>>> Corsair XMS 7 stick shootout! <<<

[THUGSROOK]

i was curious

i might try a 3200c2 but chances are ill get pretty much the same results.

 

[SupermanCK]

hmmm...you never know until you try the pc3200c2....

 

[Mikki]

excellent testing Thugs, I wish I was on the same schedule as you guys, I miss all the fun....

 

[SupermanCK]

don't feel bad mikki....this is the prime time for us....he ehee

 

[Mikki]

I'm very curious about the cas2 stuff, my opinion is that is wouldn't be much different....

 

[THUGSROOK]

Corsair XMS3200c2 512mb is on its way....

can it beat my current XMS3000 scores?

(i know SCK has paypal funds ready)

 

[Mikki]

ask and ye shall receive.....

 

[LastRide]

As always,some nice testing.It is a real tough choice on the ram.The 132FSB max with 2 sticks kind of blew me away on the P4B533.I know 1 stick is better but I was a bit surprised how 2 sticks crippled the board at 3:4.

 

[THUGSROOK]

 

[Mikki]

guess there's a method to the madness...hehe

 

[THUGSROOK]

ive been a P4 DDR "Fanboy" for the last 11 months and i have to admit....

DDR does not offer enough bandwidth for a P4 system ~ including mine!

(you can see where this is going right?)

 

[RalfHutter]

Originally posted by: THUGSROOK
well we finally found a few reasons why Intel disabled 3:4 on these boards....
cant run 2 sticks above 133fsb 3:4 and cant run beyond 165fsb 3:4 stable

I've posted about my experiences with my P4B533E and 2 x 512MB of XMS3200 at Asusboards Forums but I'll give a little synopsis because it fits the direction this thread is taking.

My old rig was a 1.8A, P4B266 and 2 x 512MB of XMS2700. I was able to run this at 128FSB (2.3GHZ, DDR340, CAS2-3-3-6) 100% stable. Was tested extensively with Prime95 and Memtest86.

I bought a new 2.26B, a P4B533E ans 2 x 512MB of XMS3200C2 figuring I could do the 3:4 trick and run at 150FSB (400DDR @ CL2) with no problem. Unfortunately things didn't work out quite that well.

After extensive testing (long sessions of Memtest86) I've maxed out at 145FSB. This gives me 2.46GHZ, DDR387. I'm running the RAM at 2.5-3-3-7, Vdimm is 2.7V. This has tested 100% stable (zero errors) after 17 hours of Memtest86 and 18 hours of Prime95.

If I run as 2-3-3-6 I can only get 138FSB before I start to get Memtest errors. 2.5-3-3-6 gives me 140FSB. 2.5-3-3-7 with 2.6V would only give me 141FSB, the extra .1V gave me 145FSB but I don't want to run the Vdimm any higher right now.

I ran Sandra and 3d Mark and found the sweet spot to be the 387MHZ setting. That's where I'll stay for now but I'm sure I'll shoot for better later on. I spent all last weekend running Memtest86 and am tired of it. I want to spend a little time in Windows for a change.


BTW - See my Rig for more system details.

 

[RalfHutter]

Originally posted by: THUGSROOK
.....DDR does not offer enough bandwidth for a P4 system ~ including mine!

(you can see where this is going right?)

I was gonna go the Rambus Way but after extensive studying decided to stick with DDR for now.

1)Try finding a good supply for quality 1066 RDRAM nowadays.

2)Try finding a stable board to run it on.There's LOTS of horror stories out there right now.

3)Seems like no more Intel development of the I850 chipset.

So I went the DDR400 route but it didn't quite pan out. I'm getting 2980+ in the Sandra Memory benchmarks so I'm *close* to 3200 at least.

 

[THUGSROOK]

just one little "tick" of a cas setting gets me 100 points on 3dmark.
3 ticks gets me 300 points and gets me a 3350 sandra score. (164fsb:437ddr)
obviously this system is screaming for more bandwidth and unless i can run 165fsb:440ddr cas222, there is no point in upgrading this system for a faster cpu.

the MHZ race is pointless w/o memory bandwidth :Q

 

[oldfart]

Bring on the dual channel DDR. Now that will be sweet!

 

[THUGSROOK]

DC DDR does look sweet but ~ we have no idea how it will work for overclockers.
memory ratios? overclockability? stability? all unknown.
im sure they are gonna spoon feed it to us very slowly too

<EDIT> we now know DC DDR can run upto about 188fsb 376 DCDDR

 

[oldfart]

memory ratios? overclockability? stability? all unknown.

Yeah, the 845G is a good example. Got a new mem ratio, but overclockability and stability are not as good as 845D/E.

 

[RaynorWolfcastle]

I got a 512 MB stick of Corsair PC3200C2 coming as well (Platinum heatspreader)... I've also got a new Athlon XP 1600+ on it's way. If I'm lucky, I can get an AGOIA core and perhaps hit 200 MHz FSB with a 8.5-9x multiplier. I would be extremely pleased with that. Hopefully my Epox 8K3A and peripherals can take that kind of high FSB without choking

-Ice

 

[THUGSROOK]

ive been bustin my hump tryin to get 166fsb:443ddr to pass on the xms3000.....

 

[THUGSROOK]

ttt for the blind ppl

 

[RaynorWolfcastle]

this thread is making me anxious to receive my RAM and CPU I'll post AthlonXP results similar to your P4 results when I get everything since I should have a pretty good OC'ing rig by the middle of this week

It will be interesting to see how an OC'd Athlon XP 1600+ (hopefully AGOIA) + Epox 8K3A + DDR400+ RAM compares to P4 rigs
Add in Canadian cold weather and I should be set with a nice OC

 

[Jarhead]

I got a 512 MB stick of Corsair PC3200C2 also, I got mine from newegg, as there are some really bad ratings on www.resellerratings.com for some of the other vendors on pricewatch that were offering it.

I was able to get another 10 Mhz out of my FSB over my premium crucial memory, now I am limited by my HD.

 

[dbwillis]

Thanks Thugs for all the testing....

 

[jmeyer2718]

thugs, i read on the corsair forums that the only difference between the black and platinum heat spreaders was the color, and that the actual spreaders were the exact same thing. this was coming from a corsair employee/tech guy. am i mistaken?

 

[THUGSROOK]

seems to be more of a difference then just color.
the black cant do cas222 at all....
the platinum cant do cas232 437ddr....
etc.

theres a big difference in these sticks.