RAID and SSD benchmarks

[FUGGER]

Here is something you will not see on a review site.

I checked out a <A HREF="http://fugger.netfirms.com/bitmicro.jpg" target="_new">BiTMICRO SSD</A> (solid state disk) this weekend with various controllers. <A HREF="http://fugger.netfirms.com/2100s.jpg" target="_new">Adaptec 2100S RAID controller</A> and Symbios Ultra3.

Strangly the Symbios card rated at 80MB/s was faster than the SCSI 160 Adaptec card. so I completed my tests with the Symbios card.

The SSD is a 68pin LVD drive, 34MB MTR in this configuration 67MB MTR with fibre controller.

Sandra 2002 scores with <A HREF="http://fugger.netfirms.com/reportme.htm" target="_new">windows ME</A>
Sandra 2002 scores with <A HREF="http://fugger.netfirms.com/reportxp.htm" target="_new">Windows XP</A>

As promised, scores from my HPT raid.
Sandra 2002 score with <A HREF="http://fugger.netfirms.com/reporthpt.htm" target="_new">HPT Raid 0</A>

Where the SSD owns is Random read/write and access time. I would need to step up to FCA (fibre channel) to get 67MB MTR (max transfer rate) Retail value of this SSD is $14,000 US. these drives in FCA RAID 0 dominate. Mhz of host had no effect on transfer rates like I hoped. Formatting with largest block size had the most performance as I was able to reach the rated 34MB MTR in this configuration.


Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[priit]

<i>Strangly the Symbios card rated at 80MB/s was faster than the SCSI 160 Adaptec card</i>

Maybe that's due to the bug in the i850 chipset thad limits the PCI bandwidth to about ~80Mb/s ?

 

[FUGGER]

Nope, no bug. Symbios card was just faster in this configuration. Im sure if I set up RAID 0 the Adaptec card would be superior.

Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[juin]

Maybe that's due to the bug in the i850 chipset thad limits the PCI bandwidth to about ~80Mb/s ?

Are you sure that I850 have a bug with PCI never hear of that or maybe you want to talk about the via who need the PCI Lantencie 0.19 pacth

http://service.madonion.com/servlet/Index?pageid=/orb/projectcompare&projectType=6&projectId=2310900

 

[Kelledin]

Are you sure that I850 have a bug with PCI never hear of that or maybe you want to talk about the via who need the PCI Lantencie 0.19 pacth

Yes, we're sure. It was linked a while back on the [H]ardForums; I'll have to dig up the link. Basically the i850 has the same sort of PCI latency issue as the VIA southbridge.

<i>If a server crashes in a server farm and no one pings it, does it still cost four figures to fix?

 

[FUGGER]

Link please

Supprised you guys didnt pick at my temps or monitor.

Im adding another drive to my raid, im at 74mb buffered read now, If I should be able to break 80Mb/s would that snuff the bug theory?

I pulled a dual xeon from service today, the motherboard is going bad it seems. Ill strip the parts and reuse the case or possibly auction it off to fatburger since he was interested in one. This is a <A HREF="http://fugger.netfirms.com/reportXEON.htm" target="_new">Intergraph TDZ2000 Dual XEON with Realizm graphics engine</A>, this system is a RIP server (Raster Image Processor). if you check the core temps and vcore voltage you can see 1 is having a problem. I already checked the HSF are operational and clear of debris. This machine also has RAID 0 setup.

Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[Kelledin]

<A HREF="http://developer.intel.com/design/chipsets/specupdt/29071501.pdf" target="_new">http://developer.intel.com/design/chipsets/specupdt/29071501.pdf</A>

Erratum #5

<i>If a server crashes in a server farm and no one pings it, does it still cost four figures to fix?

 

[FUGGER]

Thats the I860 chipset, I use the I850. Its good reading but I know some of those problems are non existant on I850 like STR and NAP.

Im looking thru I850 ICH/MCH stuffs now

Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[FatBurger]

Temperature Sensor(s)
Board Temperature 2.0°C / 35.6°F
CPU Temperature 15.0°C / 59.0°F td

Freon or liquid nitrogen?

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>

 

[FUGGER]

With LN2 I would be at ~ -175c

Im using a salvaged Cray SV1 refridgeration unit, its about the size of a 12pack of soda and zero condensation. Nothing attached to the CPU or video card other than the HSF, just very cold air in the case.

I can hit -4c mobo and +6c CPU

Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[FatBurger]

Ah, that's right. I think you mentioned that in another thread.

Does a RAID 0 of 30k strike anyone else as absurdly low for a $14,000 SSD? Or am I reading it wrong (always a high probability).

<font color=orange>Quarter</font color=orange> <font color=blue>Pounder</font color=blue> <font color=orange>Inside</font color=orange>

 

[Kelledin]

<A HREF="http://developer.intel.com/design/chipsets/specupdt/298247.htm" target="_new">http://developer.intel.com/design/chipsets/specupdt/298247.htm</A>

Sorry 'bout that. This one's for the i850. Erratum #5, same problem.

<i>If a server crashes in a server farm and no one pings it, does it still cost four figures to fix?

 

[FUGGER]

The RAID 0 is a pair of Maxtor D740X J2 drives
The single SSD scored lower on the buffered read/write but owned in the random and seek.

The SSD is prime for database or anything with lots of access to the HD. This drive has extreme temp ranges, no noise, no moving parts, and a faster interface available. I was just testing the 68Pin LVD version. I see the SSD best used in RAID configurations. It would require at least 4 SSD's to max the bandwidth of the bus with a 1ms access time.

Proving once again that <A HREF="http://www.zombo.com" target="_new">anything is possible</A>.

 

[Ncogneto]

Send me a few over so I can ty them out in my FC-AL

It's not what they tell you, its what they don't tell you!