I'm using Elixir 3200 RAM and found it ok - actually overclocks from 200MHz to 238MHz which is surprising and pleasing as it is also very cheap!

I have a 512 MB module: Serial

M2U 512 64 DS8HC3G-5T

and just bought another module

M2U 512 64 DS8HC0G-5T

The only difference being the "0" vs the "3" before the final G. Looking at the Elixir site they have a module naming rule page,

http://www.elixir-memory.com/sup_name_rul_2.asp

and the "0" vs "3" is in the module revision digit which is described as (PCB rev. & type). I presume PCB is "printed circuit board" so I guess the module revision should then have no effect regarding dual channel stability on my NF7-S (Rev 2) motherboard? I've had instability before using different die generations..

Any comments would be much appreciated.

Thanks

Andrew

Hi
i havnt had any issues with mixing Elixir
with any other brands (samsung etc) so at that price :) chance it! with a few tests
memtest , a game or two.
Ocing may throw odd stuff up but thats the chance you take with 'value' ram.

regards

Thanks, I'm running it in dual channel mode on another computer at 266 MHz and so far it is fine. Just wondering if anyone can shed light on the module revision - sounds more like the circuit board it is mounted on than anything fundamental..

Thanks

Andrew

By 266, I mean 133 x 2! ie 133 MHz bus

Hi
again
i had no problem at 166 fsb with Elixir
i ran a 2.8 athlon asus a7v8x fsb rated 333
i followed the link here ,
http://www.lostcircuits.com/cpu/amd_xp2800+/

worked a treat,athlon stable @ 2225 however when i tried other cheapo brands hynix & matrix it failed miserably,
im not really into Ocing but this was worth the effort possibly elixir is the better cheaper stuff (have since moved on to corsair in newer rig as i cant seem to get matching pairs in value sticks that run 1T.)

:)

having successfully tested at 266 MHz, I put it into the PC the memory is for and it booted up in dual channel mode at 400MHz and seems absolutely stable. So, in this instance, module revision (PCB version) seems to make no difference, probably the "generation" of memory that really matters..

Thanks all

Andrew