eva2000

thanks for the results Eric

Tony, going to try the SSDs on different boards ICH9R/ICH10R as well as 3rd party raid controllers as well and do the MFT results near the end as their trial is time limited (you only get to run full trial for 4hrs AFAIK).

eva2000

Single 32GB OCZ Core SSD NTFS 4K, 16K, 32K, 64K
Vista Ultimate SP1 64bit

Swapped over to Vista Ultimate SP1 64bit for some single non-OS 32GB OCZ Core SSD tests of different NTFS cluster sizes between 4K, 16K, 32K and 64K.

• E8600 Q820A599 **0464
• CPU Cooling: Dtek Fuzion Rad Tower Box
• Gigabyte GA-EP45 Extreme rev 1.0 F6k bios
• 512MB Inno3D 8800GT OC Edition @700/1732/2000
• 2x1GB Team PC2-10400 C6 dual channel kit
• Memory cooling: 120x25mm Thermaltake 81cfm fan
• 320GB Seagate 7200.10 SATAII OS disk
• 4x 32GB OCZ Core SSD 2.5" solid state disks single disk mode
• Pioneer 216 DVD-RW
• 1200W Silverstone OP1200
• Windows Vista Ultimate SP1 64bit - vLite + post SP1 hotfix/updates

Notes:

• Remember my results are on Vista Ultimate SP1 64bit with latest hotfix/updates integrated via vLite
• I setup OCZ SSD Core as non-OS disks as i wanted to try write testings but Everest Ultimate 4.50.1494 beta write tests throws an error so below only read test suite tested. I previously set these OCZ SSD Core disks to raid mode on ICH10R and it seems I forgot to reset them to non-raid via Intel Storage Manager so not sure if that was the problem. Will retest later.
• Seems for this motherboard + Vista Ultimate SP1 64bit, 4k NTFS clusters was best for OCZ SSD Cores compared to 16K, 32K, 64K NTFS cluster size.

Settings:

• Indexing disabled
• Superfetch enabled
• System Restore disabled
• Enabled advance performance

ATTO Bench v2.34

4K default cluster on Vista Ultimate SP1 64bit had an even faster result than my Windows XP Pro SP2 Atto Bench but the 16K-64K Vista results were similar to my Windows XP Pro SP2.

320GB Seagate 7200.10 OS disk


OCZ SSD Core NTFS 4K vs 16K cluster


OCZ SSD Core NTFS 32K vs 64K cluster

HDTune 2.55

4K vs 16K (right)


32K vs 64K (right)

Everest Ultimate 4.50.1494 Beta

Read suite tests
4K cluster had highest random read (132.1MB/s), while 16K cluster clearly had a higher buffered read (163.9MB/s). 16K, 32K, 64K all had slight faster Average read access of 0.42ns versus 4K at 0.43ns.

4K followed by 16K (left) and 32K followed by 64K (right)

PCmark05 HDD Test

16K cluster just slightly won our due to higher XP Startup speed, followed by 32K, 64K and then 4K cluster although they're fairly close when you look at the break down individual HDD test suite sub tests.

320GB Seagate 7200.10 OS


4K vs 16K NTFS cluster


32K vs 64K NTFS cluster

Chiller

this is my setup with OCZ SSD V2 in raid 0 on a Intel SRCSAS18E controller :

eva2000

hey Chiller

your image seems to be missing ? maybe try uploading to imageshack.us ?

eva2000

IOMETER 64bit 4K Random Write (100%) / Vista Ultimate SP1 64bit
Single 32GB OCZ Core SSD tests

I don't have much experience with iometer so will just post both newkleer's outlined config results and Tony's config results and let the experts chew on the results

