OK so I have various computers of varying processor speeds and memory speeds.

Now i know that various processors excel in certain tasks, I remember that long ago Athlon XP's were rated better than pentiums because pentiums were geared up for multimedia stuff whereas the athlons were good with the arithmetic unit sowere faster crunchers.

Things have really moved on since then with DDR, DDR2 and DDR3 memory and multicore CPU's all presumably fighting over memory.

Is there an ideal CPU (core) speed to FSB/memory speed ratio (or portion of for multicore CPU's) ? would this vary with task ? I seem to remember from somewhere that the CPU load figure/percentage is actually theload on the FSB not the core, obviously if the FSB is saturated the CPu is maxed out for sll practical purposes so the figure can be misleading.

Same goes i guess for video cards, I have a GT520 GPU and running GPU grid i think it only loads half the memory speed.

---

just to give an example - I run rosetta@home on my dads pc, now it has 2 cores in the proccessor (C2D @ 1.86 GHz / 1066 FSB and a dual channel 800 MHz DDR2 memory bank). now when I let it use both cores things run, if i force it to use only one core then the remaining task works faster although it has the same CPU power as before, the only difference is that it now has full access to the FSB and memory bus so i assume that 533MHz is not enough FSB throughput for a 1.86GHz core on that project, of course either way the processor "appears" to by running at 100% load but i think we are possibly talking about FSB load not core load.

---