Possible processor woes

[dentedcapacitor]

Hi All,

 

I recently managed to save up to start on my dream build. But I since hit into a bit of issues that needs some second opinions. This is by no means my first build but I would call it my first enthuasist build. I am unable to provide screenshots as I am at work while writing this.

 

Specs as follows:-

CPU: Intel i7-4790K

RAM: Kingston HyperX BEAST 16GB Dual Channel Kit (HX324C11T3K2/16)

GPU: MSI GTX 980 4G Lightning

Mobo: Asus Maximus VI Gene (Latest BIOS: 1603, BIOS updated with no previous processor using Asus USB BIOS Flashback button, original BIOS was initial release)

PSU: CoolerMaster V850

HDD: Toshiba Q Series Pro 128GB SSD

Case: Corsair 350D

Case fans: 3x CoolerMaster JetFlo 120 PWM (Two in front of the casing as intake and one in the rear as exhaust, fronts run at 1200rpm and exhaust is purely PWM with 1600rpm average)

CPU Cooler: CoolerMaster Nepton 240M in pull configuration

OS: Windows 7 Home Premium x64

 

One of the first issues I encountered was the inability for the PC to boot up with the RAM set to its maximum XMP profile which is profile 1 with DDR3-2400 CL11-13-14 @1.65V. I was only able to boot it three times out of the ten times I selected it, most of the time it would display a Q-Code of 55. As a result, I am only able to run it at its second profile of DDR3-2133 CL11-13-13 @1.6V. I have read about issues of the IMC of the 4790K having issues with higher rated RAM and as such is a common problem. I have tried raising the System Agent voltage to 1.1V from 0.836V but it still won't boot. Even if it does boot with the first profile, there is gremlin like behaviour ie. the keyboard would stop functioning and from a cold boot the same cycle persists until the motherboard says overclocked failed in the BIOS startup. I have ran Memtest x86 on both the stock JEDEC DDR3-1600 speeds and DDR-2133 and there are no errors. What was weird was that I was originally able to install Windows 7 on the machine on its first boot with the 2400MHz profile after the training mode and the two reboots during the installation was successful.

 

Second of the issues is CPU heat related. While I was hoping for a good overclocker, I am somewhat satisfied with the chip I have gotten. On first boot (cleared CMOS), the stock VCORE was 1.088V. Basically all the settings were left on auto, as a result the final VCORE after booting into Windows 7 was 1.264V. This was at stock speeds of 4.0GHz with the turbo mode activated for 4.4GHz. Since this was a fairly new build I wasn't going to try and kill it that early on in its life, the only overclocking I did was to see how low of a VCORE the CPU can use for higher multipliers. The motherboard auto sets a setting of 1.299V VCORE for a multiplier of 47x when booted into Windows. The only stress test I ran was Intel XTU for 5 minutes (like I said earlier, not trying to kill it but to get a feel for the thing). This was wise because even with my aftermarket Nepton water cooler, it was reaching a max of 79C-84C on stock speeds based on info from XTU and HWInfo64. I have since exchanged coolers between the runnings of the test as the first one was having an annoying pump rattle. And the results are still the same.

 

Ambient temps mind you in my part of the world are about 33C average on a good day (tropical weather) but I run it in an air conditioned room with an average of 24C-27C. So I wasn't expecting much in terms of idle temps but 41C average on idle with the side case panel and 39C average without the side case panel is really not much the reason of buying a liquid cooling solution. Average temps on gaming load are 60-70C. I have checked the spread of the thermal paste twice on both coolers and they covered most of the IHS properly and tightened the pump adequately without damaging the board. I am using the stock thermal paste from the cooler for the time being but the temperatures should not be getting this high.

 

I have an older rig in the room with a E6600 with a stock cooler and el cheapo generic thermal paste and it is staying at 35C on idle alone and 56C on casual load. Since this was a backup rig it sat in a corner of the room with poor ventilation and an open side case panel in an enclosed space under a desk and quieter with generic case fans and regular HDD. Of course, the E6600 has an iridium solder instead of that TIM that Intel currently uses but still. I basically have a grace period of one week on the parts I purchased for a one on one exchanged. If the problem is the processor is kinda hard for them diagnose for overheating or memory problems and being an expensive processor I highly doubt they will be very willing to give me a new one.

 

So if anyone here can give me a hand, it will be much appreciated.

[kuhchuk]

It sounds like you might've just lost the silicon lottery.  It happens.  If you can get a replacement CPU, then go for it.

[Abir Vandergriff]

There's nothing really wrong with the heat. Just a hot chip. If you manually dial your CPU for a lower voltage, it should run cooler. My 4670K is not Devil's Canyon, but I get 36 degrees on idle after finding my CPU works fine at 4GHz with 1.09V. (Stock cooler, unfortunately.) However, my ambient temp is closer to 20C.

As for the RAM, that's just bad luck, but increasing your CPU input voltage to 2.1V has been known to increase the stability of a CPU clock, often allowing for higher stable RAM clocks.

NOT REGULAR VOLTAGE SETTING. DO NOT TRY TO PUMP 2.1V TO YOUR CPU'S REGULAR VOLTAGE SETTING. YOU WILL FRY YOUR CPU. GOOGLE IT FOR YOUR BOARD.

[dentedcapacitor]

@kuhcuk, you just said what I was afraid of. Yeah, I most likely would like to get the chip changed. I have a Malaysian chip which ironically was locally made. Should have trusted the Vietnamese ones.

 

@Abir Vandergriff Thanks for your prompt reply. I would undervolt it but like I said it was unusually high temps given my E6600 runs at 1.250V VCORE max. I did not want to run high volts on the processor because I was hoping this would last me for at least 5 years and anything to decrease the lifespan was a big no-no. Should have gotten a 4670K like you at since they are from more established Costa Rican fabs.

[Abir Vandergriff]

Thanks for your prompt reply. I would undervolt it but like I said it was unusually high temps given my E6600 runs at 1.250V VCORE max. I did not want to run high volts on the processor because I was hoping this would last me for at least 5 years and anything to decrease the lifespan was a big no-no. Should have gotten a 4670K like you at since they are from more established Costa Rican fabs.

Different chip, different temps. You have to understand that the input voltage isn't what wears on the chip, that's the VCore. Input voltage typically doesn't reduce lifespan much, if by any noticeable amount at all. What I'm saying is decrease Vcore to lowest stable, and increase input voltage to increase overall stability. Reducing Vcore should increase the chip lifespan, reducing temps potentially by quite a bit, if you can manage to go down far enough. Increasing your input voltage should only increase stability. The chip should last for a long time anyway, as most people upgrade their Intel chips just because they can rather than actually needing to. A lot of people still happily use i7-970 chips overclocked and overvolted.

[dentedcapacitor]

You make a good point Abir. I was thinking of doing that from the start but the rig was still a week old and I didn't want to mess about it too much until I was 100% confident that everything works and was just thrown off by how much heat that the Devil's Canyon generates even with aftermarket cooling. Its basically still in its burn in stage to see what drivers are needed and making sure everything works etc.before installing the final programs. I was concerned of selecting the wrong voltage field in the BIOS as I was more familiar with Gigabyte BIOS from my old board than the ROG one. The ROG BIOS has an initial CPU input voltage and an eventual one so that threw me off. Will give it a try later.