• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

DDR2 800 vs DDR2 1200

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

juggernaught

Registered
Joined
Oct 5, 2007
I'm in the design stage of a new build and I'm looking at the EVGA 780i for a motherboard. This board is listed as a DDR 1200 board, but it seems like a lot of people are using DDR 800 on it. There doesn't seem to be muct DDR 1200 out there so my question is am I better off overclocking some DDR 800 Ballistix... or maybe 1066?
 
DDR2-1200? That's a 600 FSB - I doubt that any motherboard can really run that fast in real world usage. For the most part DDR2-800 is fine as it runs on a 400 FSB, which covers most any overclock (except the E8400). If you want to be safe you could get some DDR2-1000 (or 1066), but unless you plan on overclocking above 400 FSB save your money and get DDR2-800.
 
Yah I was planning on going with the E8400... think I should go with 1000 then?

Get the RAM (and mobo/proc) I got... I love my system (check sig)!

You can get the RAM for cheap on egg, has gotten a ton of good reviews (commercially and from people here), and is DDR2-1000
 
DDR2-1200? That's a 600 FSB - I doubt that any motherboard can really run that fast in real world usage
If you were to use a 1:1 ram divider, then you are correct, 600fsb. Otherwise you can use a higher divider and take advantage of that and actually run it at 1200.
 
But isn't it true that there is no real performance gain from running a pumped up FSB:RAM ratio? Wouldn't it just be better to run 1:1 with tighter timings than just throw wasted clocks at the FSB?
 
But isn't it true that there is no real performance gain from running a pumped up FSB:RAM ratio? Wouldn't it just be better to run 1:1 with tighter timings than just throw wasted clocks at the FSB?

yep, but if you have a Wolfdale, you can run these 1:1 with the FSB @ 500MHz and the multi at ether 8.5 or 8x.
 
Yeah, that I know. I figured I'm probably going to be upgrading my CPU to either a Wolfdale or Nehalem later on and wanted to be ready for the lower multiplier. That's why I bought these 1000mhz instead of 800.
 
Yeah, that I know. I figured I'm probably going to be upgrading my CPU to either a Wolfdale or Nehalem later on and wanted to be ready for the lower multiplier. That's why I bought these 1000mhz instead of 800.
I believe Nehalem is DDR3 only. :)
 
Oh well. I had planned on building a whole new rig for Nehalem anyway, just figured it was $100 less that I would have to spend. Dunno what I'll do with this computer, electricity is waaaaaaay too expensive to fold or anything.
 
the Pre-IDF thread i started has, info before the IDF conference. in which they cover neha and it is DDR3 only.

cas 4 ddr2-800 will have the same latency as ddr2-1000 cas 5. your better off just sticking with 1:1 ratio and hitting 400fsb. better yet just run unlinked mode on the nv board and do ddr2-800 cas 4.
 
Thanks for the replies.

I see most of you have p35 boards, I'm actually looking at an SLI board though... EVGA 780i...

I hear the Nvidia chipsets have trouble with high FSB, am I getting into trouble at 500 FSB?

Also is the multiplier on the E8400 locked? I'm assuming at the moment I can get the 8400 at 8x?
 
The multi on the e8400 is locked upwards, but it can go as low as 6x. So the range is from 6x-9x.
 
Hehe RAM speed.... I ran 3DM06 tonight... 2 gigs of Mushkin at 800 and at 960... I got 10 points increase. lol

800 MHz - 13073
960 MHz - 13083

Surprised me, I thought I'd get a LITTLE more than that... guess not. :p
 
I love my 2 x 1GB DDR2-1200 5-5-5-12 2.3v(I run at 2.2v, still cool to the touch at 2.3v, the amount of voltage required between 800 and 1200 is almost linear from 1.8v at 800 and 2.2v at 1200) PARTIOT memory... when I go from 800 to 1200 at the same timings I get a % or two in the more memory hungry applications.

I run 1:2 so everything works out with my C2Q at 2.7GHz on stock cooling at stock volts.
 
Hehe RAM speed.... I ran 3DM06 tonight... 2 gigs of Mushkin at 800 and at 960... I got 10 points increase. lol

800 MHz - 13073
960 MHz - 13083

Surprised me, I thought I'd get a LITTLE more than that... guess not. :p

Running my RAM at 1250 got an extra 0.02 of off a PI 1M run over 900MHz. Because of the volts needed to do 1250, I just went back to 900.
 
"Hehe RAM speed.... I ran 3DM06 tonight... 2 gigs of Mushkin at 800 and at 960... I got 10 points increase."

"when I go from 800 to 1200 at the same timings I get a % or two in the more memory hungry applications."

"unning my RAM at 1250 got an extra 0.02 of off a PI 1M run over 900MHz."


Yes, RAM speed doesn't matter much in those types of apps, and RAM speed is pretty irrelevant on a dual core system. However, on a quadcore system, depending on what applications you are running, when all 4 cores are at 100% load the difference becomes noticeable. For example, I notice about a 12% difference between DDR2 800 4-4-4-12 and DDR2 1066 5-5-5-15. There are probably programs that push that 12% a little higher even. However, the benchmark Cinebench R10 seems to get the same score regardless of how fast my memory is.
 
I bought a pair of mushkin's 1066 2x1GB awhile back. They run nicely at 4-4-4-12 1066 @ 2.35 (ouch) or 3-3-3-8 1T 800 @ 2.275 on a 780i board.
 
Back