XMS 6400C3s: DDR2-1150 3-3-3-1!

[nn_step]

pity about the board.. but that is some nice ram you have there...

[boostedevo]

Would you care to elaborate, please? What strikes me is that DDR2-1100+ 8-3-3-3.0 with 6400C3 is seen only on BadAxe board. I also have a set of COrsair 6400C3 here and testing it on AM2 and I'm really "shocked" to see such clocks, from you and Tyrou. Even given the fact that platform is different... VictorWang also has 1100+ 8-3-3-3.0 with his 2GB TeamGroup and it's on BadAxe as well... Just curious, what are those certain circumstances/BIOS settings? Thanks!

BTW, sorry about your misfortune...

Sure, 1333FSB strap + 4:5 mem ratio is the only place those speeds with the tight timings will happen on this board. I don't know what it's doing there to let it happen, but the mem OCs are insane as we are all seeing

The board *has* to be doing something we don't know about, but I have no clue what it is. Maybe altering operations/timings in the mem controller or something. Whatever it is, it sure does make for some very impressive mem OCs...too bad the board doesn't like extra Vddr and tops out at 2.2V stock.

@nn_step: thanks

[jmax_oc]

What perform theses sticks on a P5WD2-E for exemple ?
Are those impressive results due to BadAxe, Conroe or themselves ?

[boostedevo]

I have a Premium (955), not a -E (975). I've seen other people's C3s in the -Es do 1000 3-3-3-1 @ 2.4V though (Stilt). It's the BadAxe board, not Conroe. There is "something" about that strap/divider on the BadAxe and that's all there is to it.

[p360stick]

nice

[Fixxxer]

There's no doubt that the new D9 chips are good, but casually all the crazy screenshots/validations that we can see are taken with a Bad Axe.

Why not testing the same memories under other i975x boards? Then we could see what do these D9 chips really do.

[MonkSP]

I saw this mems for sell at newegg. too bad they are costing 450+. Out of my budget for memory

[mupets_revenge]

I was reading somewhere that when on the 1333 strap the bad axe increases latency in the chipset which allows for the better ram clocks. I know that sounds vague but it would make sense if the read/write to chipset is slower then however low you set the timings that ram wouldnt have to run at them cos its being limited by the chipset latency.

Well thats how i saw it.... am i anywhere near right?

[T07N]

What's the price tag on these? Where?

[thunderstruck!]

I was reading somewhere that when on the 1333 strap the bad axe increases latency in the chipset which allows for the better ram clocks. I know that sounds vague but it would make sense if the read/write to chipset is slower then however low you set the timings that ram wouldnt have to run at them cos its being limited by the chipset latency.

Well thats how i saw it.... am i anywhere near right?

Actually, the RAM is running 1:1, not 4:5.

[party animal]

What's the price tag on these? Where?

450+ for 2GB @ the egg. Waaaaay out of my price range, but very nice sticks.

[kcnyc]

Yep, DDR2 voltage above 2.2v will eventually cause failure of the VDDR upper regulating MOSFET due to overcurrent. The higher above 2.2v the faster the failure. Solution is a power MOSFET rated for a higher current...

-FCG

Do you know where to get one? We can change the power MOSFET right?

[_damien_]

Really great sticks! :thumbsup:

Nah, the memory controllers for both AM2 and Intels lack CAS 2, there's no way

The AM2 tech doc actually says both yes and no...

BIOS should not assume that a device can operate with either slower or faster CAS latency than those specified by the SPD. The DRAM controller is designed to support CAS latencies 2, 3, 4, 5 and 6.

However, CAS 2 is not listed in the supported timings table

I'm leaning towards the AM2-controller not supporting CAS 2, but even if it does, the 2-setting is unlikely to work because the memory chips too must support CAS 2. Unlike most parameters, CAS isn't a mere "time-to-complete" operation - each setting must be built into the chips in the form of pipeline stages.

[DeToNaToR.cl]

I saw this mems for sell at newegg. too bad they are costing 450+. Out of my budget for memory

This ones got D9XXX
http://www.newegg.com/Product/Produc...82E16820146106

[ScottFern]

Will these 6400C3s boot up at 1.9v? The MSI board I have coming defaults at that vdimm setting.

[dinos22]

this seems way too good to be true

your 32M score is pretty slow as well..........

[guess2098]

humm
1333 + CL3 in bad axe thread still miss leading many members.....

[jmax_oc]

What are the chips under these head spreader ?
D9GMH ? D9GKX ?

[Revv23]

There's no doubt that the new D9 chips are good, but casually all the crazy screenshots/validations that we can see are taken with a Bad Axe.

Why not testing the same memories under other i975x boards? Then we could see what do these D9 chips really do.

None of the other boards can do it.

I havent really believed any badaxe memory record for a while.

No offense to booste... But several users have pointed out that cpu-z doesnt always read mem speed correctly on badaxe (whenever you hit those magic bios settings, like oyu say.)

I dont really know enough to make judgement though, so i reserve my thoughts until someone more experienced comes around.