Needs assistance. My Athlon X4 860K OC isn't running smoothly.
Needs assistance. My Athlon X4 860K OC isn't running smoothly.
I've attempted to boost the multiplier in BIOS one step at a time while disabling Turbo. Reached 4.1 GHz, but needed to increase the CPU voltage to 1.4438 to achieve stable performance. Any further adjustments caused instability. I slightly raised the CPU number and it crashed with a BSOD. Someone with useful data would be very helpful. My goal is to hit at least 4.2 GHz consistently. Thanks.
Also note that depending on the motherboard and power phase quality, throttling might begin as early as 10-15°C TM. At 0°C TM, you should definitely anticipate throttling or shutdown, since the chip is attempting to safeguard itself. On my mATX D3h (with copper heatsinks on the VRMS), at 4.5GHz, throttling would start as early as 20°C TM. On my Crossblade Ranger, at 20°C TM, there was no throttling at all. Unfortunately, I cannot discuss anything below 20°C TM, because my cooling never allowed the chip to drop that low.
Your cooling setup is more than sufficient for this chip; unless there are other issues—such as VRMs overheating or insufficient power delivery—the results should match.
There is no guarantee you can reach 4.2. Its the silicon lottery. Your chip may be a dud.
Having said that, you still have voltage headroom (between 1.5-1.55v is acceptable). As long as your cooling is good then I would suggest raising your voltage some more to get in stable.
Another trick I like is using the base frequency. Drop your multiplier by 1 and start raising you base freq.. Note that you will have to lower your RAM multiplier and your hyper stransport and northbridge to below stock so that when yo overclock the base frequency you arent overclocking your whole system. That method got my a10 6800k from 4.7 to 4.829. Its a little boost but it helps.
I need to examine that further.
A peculiar occurrence took place regarding the temperatures later. While monitoring thermal margins in AOD, I increased the multi to 42 and maintained the CPU-NB voltage at 1.1375. I was performing IBT for stress testing. The temperature gradually began to drop steadily to 0°C! Then it entered negative values. CoreTemp, which was also accessible, followed the temperature decline by rising at a similar pace to reach a TjMax of 70°C and beyond.
The more unusual aspect is...
the rad fan on my Corsair H60 remained unchanged.
I expected it to increase to full speed if temperatures were actually rising in that manner.
(I have connected the rad fan directly to the CPU fan header (auto) and the pump to a SYS fan header, which is set in BIOS to 100% continuously.)
I understand the temperature sensors on APUs and Athlons can be unreliable, sometimes misinterpreted by software. Still, it was strange. The pump felt only slightly warm, while one hose was cool and another was only mildly warm.
I repeated the test this time, manually setting the rad fan to 100% (and loud). The same problem persisted. No crash occurred, but the temperature readings seemed completely off. I’m really questioning the accuracy of those readings.
This indicates that your system is overheating. CoreTemp isn't displaying accurate temperatures. The only way to see the actual temps is through thermal margins. If your thermal margin falls to 0°C or lower, it means your CPU is overheating. Thermal margin doesn’t actually read the temperature—it shows how much you’ve exceeded the safe limit before overheating occurs. At 0°C, you’re at your maximum CPU temperature. Therefore, if the thermal margin reads 10°C, your CPU is operating 10°C below its upper limit. If it reads -10°C, that means you’re exceeding the CPU’s maximum temperature by 10°C.
I suspect there might be a typo in those voltage settings (though I’m not certain—I haven’t tested a Kaveri yet). Most AMD CPUs have a starting voltage between 1.3 and 1.35 volts.
I’m sorry to ask, but did you apply thermal paste to your cooler? That could be the issue. Also, make sure both fans are spinning in the same direction when using a push-pull radiator setup.
It’s also possible you have a faulty pump in your cooler. I doubt it, but it’s worth checking.
I restored everything to the BIOS default and began again. The unusual temperature problem is completely resolved now. (??)
Currently at 4.0 GHz, using just the multi. TM remains around mid-30s with IBT. Core temp is also in the mid-30s, which suggests it matches the TM. Combined they reach about 70°C, which is the maximum temperature for that CPU.
