Right, back on deck and am going to give the RAM change a try!
Going to enable XMP and alter the DRAM Voltage to 1.59 - will report back with how stable it is in games etc!
Just as an aside, I've downloaded AIDA64 Extreme from cnet (the actual AIDA site only seemed to have versions for Android, iOS, Ubuntu and Windows phone!?!?) - it's given me a 30-day trial but I have no idea what to "run" to stress test/benchmark the system - any pointers?
Also, is there a decent free benchmarking tool I can use to get a base reading of my system performance to hopefully gauge what sort of difference these tweaks are making??
Thanks again for the help guys....
Mornin' Stevie
There's a few out there Aida64, Cinebench and Unigen Valley is a good free tool as mentioned above.
I have a slightly different approach to harrry for testing clocks - we both agree in principle that the day to day running being the absolute true test. The reality is that a PC can pass all the stress tests you throw at it - (e.g. 24 hr synthetic benching used to be the norm) - and yet go on to buckle when browsing or have cold boot issues etc. Overclocking can be tricky when pushing the limits (which we wont be).
There's no right or wrong approach as long as it leads to stability (and you don't slaughter your CPU in the process)- but i when i'm clocking 'virtually' via the web i've always found a 10 minutes 'blend test' (which stresses both memory and CPU synthetically) useful to quickly identify flaky clocks and especially heat issues. Yes, it's not a real world application of the CPU - but then it was designed for that very purpose - and if your system was at stock it would run the test all day long - even small ffts which crucify the CPU.
As an a side - if you think prime95 is bad harry - members use to swear by IBT about 8 years ago (never saw the point myself - like you with prime i guess
)- now that truly is a nuts synthetic test but some people still swear by it. Different strokes etc..
For example, ideally we would run your memory for a few days at it's XMP - to see if it's stable. If you started to get freezing or BSOD - we would pretty much be able to identify it as the memory straight away. Run memtest - job done.
However, as we're running your memory at the XMP (for the first time in 5 years) - doing a quick gaming test/bench test and then introducing a clock - there's going to be a nagging doubt as to which component is causing the error - the CPU, memory or combination of both (fun this clocking process).
But, if it passes blended stress test with the memory at XMP - we'll call it stable an move from there for speed.
EDIT: You've started already - i'll leave you to it. As mentioned either approach is equally valid (personally i like a mix) - as long as you're stable at the end of it. And as we've both mentioned it's the day to day stability that is the 'key' stability test.