You're gonna want to really check the temps thoroughly with that thing, I've found from past experience that thermal pads tend to really suck compared to decent thermal paste, more so if it's not covering the entire surface of the CPU (is it?)
Also, what sort of cooling do you have, is it overheating?
My rig was getting a little warm recently because dust had accumulated in the rad for my H50 - bear in mind this is only eight weeks worth, in a relatively clean flat:
Lol at "mouse surface"
Isn't exactly going to be necessary in diagnosing this...
First, have you checked that overclock is stable?
Run OCCT for an hour. 1.288v @ 4.2Ghz isn't something I'd bet on being stable.
I would personally advise trying to get it a little more stable than this (ie, it should be able to run linpack tests like linX on maximum for a decent number of runs and get no errors...)
Otherwise you're not really stable.
Sweet Jesus you got ripped off!
Also, you need to be using something other than AvP to test stability with. When that Configuration can stand up to an hour or two of OCCT linpack at maximum memory usage then you can call it stable.
It's not "on the verge of high" - it's a voltage that will probably cause degredation of the chip - shouldn't take too long either.
cadaveca even demonstrated this :p
1.5 is considered relatively safe, wouldn't go much over that...
This is probably the best OC I've seen on a graphics card using the reference cooler (although it loses stability under full load at >~1015Mhz or over 1.3v, possibly due to restrictions in the amount of power it can draw from two 6-pin sockets and the motherboard).
I advised a friend to get one...
Considered the Crucial C300?
Rated read speed of 355MB/s and Write speed of 215MB/s, not bad.
Mine's doing 3.8Ghz right now at 1.256V vcore according to CPU-Z, goes up to ~1.28v under load
It will do 4Ghz with about 1.31v idle 1.34 load, or 4.2Ghz with 1.38v idle 1.41v load.
Beyond that my cooling isn't up to the task :P
Can you still get it through furmark at those clocks?
I figured mine was using more than the 225W the connectors can supply - I can do 960Mhz in furmark at 1.237v or 1.25v, but if I go up to 1.262 it crashes immediately. However it will do lower clocks at 1.262 or 970Mhz at 1.25v without issue...
What IS a safe limit for VTT?
I've never heard of anyone really taking things above 1.4, and my motherboard says "not recommended" for any setting above 1.35, but to get my ram to run at its top speed (2133mhz) I have to push VTT to 1.39v
You're doing 950Mhz core at 1.062v? Impressive :D
If I push mine to 1000Mhz I have to go to 1.262v, at 1010Mhz it takes 1.274v and then beyond that I run into the limit of what the power connectors can supply as it's consuming ~225W under load :D
I have the same problem which started on my board (MSI P55-GD65) when I went from 4GB of ram to 8. The board would only recognise 4GB on my old speed/timings. Was fixed when I upped my timings a bit, but it is still annoying.
Now working with 8GB @ 1520Mhz 7-7-7-21