4x512MB on 875P chipset

[Azsen]

Hi, sorry if this has already been asked before, I can't find the information with search.

I've already got 2x512MB Corsair XL 2-2-2-5 in dual channel running. I bought some Corsair PC4400C25 2x512MB dual channel kit which runs at 2-2-2-5 at DDR400 I think and 2.5-4-4-8 @ DDR550.

I was originally going to just sell the XL and use the new stuff for heavy overclocking but I want to see what BF2 performance is like with 2GB. If it makes a significant difference I may keep it.

Anyway I was wondering what limitations there are when running 4x512MB in a 875P chipset? Does PAT get disabled? Do I have to run the sticks at reduced speed like DDR333?

Thanks

 

[Arcanedeath]

It depends on your motherboard, but in general it will work just fine at stock speeds even w/ double sided dimms in all 4 slots. it doesn't overclock as well w/ all the slots full tho.

 

[TStep]

Holy Grail 1
Holy Grail 2

Good reads there and gives you an idea on limitations. Since your a mixed set, its hard to predict compatibility, but since you have them already, give it a run.

 

[Azsen]

Thanks, I've read through those and this is interesting:

The fastest configuration should be 4 Double-Bank DIMMs on the Intel 875 chipset motherboard. Our first efforts to test this configuration, with 4 x 256 DS OCZ 3700 GOLD modules did not show 4 DS DIMMs any faster than 2 DS DIMMs on the ASUS P4C800-E. This was a real puzzle considering that other predicted configurations were easily verified with Quake3, Super PI and other benchmarks. When we looked deeper, the problem became clear. In every configuration that we had tested with the ASUS P4C800-E, the Performance Mode showed ?enabled? but with 4 DS DIMMs at any FSB speed faster than 800, CPU-Z 1.18C shows Performance Mode ?disabled?. Our trial was failing here because we were comparing 2 DS DIMMs with PAT on to 4 DS DIMMs with PAT off. We have asked ASUS for the reason behind why PAT appears to be disabled at 4 DS DIMMs above 800FSB.

Using 4 sticks seems to have better performance. Very unexpected. But they have to be identical...

I'm confused about what they define as "identical" dual sided sticks. I think all my sticks are TCCD based, but would running them at the same timings mean they are identical? Or do they need to be exactly the same right down to the revision/batch number to get the best performance for 4 sticks? If they are not the same do I lose 23% performance? :Q

I have the LanParty 875B Rev B. So the real test will be if PAT stays on when using 4 sticks on it while overclocking?

 

[TStep]

In the past I had 2x256 OCZ PC3500 EL and 2x256 Mushkin Level I running in an IC7 and experienced no problems at all. Not your same situation at all, but i would not be afraid to try it. I was still running with the fsb above 280 in that configuration as well with 2-3-3-6 timings, 5:4 mem ratio IIRC.

 

[Azsen]

Whoops, posted my last reply before I had finished typing it.

 

[TStep]

Are you familiar with the programs CPU-Z and memtest86+? I would test the following:

-Setup everything at stock speeds
-Use 2x512 XMS, run a couple of memory benchmarks (snadra, everest, etc), SuperPi, and a couple of game benchmarks.
-Do the same only changing to the Pc4400
-Do the same using 4x512

Compare the benchmarks and see if there is any performance lost or gained. Check mem stability with memtest86+. Use CPUz to check for PAT enabled. Overclock from there.

 

[imported_Kensai]

Shouldn't be any problems.

 

[Azsen]

Originally posted by: TStep
Are you familiar with the programs CPU-Z and memtest86+? I would test the following:

-Setup everything at stock speeds
-Use 2x512 XMS, run a couple of memory benchmarks (snadra, everest, etc), SuperPi, and a couple of game benchmarks.
-Do the same only changing to the Pc4400
-Do the same using 4x512

Compare the benchmarks and see if there is any performance lost or gained. Check mem stability with memtest86+. Use CPUz to check for PAT enabled. Overclock from there.

Yep, I can do that no problems. Used them all before. That would be the way to go about testing it.

It hasn't arrived yet, I expect it this week with some luck. Ordered it from ZipZoomFly and got a mate to send it to me. To buy that same stuff new in NZ it's twice the price no joke.

The current Corsair wouldn't go above 248Mhz, so it may be possible for me to get ~260Mhz 1:1 FSB stable with 2x512 of the new Corsair. So we'll see how 1GB, 260FSB compares to what speeds I can get with 2GB.

 

[TStep]

Recently I had some Patriot 2-2-2-5 2x512 hooked up with an IC7-Max3. I've read TCCD doesn't play well with 875, but I had some excellent results. IIRC 260ish at 2.5-3-3-7 (cpu being the limit). I think you'll be pleased with your results.

 

[SVT Cobra]

Originally posted by: Kensai
Shouldn't be any problems.

 

[Azsen]

Thought I'd give you guys an update. Got the RAM today and here are the results. You may need to scroll sideways to get them all.

With the Sandra tests I've copied how Anandtech did the unbuffered tests here. With the new RAM in I had to drop to 2-3-3-6 because the new RAM won't do 2-2-2-5 @ DDR400 according to this article. It will however do DDR500 at 2-3-3-5.

Anyway as you can see from the Sandra unbuffered tests there is better performance with the 4 dimms in at 2-3-3-6 which even beats the 2x512 in 2-2-2-6 configuration. :Q

Everest Memory Read & Latency tests show lower scores, but the write test an increase?

SuperPI 2M slightly better with 4x512 compared to the 2x512 of the new PC 4400 C25. PC Mark slightly higher memory tests for 4x512 as well.

Overclocking wise, PAT stays on over 230 FSB. Though the machine reboots at that speed. At the moment it is 47mins Prime95 stable at 220 FSB, 2.5-3-3-7, 2.8v. This is still a far cry from my best overclocks with the XL 2x512 though. The XL is dead stable to 240 FSB, 2.5-3-3-7, 2.8v.

Seems like a big drop in overclock? In the article Wesley was saying only 6-7 FSB loss from 4 DIMMs?