Athlon x4 860k not stable past 4.0GHz?

[epicman2456]

So recently, I've been trying overclocking my CPU because I got a new cooler and I just wanted more performance out of my system. I was aiming for 4GHz, and that was fairly easy so I tried to get it to 4.2GHz. When I raised my multiplier to 41, and stress tested, 1 second in, a crash WHEA_UNCORRECTABLE_ERROR so I upped the voltage to 1.386v (@4GHz it was stable @1.380v). Still crashing, this time with CLOCK_WATCHDOG_TIMEOUT, so I upped it to 1.392v, still crashing. 1.398v, crash, 1.404v, crash but this time it wouldn't stay at 1.404 during the 30sec of stress testing it did before it crashed. CPU-Z showed that it was going from 1.392-1.404v which is a whole 0.012v, mostly staying at 1.392v. I think it's a PSU problem but I'm not 100% sure. The temps are stable, I use AMD Overdrive to check temps and to do stability tests and the thermal margin stays at around 32C-28C. That's the highest I've ever seen it go, and it only goes there when stress testing. Any tips? Thanks

Specs:
CPU: Athlon x4 860k
GPU: R7 265 (OC Core: 1050MHz Mem:1500MHz but only when playing games)
Cooler: Cooler Master Hyper T4
Thermal Paste: Arctic Alumina
PSU: 500w EVGA B1 PSU
Mobo: Gigabyte F2A88XM-D3H
HDD: Seagate 1TB 7200RPM
Case: Corsair Carbide 100R

 

[Shaun o]

Yes it could be down to the power supply unit of the system not being enough.

But both of the errors you were seeing were in relation to the speed the Cpu core was running at.
More often related to the cache levels of the cpu it`s self.

Or a sync error between the Cache and the cpu core frequency set.
It can also relate to if the cpu has the DMA memory access controller on the cpu die it`s self to the main system memory of the system, again it is out of synchronization.

So you would have to in some cases lower the speed of the main system memory of the system in the bios.
Or as a result choose a lower HT speed.
To try to keep things in sync to avoid the error, adding voltage to the cpu core is not the answer.
Keeping the synchronization of the cpu HT and memory speed is.

It`s like you throwing a ball for example to a friend, but he is looking away at the time and the ball is not caught by him because the timing was off. So the ball drops or is not caught. Because one of you is out of sync.

The result is the data is missed or it can be corrupted in the process in relation to data transfers and requests, from cpu to main system memory, and back again.

 

[epicman2456]

Yes it could be down to the power supply unit of the system not being enough.

But both of the errors you were seeing were in relation to the speed the Cpu core was running at.
More often related to the cache levels of the cpu it`s self.

Or a sync error between the Cache and the cpu core frequency set.
It can also relate to if the cpu has the DMA memory access controller on the cpu die it`s self to the main system memory of the system, again it is out of synchronization.

So you would have to in some cases lower the speed of the main system memory of the system in the bios.
Or as a result choose a lower HT speed.
To try to keep things in sync to avoid the error, adding voltage to the cpu core is not the answer.
Keeping the synchronization of the cpu HT and memory speed is.

It`s like you throwing a ball for example to a friend, but he is looking away at the time and the ball is not caught by him because the timing was off. So the ball drops or is not caught. Because one of you is out of sync.

The result is the data is missed or it can be corrupted in the process in relation to data transfers and requests, from cpu to main system memory, and back again.

So that out of sync thing, it would explain most of the errors I got?

But what about the CPU voltage not being constant when under load? Would lowering HT speeds fix it? And how would I lower memory speeds? Does these mean underclocking the RAM?

Also I forgot to add one thing. My RAM is 8GB, but it's only 1 stick of 8GB, not dual-channel.

 

[epicman2456]

Yes it could be down to the power supply unit of the system not being enough.

But both of the errors you were seeing were in relation to the speed the Cpu core was running at.
More often related to the cache levels of the cpu it`s self.

Or a sync error between the Cache and the cpu core frequency set.
It can also relate to if the cpu has the DMA memory access controller on the cpu die it`s self to the main system memory of the system, again it is out of synchronization.

So you would have to in some cases lower the speed of the main system memory of the system in the bios.
Or as a result choose a lower HT speed.
To try to keep things in sync to avoid the error, adding voltage to the cpu core is not the answer.
Keeping the synchronization of the cpu HT and memory speed is.

It`s like you throwing a ball for example to a friend, but he is looking away at the time and the ball is not caught by him because the timing was off. So the ball drops or is not caught. Because one of you is out of sync.

The result is the data is missed or it can be corrupted in the process in relation to data transfers and requests, from cpu to main system memory, and back again.

Quick Update: I ran BlueScreenView to see what caused the crashes and this is what came up:
http://imgur.com/a/2dBGs

Going bottom to top is each time I increased the voltage by 0.006v. Most of them are caused by atikmdag.sys which when I looked it up, it had to do a lot with my GPU. I'm still not sure what.