Running K8T Neo2 in single channel mode.

bob661

Senior member
Oct 20, 2004
425
0
0
Does anyone know if there is a performance hit when running in single channel mode on a 939 board? Thanks.
 

joelslaw

Senior member
Dec 9, 2004
466
0
0
there is always a performance hit running single vs. dual. (that's why they call it dual :D ) that's one of the main advantages of 939 over 754.
 

bob661

Senior member
Oct 20, 2004
425
0
0
Well I meant a noticeable one. A friend is building one of these for one of his friends and the other guy can't quite swing for the extra ram. I was wondering if it would run like a dog or not. Thanks for the reply.
 

joelslaw

Senior member
Dec 9, 2004
466
0
0
that's a tough question, since a lot of this is subjective. A lot also depends on what you'll do with it. Some programs like memory bandwidth more than others. But I think most everyone (no matter how big of a power user) would agree that it won't "run like a dog". In most cases (if I remember correctly) the biggest difference you'll see is about 5%. In games, the effect is negligable. I believe encoding programs (mp3 rippers, dvd rippers, trans-coding software, etc.) are where you would see the biggest benefiet.

Something you might think about if you really want to run dual channel, is getting the same amount, just in two sticks instead of one. For example: the cheapest 512 stick of pc3200 I found on newegg was around $60, where as the cheapest set of 2x256 pc3200 sticks were around $75. So you only have to pay $15 extra bucks for dual channel. Just have to ask is %5 worth the $15 bucks for you. Maybe it is, maybe it isn't.

I'm the kind of person who would pay the extra $15 just to make sure I was going to be happy with what I got, instead of always wondering if I would have liked dual better than single channel. But it's equaly reasonable to think that you could save the $15 and run at a speed so close to dual, that only a benchmark can tell the difference.

n e way, that's enough babbling from me. I have been looking around trying to find benchmarks that support the "5%" thing, but for some dumb reason can't find any. I'll keep looking and if I find some, I'll post them. HOWEVER, it has been shown that senthetic benchmarks can sometimes exagerate the real world effects of better (ie. more expensive) ram, or any other component for that matter! To see a good example of this, look at Zebo's sticky in CPU/OC, here:
http://forums.anandtech.com/me...5190&enterthread=y
 

bob661

Senior member
Oct 20, 2004
425
0
0
Thanks for the info. The buyer doesn't do anything that would push the machine currently. But that may change once he notices how much quicker his new computer is.
 

joelslaw

Senior member
Dec 9, 2004
466
0
0
Originally posted by: bob661
Thanks for the info. The buyer doesn't do anything that would push the machine currently. But that may change once he notices how much quicker his new computer is.

heh heh heh, didn't we all!