Lowering the CPU Multiplier

[Usaybis]

I know that many people (including myself) lower their multiplier in order to acheive higher FSB ratings. Does lowering the CPU multiplier have any down side? Are there any performance costs?

 

For instance if I had 2 CPU's both running at 3.0 Ghz and one was using a multiplier of x8, and one was using a multiplier of x11, would one perform better?

 

If the the multiplier has no impact on performance, than you could for instance buy a cheaper C2D like the E4300 up it's FSB to 1066 in the BIOS and reduce it's multiplier to x7 and have a yourself an E6300?

 

When overclocking both would there be any real difference? Does pushing the FSB beyond the CPU's native support place more stress on the chip? Allow me to illustrate:

 

You take an E4300 and an E6300 and clock both at 375 (1500 FSB) x7 to give you 2.62 Ghz. Each chip has the same clock speed and same FSB throughput. It seems you could get the exact same results but with less cost.

 

Since both of these chips are designed to run at the same speed, just a different native FSB's is there any real difference here?

[road-runner]

I know that many people (including myself) lower their multiplier in order to acheive higher FSB ratings. Does lowering the CPU multiplier have any down side? Are there any performance costs?

 

For instance if I had 2 CPU's both running at 3.0 Ghz and one was using a multiplier of x8, and one was using a multiplier of x11, would one perform better?

 

If the the multiplier has no impact on performance, than you could for instance buy a cheaper C2D like the E4300 up it's FSB to 1066 in the BIOS and reduce it's multiplier to x7 and have a yourself an E6300?

 

When overclocking both would there be any real difference? Does pushing the FSB beyond the CPU's native support place more stress on the chip? Allow me to illustrate:

 

You take an E4300 and an E6300 and clock both at 375 (1500 FSB) x7 to give you 2.62 Ghz. Each chip has the same clock speed and same FSB throughput. It seems you could get the exact same results but with less cost.

 

Since both of these chips are designed to run at the same speed, just a different native FSB's is there any real difference here?

The problem you will run into with the lower multi will be the need for higher clocking RAM. It is also harder on the motherboard and RAM as both will probably require more volts to run the higher FSB.

[Verran]

For instance if I had 2 CPU's both running at 3.0 Ghz and one was using a multiplier of x8, and one was using a multiplier of x11, would one perform better?

There are differences, but they are very small. In almost every case, you'll never know the difference. I have benched my chip at 9x, 10x, and 11x, all at the same total speed, and the differences are pretty much imperceivable, even in synthetic benchmarks. In a game, for example, I can pretty much promise you you'll never see the difference.

 

For me, I used a lower multi because it let me overclock higher. My chip OC's better with 10x than 11x, so I go with 10x

 

You take an E4300 and an E6300 and clock both at 375 (1500 FSB) x7 to give you 2.62 Ghz. Each chip has the same clock speed and same FSB throughput. It seems you could get the exact same results but with less cost.

Actually, that was one of the biggest appeals of the E4300 when it came out. The lower stock FSB left more room to OC, and it still had the 9x multi, so really it could clock to speeds very close to what an E6600 would, and for about half the price. The only problem is that not all chips OC like mad. While some E4300s clocked like mad, others didn't. You're just never quite sure what you'll get when you buy a cheap chip in hopes of a 50% OC.

 

The other thing to keep in mind with your example is that the E4300 had less L2 cache. So even clocked like a E6600, it never fully matched it, which was one of the other reasons it was notably cheaper.

[Usaybis]

Thanks for the information. I appreciate the quick responses.

 

The problem you will run into with the lower multi will be the need for higher clocking RAM. It is also harder on the motherboard and RAM as both will probably require more volts to run the higher FSB.

 

Got ya. I'm not to concerned about added stress on my Mobo, the 680i seems to handle it quite well.

 

Actually, that was one of the biggest appeals of the E4300 when it came out. The lower stock FSB left more room to OC, and it still had the 9x multi, so really it could clock to speeds very close to what an E6600 would, and for about half the price. The only problem is that not all chips OC like mad. While some E4300s clocked like mad, others didn't. You're just never quite sure what you'll get when you buy a cheap chip in hopes of a 50% OC.

 

Seems that way to me as well. Could be why Intel released the E6320 with the full 4mb of L2 which is actually selling for less than the original E6300.