Micron D9 GMH\GKX\JKH etc club

[Dee]

650? Very nice!

[CryptiK]

Well, 649.5 & 649.6 to be accurate, but still a lot better than 605 which was my previous max on this board with the same ram kit. Thanks mate.

[Shiranui Gen-An]

I just figured out how to clock ram high on my M2F. The setting that made no difference at all on my P5Q-Deluxe, made all the difference on my M2F.

What setting was that?

[Dee]

I was also about to ask.

[CryptiK]

You're both going to laugh. I sure did. The setting is "Memory OC Charger" and it's description is "Enabled: Increase ram overclockability". It didn't make any difference at all on my P5Q-Deluxe, so I just set it to disabled on my M2F and completely forgot about it.

I also set the strap manually to 266 and PL 10.

[Ket]

@ Dee, either a 7200 or 8400, maybe 8500 depending on price diff.

Cryptik, nice results, but what was that one setting?

ED - dho.. thats always happens I ask something and the reply comes in as im posting lol

[Dee]

This RAM rules. DDR1200 @ 1.9v! Runs PI just fine, too. Booted into Windows at DDR1302 using 2.00v but wasn't stable for long!

[zsamz_]

This RAM rules. DDR1200 @ 1.9v! Runs PI just fine, too. Booted into Windows at DDR1302 using 2.00v but wasn't stable for long!

chek your real volts
i need a minimum of 1.98v real for 1200mhz

[CryptiK]

I reduced the tRD (PL) and got some better bandwidth & lower latency:

Everest @ 651 MHz, tRD 6 @ 2.2v real


SPI1M @ 651 MHz, tRD 6 @ 2.2v real

[Dee]

chek your real volts
i need a minimum of 1.98v real for 1200mhz

As stated earlier, I cannot check "real" volts.

And what if my board is undervolting the RAM? Without checking with a multi-meter no one knows. Although I doubt it's THAT far off what's stated in BIOS anyway. Even if it's 1.96v, that's pretty good for DDR1200 speeds.

[zsamz_]

As stated earlier, I cannot check "real" volts.

And what if my board is undervolting the RAM? Without checking with a multi-meter no one knows. Although I doubt it's THAT far off what's stated in BIOS anyway. Even if it's 1.96v, that's pretty good for DDR1200 speeds.

you got a nice set
one of the rare d9gkx enjoy

[Dee]

Yes, finally! I always have bad luck with hardware, so maybe the memory god felt it was time for something nice to happen to me!

I love CellShock. This brand has quickly become my favourite, and man I was a BIG Crucial fan for the last 4 years. I gotta hand it to MSC. They put out some amazing products. Actually, I owe you a "thank you", zsams, you were the one who suggested I take a look at CellShock. Cheers!

EDIT: Just a thought -- what's the max recommended voltage without active cooling? I know opinions on this vary somewhat but I'm interested in other D9 users thoughts.

[zsamz_]

Yes, finally! I always have bad luck with hardware, so maybe the memory god felt it was time for something nice to happen to me!

I love CellShock. This brand has quickly become my favourite, and man I was a BIG Crucial fan for the last 4 years. I gotta hand it to MSC. They put out some amazing products. Actually, I owe you a "thank you", zsams, you were the one who suggested I take a look at CellShock. Cheers!

EDIT: Just a thought -- what's the max recommended voltage without active cooling? I know opinions on this vary somewhat but I'm interested in other D9 users thoughts.

lowest volts for max bandwidth + latency
i personally get best results @ 1200+ 2.0-2.05v cas 5
cas 4 1000+ i need 2.2v+
so i chose 1200+
no matter what voltage or speed i cool mine with 2 90mm fans

[WaterFlex]

That`s mine lovely Shocks:

[CryptiK]

Nice work there mate

[Dee]

Nice!

[Ket]

Done some more tweaking, can't say this is bad for a poxy 3.3GHz

[WaterFlex]

CryptiK, Dee
Thank you!

[CryptiK]

Nice Ket, love the 108% FSB overclock

Can you repeat that test at exactly the same settings except for setting 5-5-5 for the first 3 primary timings? I have found through my testing that CL 4 has very little gain in bandwidth & causes a small reduction in latency, much less than it used to with other platforms (P35, P965). I haven't seen what a P5Q-Pro does. Seems to be a strange limitation/behaviour of the P45.

@ Dee - nice work from you above too, sub 2.0v real 600 MHz is very nice indeed

[Dee]

Thanks C.

Wow Ket! 108% FSB overclock!? That's insane! Nice work brother.

[Ket]

FSB has never been a problem for my E4400 I hate how my CPU gets insanely unstable past the settings I'm using atm tho It has to be my CPU, three RAM kits cant all crap out at exactly the same area. I'll test 5-5-5-12 settings in a bit.

ED - heres the 5-5-5 results, practically no difference at all.

[CryptiK]

There is no way your kits wont do 1200+, Dee's kit does 1200 sub 2.0v, mine do 1300 on 2.2v, it probably has to be your board.

[Ket]

Nah can't be my board either, 1066FSB CPUs would be screwed otherwise. I don't care how crappy a manufacturers quality control is theres no way (or very very unlikely anyway) a board would make it out the door that cant even handle a stock 1066FSB CPU running 1:1 with some PC8500 memory

[CryptiK]

Nah can't be my board either, 1066FSB CPUs would be screwed otherwise. I don't care how crappy a manufacturers quality control is theres no way (or very very unlikely anyway) a board would make it out the door that cant even handle a stock 1066FSB CPU running 1:1 with some PC8500 memory

Yeah I suppose so, hope your new cpu allows you to have some ram oriented fun. Thanks for doing the CL 4 vs CL 5 test, as I suspected, there is very little difference. It must be a P45 quirk as I have not seen any of them do much better on CL 4 vs CL 5 which is a little irritating.

Just did this with my TG GMH @ 2.24v This board isn't too bad now I've figured it out.

[Ket]

Not bad I just done a little more investigating and my CPU is deffinately the issue. Whacked everything back to stock and just set FSB to 320 to give a mem frequency of 1280MHz.. or was it 1320? well something like that , went right through POST then crapped out. With any other OC settings the system won't even POST with 1200MHz mem frequency nevermind 1300+. Bad, bad E4400