DDR2 800 Bottle Neck

[411tech]

Want to upgrade to A Phenom II X4 944 be.

I have 4gigs of ddr 2 800 in my present system. In the interested of being the tight fisted money grubber that I am :sneaky:saving money on recycling this Ram into my next build would be my preference.

I have searched around but haven't found any articles or benchmarks on whether or not ddr2 800 will kill the performance of phenom II x4 955.

If I get a am2/am2+/am3 Mother Board drop my old ddr2 800 in it and throw in the Phenom II X4 944 be. am I going to bottle neck my CPU.

I also will be moving my ATI 5850 to this new system.

Thanks in advance!! :hmm:

 

[VirtualLarry]

nah, I don't think that it will bottleneck it much.

 

[411tech]

nah, I don't think that it will bottleneck it much.

I don't mean any disrespect but I think that is the most technically inclined explanation of a point of opinion ever devised. :|

 

[dorion]

I upgraded my Phenom II 965 from 4x2GB DDR2 800 to 2x4GB DDR3 1600. The performance I can see from the memory upgrade is minimal and probably more to me running the memory at 7-7-7-22 1T. The stock settings performance on either type of memory were pretty much identical(800 5-5-5 2T vs. 1333 9-9-9 2T).

On the other hand the increase in motherboard quality and addition of an 8-pin CPU power plug on the DDR3 motherboard allowed a 200MHz overclock, and my above mentioned tight ram timings.

 

[411tech]

What were your original ddr2 timings?

Thanks for the replay, I appreciate it ()

 

[Voo]

You'll see DRAMATIC improvements.

Oh wait you didn't say you run Super Pi for a living? Ah my bad No, the difference will be minimal - an opinion which as it happens to be is based on real world benchmarks and experiences which are quite more accurate than pure technical explanations, which for such complicated situations are always vastly simplified or completely non comprehensive.
But we can try anyhow: Most user programs contrary to HPC stuff aren't particularly bandwidth limited and caches are large enough that we only get relatively few cache misses which are easily masqueraded.