New
#1801
Any time mate :)
Very true. Browsing, Word, e-mail doesn't care an iota for OCing. A few apps will but on the whole most light everyday tasks won't.
Since I game - I do notice a difference in some of them. But on the whole, like pretty much everyone else - it's only done for the fun/hobby aspect.Personally I've not seen any real world increase in performance with any overclock in CPU clocks or RAM fiddling. These things will get better scores in bench tests though.
RAM and BCLK fiddling however - essentially none at all. Benchmark increases are nice, but ultimately they are just for fun and shouldn't be taken to seriously.
* Just to clarify - that comment was not directed at anybody in particular :)
At the end of the day our PC's are just that - Personal Computers.
There is no one 'true' method or answer when it comes to testing.
I personally advocate my methods because I've found it's what works for me.
It also can depend on the OC you are aiming for. 'Light OC's' don't need as much testing. But higher OC's where more variables and potential weak spots are introduced do make it easier to narrow down culprits.
Case in point for the pro overkill testing - when I failed the 8hr prime the other night, I knew I was on the right track with my target - but not quite there.
I adjusted settings (more vcore basically and lowering BCLK), did a 10 pass IBT max run. Passed. Even did a 7 pass memtest run to make sure it wasn't the extra mhz on the RAM. passed.
Did a 99 pass Crysis harbour benchmark run (to simulate a more 'realistic load scenario) passed.
I'm now at 10 or so hours into another Prime run. The two cores that failed on two runs are now still going strong.
By rights, the machine is now more 'stable' than it was before. And unless it is a b*stard and x124's randomly, then I'll use the machine as is and 'real world' test it.
By overkill testing, I've narrowed down the weak point - voltage.
It's a method I've been using for years and have personally found works for me for the most part compared to quick and dirty testing.
Extended testing nailed the 'round off' bug.
11hrs - max vcore, temps seen
No errors :)
Now to sit back and 'real world' it
One thing I've learnt from this exercise - BCLK OCing is not really worth it.
The fact that PCI-E lanes (GPUs affected and other devices on PCI-e slots), sata, (possible odd drive flap ups) and Potential memory problems reinforce the idea the sticking to a BCLK of 100.00 and mainly adjusting via multis with SB is valid.
Decent score - but real world notice? Zilch
But overall, as long you keep it as hobby - go nuts!
(of course now that it's done - I want to try for higher simply because I can Plenty of temp /vcore headroom... )
One thing I've noticed since adjusting the BCLK is the infamous Asus POST turn on/turn off/ turn back on normal start.
Never occurred when I solely used the multi
Hmm, now that you mention it - mine has done it a few times, but it was very irregular. Certainly not regular enough to be noticeable. I noticed it most on cold boots after the power had been disconnected.
Now it's every cold boot, or restart.
My previous my P5Q PRO did it all the time. I don't recall my P6X58D doing it though. Initializing/ Pre-Post time has always been long with Asus.
They always fiddle and fart and multiple restart after certain changes. (Multi, volts and RAM especially). However sometimes they don't.
When I finish adding my new drives etc - I might reset the bclk back to 100.00 to see if changes anything.
It's not like there's any noticeable difference with 40 odd extra mhz anyway. I'd rather lose the minute mhz increase than double power on every single time
It just makes me curious as to what it's polling/prepping for the 2nd reboot - if that's what it's doing.
***EDIT***
Reset BCLK back to 100.00 and it appears to back normal.
So while you can 'safely' adjust SB using the BCLK with small increments - overall it's simply not worth it.