• E8600 Q820A599 **0464
• CPU Cooling: Dtek Fuzion Rad Tower Box
• Gigabyte GA-EP45 Extreme rev 1.0 F6k bios
• 512MB Inno3D 8800GT OC Edition @700/1732/2000
• Memory cooling: 120x25mm Thermaltake 81cfm fan
• 32GB OCZ Core SSD OS disk
• 3x 32GB OCZ Core SSD 2.5" solid state disks single disk mode (bios set with RAID)
• 640GB Samsung HD642JJ
• Pioneer 216 DVD-RW
• 1200W Silverstone OP1200
• Windows Vista Ultimate SP1 64bit

• Newkleer config tests on SSD2
• Tony config tests on SSD3

Newkleer 4k random write 100% config
vs
Tony 4k random write 100%
Only diff is maxdisk size on Tony's config is = 0



iobw.tst file sizes

• newkleer config = 10GB if you leave iometer to prepare then bench drive
• tony config = 1.24GB if you abort preparation of drive after 10 seconds

Newkleer config result on SSD2
vs
Tony config result on SSD3



Managed to get Everest Ultinmate 4.60 Write tests to run if the SSD disk is not formatted at all.



Atto Bench v2.34
- both 32MB and 256MB length tests



Regardless, to this day, i haven't experienced any stuttering or freezing or problems installing WinXP Pro SP2/SP3 and Vista Ultimate SP1 32bit or 64bit in single drive or raid 0 for my 4x32GB OCZ SSD Core.

eva2000

IOMETER 64bit 4K Random Write (100%) Vista Ultimate SP1 64bit 2x32GB OCZ Core Raid 0

This time converted SSD2 & SSD3 drives into a raid 0 128K stripe volume using Intel Storage Manger within windows and redid the Iometer 64bit 4K random write (100%) tests along with some Everest Utlimate 4.6, HDtune and Atto Bench 2.34 benchmarks. Will be using newkleer's outlined config

• E8600 Q820A599 **0464
• CPU Cooling: Dtek Fuzion Rad Tower Box
• Gigabyte GA-EP45 Extreme rev 1.0 F6k bios
• 512MB Inno3D 8800GT OC Edition @700/1732/2000
• Memory cooling: 120x25mm Thermaltake 81cfm fan
• 32GB OCZ Core SSD OS disk
• 2x 32GB OCZ Core SSD 2.5" solid state disks Raid 0 on ICH10R 128K Stripe
• 32GB OCZ Core SSD data disk
• 640GB Samsung HD642JJ
• Pioneer 216 DVD-RW
• 1200W Silverstone OP1200
• Windows Vista Ultimate SP1 64bit

Default volume write back cache is disabled, so enabled it by checking the 'enable advance performance box' in volume's properties tab or via right clicking the Raid 0 volume in Intel Storage Manager application.



Iometer 64bit - 4K random writes (100%)
2x32GB OCZ Core SSD raid 0 doubled the random 4k writes I/O per second and halved the average I/O response time but increased maximum I/O response time by ~50%.



Atto Bench 2.34 - 32M and 256MB length


HDtune 2.55
Weird how write cache is reported as unchecked here.



Everest Ultimate 4.60 Read suite & Random Write & Average Write Access
Looks like 2x32GB OCZ SSD Core raid 0 improved random write averages three fold from single 32GB OCZ SSD Core disk from ~25MB/s single to ~76MB/s in raid 0 and improved average write access times by just over 36%.

Wolf-Tech

I looked at your screenshot's and Noticed why the performance would be so slow. In your booting screenshoot you have the drives configured port 00 01 04 and 05. The intel controler has problems with any drive configuration with split numbers. Put the drives in 01 02 03 04 ports and put your dvd drive in the last port. Then raid 0 them your performance should be 4x the read speed on 1.

I should know since I have run many raid's on the intel ich8 9 and 10 controllers. I did the same thing you did once by mistake with 2 raptor drives and the performance was sick. Then I put them in 00 01 and raided them got my double performance I was looking for.

eva2000

wow thanks for the informative insight !

welcome to i4memory.com Wolf-Tech