bornfree wrote:
>Fishface wrote:
>> But I thought you had one stick of "PC2-4300" rated for 266 MHz.
>
> Yeah I do. Shall I ditch it as a temporary solution until I run
> memtest.. That way I can just use the 400Mhz RAM, at 400Mhz.
I'd probably try it first. Bump up the voltage, try 333 MHz and jump to test
#5. If you don't get any errors highlighted in red after a couple passes,
try 366. Then 400. If you get errors, drop it by half the previous jump to
zero-in on the highest stable speed. Finally, I ususally let it run overnight
or all day while I'm at work (on all tests) to be certain. By knowing the limit,
you can be more confident as to how close you are actually treading from
the edge. Then if Orthos gives errors in Windows, it is usually more CPU
voltage that is required and not something to do with the memory stability.
As I said previously, the inexpensive memory is often the same as the higher
speed memory, only untested or operating at a reduced voltage. It will often
perform much better than its rated speed. This will, of course, vary by brand,
part number and batch of the chips.
If your lower speed stick proves to be very limiting, you could remove it, but
you would be losing a lot of memory bandwidth because of the dual channel
nature of the memory controller. That will hurt in some applications, and not
make a bit of difference in others. Usually, I believe you would benefit more
from the higher CPU speed.
>> Stay informed about: Is this safe?