• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

dual channal

  • Thread starter Thread starter CU
  • Start date Start date

CU

Platinum Member
I just upgrade to an ASROCK 939DUAL SATAII and opteron 144. I cannot get my memory to boot in dual channel mode at 200mhz, it will work at 174mhz. Both memory modules are Corsair XMS3200c2, but they are different versions. After reading Corsair's support forum that seems to be the problem. Are all memories like this ie, OZ, Patriot, Kingston, etc.? This makes it very hard to upgrade ram later on if you have to sale your old ram since the new ram will probably be a newer version than what you had.
 
Dual channel RARELY works unless every aspect of the memory chips is identical.

Think about it; the chipset has to move data across both sticks of RAM simultaneously. So the RAM should be the same clock speed, same CAS, same RAS, same CTR, same density... or the data is not going to move across the RAM at the exact same rate.

Naturally, if one stick has a CAS, for example, that's a little slower than the other, then lowering the clock speed on both sticks may allow both to run at the same refresh rate and therefore run in a dual channel configuration.
 
I figured it would have to be the same speed, cas, etc. But Corsair says it has to be more than that. It has the be the same version also meaning same chips and pc board. I can set the speed, cas, etc. in the bios and it still will not work in dual channel which is what Corsair says. That just surprised me since I didn't know it. I am not even sure why it is true because if the memory is accessed at the same speed, cas, etc. why does it care what chips or pc board the memory is built from?
 
Well, they can say that it has to be the same version as a CYA. You see, sometimes chips can be the exact same, but the SPD (a BIOS chip for DIMM's, so to speak) may be programmed differently from version to version. But the solution to that is easy: Manual memory timings. If the BIOS doesn't read the SPD, then the SPD programming doesn't make a difference. But if Corsair supported every Tom, Dick and Harry that wanted to manually set their memory timing, they'd lose money in support! 🙂
 
I have tried manual memory timings and that does not fix the problem. Corsair also says this will not always work. I do have I thought why it does not work though. My guess is that the trace lengths on chip A compared to chip B and on pc board A compared to pc board B may vary therefore causing a slight timing difference. The timing difference would be a problem when you are trying to get data from chip A pc board A and chip B pc board B at the same time (dual channel) because it would a arrive at slightly different times. This would difference would be a bigger problem as you increase in speed. Does that make sound right?
 
Nope, it's propbably because the signal is lagging too much at higher speeds, if both sticks lagged equally it wouldn't be a problem but if one lags and the other doesn't...

Complete guesswork, but the path length isn't that much further and the signals are so low frequency that it's irrelevant.
 
Originally posted by: CU
I have tried manual memory timings and that does not fix the problem. Corsair also says this will not always work. I do have I thought why it does not work though. My guess is that the trace lengths on chip A compared to chip B and on pc board A compared to pc board B may vary therefore causing a slight timing difference. The timing difference would be a problem when you are trying to get data from chip A pc board A and chip B pc board B at the same time (dual channel) because it would a arrive at slightly different times. This would difference would be a bigger problem as you increase in speed. Does that make sound right?

No; the DIMMs have to respond within the clock rates of the memory controller -- as long as they do that, it should work. Basically, all a dual-channel memory controller does is to interleave your system's memory addresses between two or more memory modules. They're accessed completely independently.

It shouldn't matter beyond the modules being the same size and being able to run the same timings. You could also have problems if one is single-sided and one is double-sided, or the chip densities are different. But different PCBs or DRAM chips? Shouldn't be a problem.
 
Originally posted by: Matthias99

No; the DIMMs have to respond within the clock rates of the memory controller -- as long as they do that, it should work. Basically, all a dual-channel memory controller does is to interleave your system's memory addresses between two or more memory modules. They're accessed completely independently.

It shouldn't matter beyond the modules being the same size and being able to run the same timings. You could also have problems if one is single-sided and one is double-sided, or the chip densities are different. But different PCBs or DRAM chips? Shouldn't be a problem.

Well it looks like I was wrong about the trace lengths since no one agrees with me. I do agree that different PCBs or DRAM chips shouldn't matter, but Corsair says it does. Just wondering why. By the way here are the numbers off my two 512meg sticks:
xms3202v3.1
0411040-0

xms3202v1.1
0252042
 
I did some more testing and each module will run at 200mhz 2-3-3-6 2T in any slot alone. But when I put them together in dual channel or single channel the pc will not post at that speed. So the problem is not with duel channel, but with the ram just not working together. For now I have them running at 176mhz 2-3-3-6 1T dual channel. Do you think corsair would trade me for matched pairs so I could run them at their designed speed together?
 
Back
Top