I believe the previous abnormal readings were inaccurate. Then, AOD would display the idle TM at 70°C and then drop to 0°C when loading. As I mentioned earlier, the pump stayed cool and the hoses were fine. The radiator fan didn’t increase speed. It was difficult to tell which hose was intake and which was discharge because the temperatures were so low.
I only have one fan on the cooler. (Standard setup for the H60) The fan is intake through the radiator into the case. The top fan is exhaust. The front fan is intake.
Thermal paste is the original manufacturer’s product.
Check if the 2133 MHz RAM with 1.6V voltage works with your Athlon memory controller.
I don’t believe the RAM would lead to any recognized issues.
I’ve mentioned this before and I’ll restate it; CoreTemp isn’t compatible with fm2/fm2+ APUs or CPUs. Avoid using it. The only reliable method is using Thermal Margins in CCC. You might also try your motherboard’s hardware monitoring tool to track socket temperatures, which should align closely with CPU temperatures (within a 5°C range).
I’ll remind you once more just in case you weren’t clear before.
Your thermal margins don’t display your CPU temperature! They reflect the difference between your CPU’s maximum temperature and its current temperature. For example, if your CPU can reach 70°C and you’re at 60°C under load, the margin will be 10°C. That’s straightforward math.
CoreTemp has a different approach—it measures your actual CPU temperature. The issue arises because CoreTemp doesn’t correctly interpret the temperature sensor on fm2/fm2+ CPUs. This can lead it to report a CPU temperature of just 30°C during overclocking, which is inaccurate.
You should generally rely on your motherboard’s hardware monitor for idle temperatures, as AMD Overdrive doesn’t measure them accurately. That’s why it shows a thermal margin of 70°C at idle, implying your CPU would run at -10°C under load—impossible without extreme cooling.
Please review the link I shared; it clearly explains what I just explained. You’re likely misinterpreting your readings.
Memory speed capabilities rely on your specific Mobo model, though it seems nearly all FM2+ boards can handle up to 2133mhz. Numerous cases have shown that overclocking with 2400MHz RAM can lead to instability, but adjusting to 2133mhz using XMP or standard settings often resolves these problems.
The temperature concerns might have been resolved after a BIOS reset. If the fan speed setting in BIOS was incorrect, it would not have increased beyond its minimum limits. The way you note the changes before and after the reset matches what would happen if that were the situation. I’m not accusing you of lying about your BIOS fan settings, but it’s possible something was overlooked that kept the fan at idle speeds.
The unusual readings might actually be accurate. During idle with little fluid flow and minimal fan activity, at the speeds and voltages mentioned, your chip could still have full TM available, and at those exact conditions, you’d observe the TM dropping to zero as it overheats quickly under load. A draw of 4.2 is relatively low; these chips usually need a much higher voltage increase once you exceed 4.2. On my system, I can run 4.2 with an undervolt of 1.275, but 4.3 needs 1.425V, and 4.5 requires 1.51V to stabilize. The chip’s characteristics mean its stock settings are already pushed to their limits when targeting TDP.
In short, it looks like you’re making progress—keep pushing, and good luck! In the worst case, you might end up with a defective board, but even at 4.2 you’ll still achieve solid performance compared to the standard configuration.
Should also be noted that depending on the motherboard and power phase quality, throttling might begin as early as 10-15°C TM. At 0°C TM, you should definitely anticipate throttling or shutdown, since the chip is attempting to safeguard itself. On my mATX D3h (with copper heatsinks on the VRMS), at 4.5ghz, throttling would start as early as 20°C TM. On my Crossblade Ranger, at 20°C TM, there was no throttling at all. Unfortunately, I cannot comment on anything below 20°C TM, because my cooling system never allowed the chip to drop that low.
My cooling setup is quite strong for this chip, so you should experience similar temperature thresholds unless there are other issues, such as VRMs overheating or insufficient power delivery capability. My board has a 4+2 configuration, which is typical for mATX boards (there are no mATX boards with better than 4+2), possibly contributing to your limitations.
Also, a rare feature among these chips is the ability to optimize the NB with stability. I’ve experimented with several different 860Ks, and only one (the one I kept for the XBR) can stabilize the NB at all, even then only up to 2000.