIOS.

Although, honestly, I can't imagine a lot of people who are dropping the big bucks on SSD arrays are crippling them by running them on the onboard soft-RAID anyway.

Be curious to see what my 1+0 WD Black array does, when I get it together later this weekend.

ICH10R is better for SSDs than many hardware RAID cards

[CodeNinja]

ICH10R is better for SSDs than many hardware RAID cards

Really? That seems counter intuitive to me. Good info to know, though.

[cstkl1]

just curious
whats ure everest buffered scores are

my velociraptor raid 0 is reading on 7800-8000mb/s

[cstkl1]

hdtach burst speeds are around 4k.

edit i think i know why ts having a poor result in his x58

dude enable volume write back

my result on burst especially
8k with write back
440mb with disabled

[XS Janus]

Looks like a cache benchmark to me...

[TimoneX]

ICH10R is better for SSDs than many hardware RAID cards

How so? Which SSDs?

[One_Hertz]

How so? Which SSDs?

Intel X25-E for example.

[TimoneX]

Intel X25-E for example.

I figured you'd say that. I saw one of your own reviews that showed large gains in some operations with hw raid. I seem to recall you stating that the controller added latency though. My HW raid setup is adding latency as well...a whopping .2ms of it. .4 vs .2 according to HDTune. I haven't tried it on the ICH10R yet. These are cheapy patriot warps though($50 each). The Intel XM25-E is the best SSD available IMO...unfortunately I am not willing to spend $1000+ on a couple of them.

Not arguing with you btw...never used the extremes. I'm a cheap bastahd!

I did use OCZ solids on both ICH10R & the adaptec 3085 and it simply wasn't acceptable. Never entirely got rid of the stutter. These warps are lightning fast and very smooth. HW raid and cheapy SSDs are in fact a viable alternative to pricey(quality) SSDs IMO.

[archy121]

*removed*

[Stealth187]

So what happened to this discussion? Was there any resolution?

I've discovered the degraded performance on my x58 Rampage 2 Extreme with Intel G2 80gb SSD. Its mainly 4k read/writes but results are down across the board compared to ICH7R.

Using an old bios (0804) I get proper performance but that is hardly ideal.

[Andrea deluxe]

my 3 disk in raid0 stripe 128kb.

first test cache off

second test cache on

[HydrogenAlpha]

I am not sure if this information alone can help us get to any conclusions, that's why I insisted on owners of fast SSD storage.
The idea behind this, is to post some numbers involving small reads/writes (where the difference is more obvious) along with your motherboard and storage specs and in combination with information from Everest (or any other program that provide such information).
This way we could cross reference the results with existing ones on the net and see whether your X58-based ICH10 offers similar or seriously degraded performance.

Hey Chosen,
I have 3x 160GB G2 drives in RAID 0 on the ICH10R of a P6T deluxe board. Would you like IO stats on this array? I know you've already got P6T results so there's no point in posting unless you think it would be helpful.

[HDCHOPPER]

Please do I just got a Asus P6T SE lol ....

[uOpt]

There is no way that Intel put the disk controller behind the traditional PCI bus. It just looks like it to random software utilities walking what is presented to them looking like PCI busses.

However, I remember that a similar problem existed with 975x versus 965p. Clocking and features aside, a couple things were just going quicker on the cheaper chipset, which also came later.

Possible reasons could exclude:

• More checksumming or other error correction or other safety going on in the disk subsystems in x58. Intel has a history of putting all the safety features only into the more expensive board chipsets.
• Same as above, but on a Windows driver level (anybody tried Linux?)
• In general, driver differences between the two chipsets' Windows drivers.

My money is on simply faster microarchitecture on the simpler chipset. (did it also come later?)

Not having sample boards I have no way of investigating, nor am I interested given general Intel antics lately (ECC only with Xeon and on some boards? Time to try AMD again).