M2N32-SLI DLX + OCZ 2GB PC7200 XTC Plat. = DDR1200 goodness

[3oh6]

OCZ2N9002GK - 2X1GB DDRII PC7200 4-4-3-15 2.1v (D9GKX)

More eye candy available upon request.


I finally have my first kit of DDRII to play with on the M2N32-SLI so bear with me as I drag my feet through this. This kit is a review sample I received from OCZ but in the past all other samples from OCZ I have received for review perform on par with end-user results.

the setup:
Asus M2N32-SLI bios 0501
AMD64 3200+ 0618XPDW 'naked' @ 300*10 w/1.64v
3/8" Asetek WaterChill
OCZ DDRII 2X1GB 7200 Plat. XTC @ 2.4v
HIS X800GTO
SIlverStone Z56ST
Raptor 36GB
Win XP Pro SP2



---DDR1200 5-5-4-5 VDIMM 2.4V - CPU Validation---


---DDR1200 5-5-4-5 VDIMM 2.4V - 32M / Prime Stable---


---DDR1200 5-5-4-5 VDIMM 2.4V - Memtest Stable---


---DDR1200 5-5-4-5 VDIMM 2.4V - 3D Stable---


---DDR1200 5-5-4-5 VDIMM 2.4V - Sandra Buffered---


---DDR1200 5-5-4-5 VDIMM 2.4V - Sandra Un-Buffered---


As you can see this setup is rock solid stable and the memory is doing just fine. My only concern thus far is the Sandra Bandwidth numbers and for some reason I was thinking they would be higher at these settings. Any comments on that?

---DDR1100 4-4-3-5 VDIMM 2.5V - 32M / Prime Stable---


---DDR1100 4-4-3-5 VDIMM 2.5V - Memtest Stable---


---DDR1100 4-4-3-5 VDIMM 2.5V - 3D Stable---


---DDR1100 4-4-3-5 VDIMM 2.5V - Sandra Buffered---


---DDR1100 4-4-3-5 VDIMM 2.5V - Sandra Un-Buffered---


At 4-4-3-5 I couldn't get 280*10 to run 32M so I will have to settle at DDR1100. Of course the limitation in voltage is the motherboard as I haven't done a volt mod just yet. I still might not keep this M2N32 so I want to keep it original. Playing withing timings or secondaries couldn't get 32M stable above 275*10...close but not quite.

---DDR850 3-3-3-5 VDIMM 2.5V - 32M / Prime Stable---


Some adjusting of secondaries was required to stabalize the memory at 3-3-3-5 DDR850 but I finally pulled it off sans windows memtest so it looks like 3-3-3-5 at these frequencies are for benching only. I would have liked to get that magic DDR900 at 3-3-3-5 just because but that just won't happen @ 2.5v. This will pretty much wrap up my work with these fine modules unless there are any requests for something specific, just ask.

[RyderOCZ]

[bachus_anonym]

How high have you managed to push it at 8-4-4-4.0 ? Did you clear DDR-1100 for 32M or SP2004?
Thanks for sharing your results, man

EDIT: NVM, just noticed you will update this weekend

My only concern thus far is the Sandra Bandwidth numbers and for some reason I was thinking they would be higher at these settings. Any comments on that?

Those number look pretty normal to me - Buffered @ DDR-1180 8-4-4-4.0. If you had dualcore your Buffered and Unbuffered would be much higher

[3oh6]

DDR1100 @ 4-4-3-5-11 is certainly 32M stable and a couple hours of SP04 without breaking 2.5v, not sure about Win Memtest but I will certainly find out soon enough.

something kind of major just happened and something tells me my free time just expanded by a good 40-50 hours a week *nudge *nudge

**edit**
Okay, thats what I figured but wasn't sure that the dualies had such an effect on bandwidth numbers. I guess I will get more first hand experience with that when they come down to a reasonable price that a now un-employeed bum can scrape together
Thax for the heads up and I fired you a PM too BA.

[hipogtr]

Great results. Keep om tweaking for some 4-4-4 action.

[DaWaN]

What's under the spreaders ? D9GMH ?

[3oh6]

What's under the spreaders ? D9GMH ?

aha...there is something i was forgetting. not D9GMH, but D9GKX. first post will be updated, sorry guys.

and 4-4-3-5 is coming up soon enough, just priming for an hour or so then some win memtest and bandwidth numbers. 1100 seems to be the tops at C4 w/2.5v.

[3oh6]

first post updated with 4-4-3-5 timings. just waiting on memtest to reach a good 500% before throwing that SS up there. next up will be 3-3-3-x so expect results around this time tomorrow.

on a side note, this memory went from $319 @ Newegg last week to $499 so it went from "amazing deal" to "tough to swallow" in a hurry. i assume an error was made because that is quite the jump.

[SMa]

It seems like all D9G is good.
D9GMH, D9GKX, ... (are there more?)

