2GHZX GST Quick Test Database

[GSKILL TECH]

test under

DFI NF4 Ultra-D Bios 702-1
FX60
430W PSU
S3
Hitachi 60GB SATA2

DRAM FREQUENCY SET: 200 (1:1)
COMMAND PER CLOCK: ENABLE (1T)
CAS: 3
TRCD: 4
TRAS: 8
TRP: 4
TRC: 7
TRFC: 13
TRRD: 2
TER: 2
TWTR: 2
TRWT: 3
TREF: 3684
TWCL: 1
DRAM BANK INTERLEAVE: ENABLE

DQS SKEW CONTROL: AUTO
DQS SKEW VALUE: 0
DRAM DRIVE STRENGTH: 7
DARM DATE DRIVE STRENGTH: 3
MAX ASYNC LATENCY: 8NS
DARM RESPONSE: NORMAL
READ PREAMBLE TIME: 5NS
IDLECYCLE LIMIT: 256
DYNAMIC COUNTER: ENABLE
R/W QUEUE BYPASS:16
BYPASS MAX: 7
32 BYTE GRANULARITY: DISABLE



for 0602 (2006 FEB) 5 kit tested at 2/17/06



5 kits


1 kits: looks like it stable around 286~284









2nd kit with more detail how i got the max overclock
around 274~272
















3rd kit; around 266~264










next time i am going to post only the qucik max clock found

[palehorse]

I dont understand your testing..

for instance, with the 1st kit, how can they be labeled "stable" at 286 if they report errors in memtest? Is there a certain number of errors you find exceptable to still consider them stable?

i guess im just not understanding how you're testing... you appear ot keep the timings static throughout the tests, and you're then just changing the clocks until...what? what tests are you using for stability? memtest? if so, then it looks like each kit is getting errors at the clocks you're describing as "stable"... so i dont get it...

Thanks for your efforts, but can you please explain your process here?

[GSKILL TECH]

I dont understand your testing..

for instance, with the 1st kit, how can they be labeled "stable" at 286 if they report errors in memtest? Is there a certain number of errors you find exceptable to still consider them stable?

i guess im just not understanding how you're testing... you appear ot keep the timings static throughout the tests, and you're then just changing the clocks until...what? what tests are you using for stability? memtest? if so, then it looks like each kit is getting errors at the clocks you're describing as "stable"... so i dont get it...

Thanks for your efforts, but can you please explain your process here?

check my thread 1st, the errors you see was happened in unstable speed

http://www.xtremesystems.org/forums/...ad.php?t=89366

[GSKILL TECH]

let me explain it how i did it.

1st I boot up at 250 with
DRAM FREQUENCY SET: 200 (1:1)
COMMAND PER CLOCK: ENABLE (1T)
CAS: 3
TRCD: 4
TRAS: 8
TRP: 4
TRC: 7
TRFC: 13
TRRD: 2
TER: 2
TWTR: 2
TRWT: 3
TREF: 3684
TWCL: 1
DRAM BANK INTERLEAVE: ENABLE

DQS SKEW CONTROL: AUTO
DQS SKEW VALUE: 0
DRAM DRIVE STRENGTH: 7
DARM DATE DRIVE STRENGTH: 3
MAX ASYNC LATENCY: 8NS
DARM RESPONSE: NORMAL
READ PREAMBLE TIME: 5NS
IDLECYCLE LIMIT: 256
DYNAMIC COUNTER: ENABLE
R/W QUEUE BYPASS:16
BYPASS MAX: 7
32 BYTE GRANULARITY: DISABLE

then, i started memtest two with each one 500mb
i rise mhz and check any error appear,
i rise mhz til 291 and get 1st error pop out, then i lower the mhz til about not more memtest error pop out for like 10 to 15 min. i said this is quick tests but i can very sure to tell you those 3 kits is 99% stable at the speed i stated

[blazin-asian]

how come you have the memtests running 2 x 500? isnt that testing only 1gb? sorry, im new at this.

[GSKILL TECH]

cuz i am runing a quick test
1GB is good enough to see anything wrong with the kit or not

[palehorse]

[Edit: ok, nevermind. I see that if you try to go over 500, it says: "Your version of windows limits the ammount of RAM you can allocate for a single program....etc etc" that explains why you have to run two instances of it.

Then again, only the second instance gives me errors... the first 500 never does.

and this mean that we have to run four instances of the program to test all 2000mb?! weird...

[palehorse]

ok... i managed to use your technique to get 280 stable according to your standards. They even proceeded to pass memtest86 test #5 several runs at bootup.

However, they do not appear to be "game stable." since setting them at that speed cuased a reboot 1 hour into BF2. My temps were not over 50C, so the reboot appears RAM related, and does not occur @ 270. They also failed to pass 32M (two instances) AND they failed SP2004 memory stressing.

So, the bottom line is that I'm not sure of the validity of just using windows Memtest to claim stability at any given speed.

see pic:

[GSKILL TECH]

ok... i managed to use your technique to get 280 stable according to your standards. They even proceeded to pass memtest86 test #5 several runs at bootup.

However, they do not appear to be "game stable." since setting them at that speed cuased a reboot 1 hour into BF2. My temps were not over 50C, so the reboot appears RAM related, and does not occur @ 270. They also failed to pass 32M (two instances) AND they failed SP2004 memory stressing.

So, the bottom line is that I'm not sure of the validity of just using windows Memtest to claim stability at any given speed.

see pic:

i stated on top, that is a quick test.
if you want to make sure that is 24/7 stable, you need to do at least 3 memtest with 512MB each and around 300~400% stable.
that is why i adjust stable speed is about 2 to 3 mhz lower than max

[skidpl8]

So GSTech, if I buy this 2GB kit and I can only get 250 on my ultra-d with these settings, can I RMA them for new sticks?

[GSKILL TECH]

So GSTech, if I buy this 2GB kit and I can only get 250 on my ultra-d with these settings, can I RMA them for new sticks?

if 250 stable we can not do anything about it! sorry