3070 OC benchmark

[Scott]

4.6Ghz I think is fairly typical of the 5600X without any tweaking at all, so I think on that front it's OK.

What are the peak temps on the GPU? I would force the GPU & CPU coolers to max and then re-run. That should get you the best case result with everything stock.

From there I would bump up the core clock of the GPU 100Mhz (Most can take that without issue) and bump up the power to 110% with the temps around 90C.

 

[DylBee]

4.6Ghz I think is fairly typical of the 5600X without any tweaking at all, so I think on that front it's OK.

What are the peak temps on the GPU? I would force the GPU & CPU coolers to max and then re-run. That should get you the best case result with everything stock.

From there I would bump up the core clock of the GPU 100Mhz (Most can take that without issue) and bump up the power to 110% with the temps around 90C.

I decided when I started to experiment with undervolting the GPU that I could handle the GPU fans running at full blast whilst gaming so that's all good. Peak temps are nice and low, can't remember at the moment but 40s/50s max. I've got to 2050mhz GPU clock @950mv so far without causing crashes

 

[SpyderTracks]

I decided when I started to experiment with undervolting the GPU that I could handle the GPU fans running at full blast whilst gaming so that's all good. Peak temps are nice and low, can't remember at the moment but 40s/50s max. I've got to 2050mhz GPU clock @950mv so far without causing crashes

Are you testing the undervolt with furmark? That's the best tool to highlight any GPU instability

 

[DylBee]

Are you testing the undervolt with furmark? That's the best tool to highlight any GPU instability

I hadn't been but will take this advice and start to use that also

 

[Scott]

You need to have a methodical approach to this so not sure if you are following a guide or etc. Some points:

Max fans is bad, you don't want to operate like this as you will massively decrease the lifespan of them. Max fans during testing is OK as you can get a constant baseline of cooling potential. It also allows for comparison when creating a fan curve.

Every time you change the OC you should do 2 things. Check results using 3Dmark to ensure it's had a positive influence on the score (you might be surprised) check with Furmark for stability.

During all testing you should always know what your max temps are, it shouldn't be a guess. I tend to either refer to the 3Dmark result for the benchmark or use HWMonitor throughout testing to note my peak.

In the link you posted the average GPU temp was 65C, so certainly not a peak of 50.

To be blunt, your approach seems to be very hap hazard and the responses to the queries when trying to assist have been vague at best. If you don't want any further help with the OC and just want to go about it yourself then that's OK, please let us know though as spending time trying to guide you down the correct path only to be glossed over is frustrating.

 

[DylBee]

You need to have a methodical approach to this so not sure if you are following a guide or etc. Some points:

Max fans is bad, you don't want to operate like this as you will massively decrease the lifespan of them. Max fans during testing is OK as you can get a constant baseline of cooling potential. It also allows for comparison when creating a fan curve.

Every time you change the OC you should do 2 things. Check results using 3Dmark to ensure it's had a positive influence on the score (you might be surprised) check with Furmark for stability.

During all testing you should always know what your max temps are, it shouldn't be a guess. I tend to either refer to the 3Dmark result for the benchmark or use HWMonitor throughout testing to note my peak.

In the link you posted the average GPU temp was 65C, so certainly not a peak of 50.

To be blunt, your approach seems to be very hap hazard and the responses to the queries when trying to assist have been vague at best. If you don't want any further help with the OC and just want to go about it yourself then that's OK, please let us know though as spending time trying to guide you down the correct path only to be glossed over is frustrating.

Scott, thanks for this - i didnt mean to be vague, I was replying to some comments when away from my system so didnt have access to the exact figures. I appreciate the frustration though when people are trying to help. I'm learning all the time about how to OC in general (from a variety of different sources/guides) so I appreciate the clear steps you've listed above, its helpful and I will follow then from now onwards. For info: Ive pasted below another 3DMark run.

I scored 12 802 in Time Spy

AMD Ryzen 5 5600X, NVIDIA GeForce RTX 3070 x 1, 16384 MB, 64-bit Windows 11}

www.3dmark.com

From Furmark- the Max GPU temp was 59 and from afterburner the max CPU temp was 56. As per the thread on 5600x OC I've also enabled PBO and manually set the Power limit to 90w. This changed the average clock frequency to 4542mhz (up from 4286mhz)

Max fans was inaccurate of me to say so I've also posted a picture of my GPU fan curve from afterburner, for reference.

 

[Scott]

Scott, thanks for this - i didnt mean to be vague, I was replying to some comments when away from my system so didnt have access to the exact figures. I appreciate the frustration though when people are trying to help. I'm learning all the time about how to OC in general (from a variety of different sources/guides) so I appreciate the clear steps you've listed above, its helpful and I will follow then from now onwards. For info: Ive pasted below another 3DMark run.

I scored 12 802 in Time Spy

AMD Ryzen 5 5600X, NVIDIA GeForce RTX 3070 x 1, 16384 MB, 64-bit Windows 11}

www.3dmark.com

From Furmark- the Max GPU temp was 59 and from afterburner the max CPU temp was 56. As per the thread on 5600x OC I've also enabled PBO and manually set the Power limit to 90w. This changed the average clock frequency to 4542mhz (up from 4286mhz)

Max fans was inaccurate of me to say so I've also posted a picture of my GPU fan curve from afterburner, for reference.

That's looking far healthier now. Right about where you should be for a mild OC on the GPU and PBO on the CPU.

I would adjust the fan curve though, that's far too extreme in my mind. I would aim for 100% fans around the 70c mark. 40-50 would be my steepest climb, aiming for around 60% fans by then, and rounding it over to the 100% over the next 20 degrees.

I would then run a before and after Timespy run to make sure you weren't taking too much of a hit on performance.

I would have a similar curve on the CPU, only shift across 10 degrees so that the the 100% is around 80C.