Crucial Ballistix DDR2-1000 (Micron D9GMH) OC'ing Results
 

Crucial Ballistix 2GB (2 x 1GB) 240-Pin DDR2 1000 (PC2 8000)
Model #: BL2KIT12864AA1005
http://www.newegg.com/Product/Produc...82E16820146563
5-5-5-15 2T 2.2v
Micro D9 chips (D9GMH)
Comprehensive D9GMH Chip guide: http://ramlist.ath.cx/ddr2/


I'm not impressed with the results of messing with the timmings. 3 FPS at best? Maybe memory is just not a bottleneck on my system.



FAILURES
(Decided NOT to got above 1.375v)
4-4-4-08 1T at 1.375v/0800Mhz (400Mhz) would not pass SuperPI in XP
5-5-5-15 1T at 1.375v/1000Mhz (500Mhz) would not POST
3-3-3-08 2T at 1.375v/1000Mhz (500Mhz) would not POST

CRUCIAL LOVE
Memory survived 6 unstable Windows XP boots, 10 BIOS Post Failures, and 2.375v for an extended time without a problem. After testing I rechecked 4-4-4-12 (2T) at 1000Mhz with Memtest86+ overnight and it was still good. I'm gonna leave it at these timings but I want to do some more checks with 3DMark06 because it's results for these timings didn't make sense.


RIG
EVGA 680i
E6600 Intel C2D
Memory - Duh
EVGA 8800GTS
PC POWER AND COOLING 750W SILENCER

CPU OVERCLOCK
CPU: E6600
CPU SPEED: 3.375 Ghz
CPU MULTIPLIER: 9
FSB SPEED: 1500 Mhz (~ linked 1:1)
CPU vCore: 1.35v
CPU FSB Voltage: [DEFAULT]
NFORCE SPP: DEFAULT]
COOLING: Zalman 9700 / x4 Case Fans (21db, 27.8 CFM)
CPU TEMP: Using TAT

Desktop Idle 43C
Gaming 51C
Prime95 Average 56C
Prime95 Max 60C

STABILITY VERIFIED BY: Dual-Prime95 12 Hours



SETUP
3DMARK - standard 1280x1020 run with no AA
HL2 LOST COST -1680x1050, x4 FSAA, Grahpics MAX except shadows & water reflection

SETUP SCREW UP
For SuperPI, a 2M digits test was not enough to really show much of anything. Live and learn.

RESULTS

HL2 Lost Cost Stress Test
(0800Mhz) 5-5-5-15 (2T) 2.200v 166.37 FPS

(1000Mhz) 5-5-5-15 (2T) 2.200v 168.12 FPS

(0800Mhz) 4-4-4-12 (2T) 2.200v 169.03 FPS

(0800Mhz) 3-3-3-08 (2T) 2.325v 168.54 FPS

(1000Mhz) 4-4-4-12 (2T) 2.200v 169.66 FPS

SiSoft Sandra memory Sub Systems

(0800Mhz) 5-5-5-15 (2T) 2.200v 6562 MB/s Bandwith Int Buff'd
(0800Mhz) 5-5-5-15 (2T) 2.200v 6524 MB/s Bandwith Float Buff

(1000Mhz) 5-5-5-15 (2T) 2.200v 6945 MB/s Bandwith Int Buff'd
(1000Mhz) 5-5-5-15 (2T) 2.200v 6928 MB/s Bandwith Float Buff

(0800Mhz) 4-4-4-12 (2T) 2.200v 7049 MB/s Bandwith Int Buff'd
(0800Mhz) 4-4-4-12 (2T) 2.200v 7041 MB/s Bandwith Float Buff

(0800Mhz) 3-3-3-08 (2T) 2.325v 7304 MB/s Bandwith Int Buff'd
(0800Mhz) 3-3-3-08 (2T) 2.325v 7365 MB/sBandwith Float Buff

(1000Mhz) 4-4-4-12 (2T) 2.200v 7712 MB/s Bandwith Int Buff'd
(1000Mhz) 4-4-4-12 (2T) 2.200v 7643 MB/sBandwith Float Buff

SuperPI 2M
(0800Mhz) 5-5-5-15 (2T) 2.200v 40s

(1000Mhz) 5-5-5-15 (2T) 2.200v 38s

(0800Mhz) 4-4-4-12 (2T) 2.200v 40s

(1000Mhz) 4-4-4-12 (2T) 2.200v 38s

(0800Mhz) 3-3-3-08 (2T) 2.325v 38s

3DMark06
(0800Mhz) 5-5-5-15 (2T) 2.200v 8942

(0800Mhz) 4-4-4-12 (2T) 2.200v 8999

(1000Mhz) 5-5-5-15 (2T) 2.200v 9007

(1000Mhz) 4-4-4-12 (2T) 2.200v 8967 (very odd .. 2nd run confirms. 8966)

(0800Mhz) 3-3-3-08 (2T) 2.325v 9014

Re: Crucial Ballistix DDR2-1000 (Micron D9GMH) OC'ing Results
 

Yeah the core 2 like the Athlon is not at all bandwith sensitive unlike the P4, due to the clock cycles of 11 Vs the Prescotts 38 in this case due to them being slower they try to cache most of it as such is very bandwith heavy