[b0bd0le]

how much will this ram cost?

[3oh6]

how much will this ram cost?

that is tough to say, Newegg had it for $319 last week and this week it is $499USD. i have also found it on a couple of Canadian sites for around $600CDN which is in-line with the Newegg price but I also just found it at ZZF for $320USD (http://www.zipzoomfly.com/jsp/Produc...ductCode=85247).

who knows if it will jump in price there as well or if the Egg and Canadian stores are off their rocker. oh, and first post has been updated with a pinch of 3-3-3-5 @ DDR850 action.

[thunderstruck!]

Very nice review.

[Vito]

I am at 540 atm, stock voltage, stock timings, I also have the M2N32, I am a total noob at DDR2, 1. Why is 1T such a struggle with DDR2? 2. what is the stock voltage, I tried to set it to 2.1v and it didn't post, too low?

[3oh6]

thanx guys.

stock voltage for the kit is rated at 2.1v, mine definitely run stock speed/timings at stock volts. i always test that right out of the package, no problems there.

as for why 1T is so hard, i imagine it has to do with the IMC and just the stress to run high frequencies at 1T.

[|SiLA|]

nice nice..going to replace with this ram the crap kit i bought yesterday pc6400 44415 xtc platinum

going to test it later on intel system

[|SiLA|]

not happy yer of the results i'm having..maybe because i can only give 2,3

any chances to try them with 2,3 by you?

[3oh6]

what type of results are you looking for at 2.3v? i think i might have some screenshots of 2.3v runs as that is the 'technical' warranty limit for voltage, but i definately have results in my books and can test further if need be. what timings and/or frequency are you trying to achieve? also a list of your hardware...if there is an X2 in there then that may be the reason for differing results.

[chema]

can yo do a 3dmark06? what do you obtain in superpi 1mb??

[|SiLA|]

what type of results are you looking for at 2.3v? i think i might have some screenshots of 2.3v runs as that is the 'technical' warranty limit for voltage, but i definately have results in my books and can test further if need be. what timings and/or frequency are you trying to achieve? also a list of your hardware...if there is an X2 in there then that may be the reason for differing results.

4444 and 4434

since i can't even load windows at 1000 4444 i can't load windows but with 4445 but only few seconds of prime

TIA

edit

let's see what timings u can have with 2,3v @1000

edit 2

P5WD2 Premium
PD 930 @4,6(can do 4,8)
these ocz

tagan 580
x1800xt 512

i've checked the chip and it's D9

[3oh6]

can yo do a 3dmark06? what do you obtain in superpi 1mb??

---3GHz (single core) DDR1000 4-4-3-5 VDIMM 2.2V 3DMark 06 ---


BFG 7900GTXs SLI just above stock on air, turned out they are both bunk so there will be no fun with them for a bit. not sure how much memory plays a role in 3DMark 06 but this is pretty much the only work i have with the PC7200s and 3DMark 06. as for 1M super pi...

I did a direct comparison at 319*9 2.87GHz and got these results just cycling through the top 3 dividers and the best timings with VDIMM 2.3v:

CPU 319*9 = 2871MHz // HT 319*4 = 1276MHz

Speed        DDR820     DDR956      DDR1148
Timings      3-3-3-8    4-4-3-5     5-5-4-8 2T
SuperPi 1M   29.972     29.494      28.875

These are averaged results from multiple runs. The only other 1M runs I have done were geared at the best 1M time at 3GHz or below. I cracked 26 seconds on 939 with TCCD but the best I have managed with the PC7200 on AM2 was 27.5 seconds DDR1200 5-5-4-5. I don't have a screen shot because it won't be worth it until I crack 26 seconds

Let me know if there is anything else and I will get around to posting up here.

[|SiLA|]

how come 1 stick only?Hehe

[3oh6]

let's see what timings u can have with 2,3v @1000

P5WD2 Premium
PD 930 @4,6(can do 4,8)
these ocz

tagan 580
x1800xt 512

i've checked the chip and it's D9

Ahhh, well I don't think my results are even relevant seeing as you are running Intel. Either way here is a screen shot I have of 2.3v at the highest I could get 4-4-3-5 with tight secondaries...happens to be just over DDR1000

---2.5GHz (single core) DDR1018 4-4-3-5 VDIMM 2.3V Stable---

how come 1 stick only?Hehe

Haha, woops. I forgot that run was with a single stick. I was short on DDR2 for a few days and needed two AM2 systems running...cheap and easy was to split it up.

[|SiLA|]

tnx a lot

imma try these rams on AM2 system saturday..same mobo b ut ultra not sli..and 3500+ cpu

[chema]

thanks 3oh6 for the 3Dmark and superpi.
great results!

[ozzimark]

It seems like all D9G is good.
D9GMH, D9GKX, ... (are there more?)

same chips, different rated speed. D9GMH = B6-3 (333mhz @ cas5), D9GKX = B6-25E (400mhz @ cas5)