Unusually High Temperatures 4790k + NH-D14 (Need Help)

[TroubleKlef]

So I just completed my PC build today and I'm in the process of updating drivers, bios, and installing programs.

 

One of the programs I installed was Aida64 as I wanted to see what temperatures I would be getting with my setup.

 

I ran only the FPU test and here were the stats.

 

CORE1: Min 31 Max 91

CORE2: Min 31 Max 93

CORE3: Min 28 Max 90

CORE4: Min 29 Max 84

 

My CPU is completely stock although I did set my XMP profile to #1 which should have only clocked my memory to 1600mhz.

 

I do have a pretty small case (Bitfenix Phenom M) with 2 Noctua NF-F12s from the top intaking air and a Noctua NF-A14 for an exhaust. These temperatures still seem extremely high and I'm wondering if I need to reseat the cooler and try the dot method instead of the line for thermal paste.

[MegaDave91]

Likely that the air isn't escaping quick enough.

[aidenrelkoff]

Likely that the air isn't escaping quick enough.

that, and maybe too little thermal paste?

[Cheese4395]

Is the cooler seated properly? 

[Cheese4395]

Likely that the air isn't escaping quick enough.

 

If u want to test this feel your case while running the test. If it's hot then this is true. If it's not then the cooler isn't seated properly 

[TroubleKlef]

Linus's "ultimate sli build" had 2 780tis had only 1 120mm fan from the bottom intaking air and he is getting temperatures of 90c with Prime95.

 

Shouldn't I be getting better temps since I have no graphics card along with 2 NF-F12s and the D-14 being superior to the BeQuiet?

 

I didn't benchmark long enough for the case to get hot (only about 5 min) as 90c was a bit overwhelming at stock speeds. I did accidentally put a bit more thermal compound than I would have since the compound released faster than expected.

 

Also I lifted up the CPU cooler after it touched a little bit of the compound so that may have introduced air pockets?

[MegaDave91]

I didn't benchmark long enough for the case to get hot (only about 5 min) as 90c was a bit overwhelming at stock speeds. I did accidentally put a bit more thermal compound than I would have since the compound released faster than expected.

 

Also I lifted up the CPU cooler after it touched a little bit of the compound so that may have introduced air pockets?

Too much thermal paste can do the exact opposite of its intended purpose, so there's a clue. Try reapplying it and use only the amount the size of a grain of rice. I know Linus says uncooked grain of rice, but I really think that's too little. 

[TroubleKlef]

Too much thermal paste can do the exact opposite of its intended purpose, so there's a clue. Try reapplying it and use only the amount the size of a grain of rice. I know Linus says uncooked grain of rice, but I really think that's too little. 

Alright I will reseat the heatsink and this time will try the dot method of a grain of rice.

 

What is the best way to clean the CPU? 90% alcohol and a coffee filter?

[MegaDave91]

Alright I will reseat the heatsink and this time will try the dot method of a grain of rice.

 

What is the best way to clean the CPU? 90% alcohol and a coffee filter?

 

90% is great and a coffee filter will work. Toilet paper is another great one. 

[rmac52]

I'd reseat the cooler.  You should be getting better temps than that.

[TroubleKlef]

Alright so I reseated the cooler and tried the dot method and these are my new temps.

 

CORE1 min 33 max 90

CORE2 min 35 max 92

CORE3 min30 max 89

CORE4 min32 max 84 

 

The results are near identical. Do I just have a really bad chip?

[TroubleKlef]

So I redid the Aida64 stress test but this time I did the CPU stress rather than the FPU stress and these were my temperatures

 

CORE1 max 58

CORE2 max 61

CORE3 max 57

CORE4 max 53

 

Each core under 100% load. So which stress am I supposed to base my overclocks on? CPU stress? or the FPU stress?

[TroubleKlef]

So I did a Prime95 blend test while I was asleep and my CPU hit 100c! Yikes! Will my CPU be damaged? The CPU would throttle itself or shut down to prevent damage, right?

 

It seems to me that floating point stress testers really put some stress on my CPU. When overclocking, should small FFTs on P95 and FPU Stress on Aida64 even be used? 

[TroubleKlef]

I tried OCCT today and I got some fairly good temperatures.

 

Each core was around 68c.

 

By the way my CPU is turbo'ing to 4.4ghz at like 1.162v. Are these good temperature and will I have headroom to overclock?

[TroubleKlef]

Could anyone clarify which programs/tests I should base my overclocking on?

 

Using Prime95 small FFT or Aida64 FPU stress tests my CPU will hit 100c.

 

Using Aida64 CPU stress or OCCT my CPU will hit ~70c.

 

So if I base my overclock on floating point tests I have no room for overclock but with regular stress tests I have a little bit of wiggle room.

[TroubleKlef]

I don't mean to beat a dead horse but I can't find this information elsewhere

[foamygopher]

Prime 95 and Aida 64 both stress AVX. Wich would explain the massive temps. With occt AVX isn't enabled by default. You should be ok if your temps in occt look good.

[foamygopher]

Also haswell + prime 95 = bad. Don't use prime 95.