Clock speed versus memory throttling tradeoff?

[SILENCER00]

I'm just curious what will yield higher performance overall: lower clock speed with memory running 1:1, or higher clock speed with a memory divider?

Example scenario being: 165 Opterpwn at 9x275 @ 2.48Ghz with a 1:1 memory timing or 9x310 but knock the memory divider back to 166 so it runs lower at 253 but with a higher cpu clock speed of 2.79Ghz? (full bandwidth mem vs. 300meg overclock and 17% mem bandwidth loss)

 

[J-Mag]

I'm just curious what will yield higher performance overall: lower clock speed with memory running 1:1, or higher clock speed with a memory divider?

Example scenario being: 165 Opterpwn at 9x275 @ 2.48Ghz with a 1:1 memory timing or 9x310 but knock the memory divider back to 166 so it runs lower at 253 but with a higher cpu clock speed of 2.79Ghz? (full bandwidth mem vs. 300meg overclock and 17% mem bandwidth loss)

Higher clock speed is almost always better, the AMD chips aren't bandwidth starved really. There might be some niche applications out there which show better performance w/ high bandwidths vs clock speed (like Sandra unbuffered benches, but who cares?)

 

[SILENCER00]

I like that answer. Thanks!