New 5600X build, baseline benchmark test

[i6pwr]

Finally have this build finished, want to get a few baseline results then explore the option of OC'ng if necessary, I know...never "if".

Specs are:

Gigabyte Aorus X570 Master
Ryzen 5600X
G-Skill Trident Z Neo 32GB (8x4) 3600
Samsung 980 Pro NVMe 4.0 1TB
Noctua D15 Chromax
Gigabyte Aorus 1080Ti Rev. 1.0 Specs are here, card is running stock. https://www.gigabyte.com/Graphics-Card/GV-N1080AORUS-8GD-rev-10-11#kf
Alienware 34" @ 3440x1440

Everything is running stock, including XMP profiles. I'm sure there might be some tweaks I should do for running stock but haven't done anything yet.

I will run some additional benchmarks but have only run R23 so far. Not sure how it compares, but needless to say it's much quicker than my previous 4790K which is still in use for one of the kids. Hopefully the 3060Ti wait list will shrink more in the coming months.

CPU idle temp is around 25 C and during these two tests, CPU was around 43 C and the GPU was 56-58C.

 

[Zerileous]

When you say including XMP profiles, I assume you mean that they are enabled. If not that's a big thing to do.

 

[i6pwr]

When you say including XMP profiles, I assume you mean that they are enabled. If not that's a big thing to do.

Correct. XMP is enabled.

 

[i6pwr]

I ran 3DMark and had some interesting results. After three tests on the trial mode, I thought I would load Aorus Engine and modify the fan curve. That worked for one run, which peaked around 88C, but the subsequent tests seem to peak the GPU temp at 60 C, regardless of settings. I have also tried Afterburner after uninstalling Aorus Engine. I've heard the Aorus Engine is buggy, but could be something I'm overlooking.

Either way, in stock form, this is 3DMark. Might purchase the advanced version if recommended.

Curious as to what to look for. Perhaps it possibly had something to do when I loaded up the Aorus Engine.

So before I changed anything, my first 3 results were within a few points of the previous score, once I tried changing the fans, even after setting a steep curve, they seem to now be maxed at about 60 C and the score dropped significantly which is expected.

 

[EarthDog]

If you raise the fan speed on the card and temps lower or stay the same, why is a significant drop in score expected? I would think the other way around, honestly.....

 

[i6pwr]

If you raise the fan speed on the card and temps lower or stay the same, why is a significant drop in score expected? I would think the other way around, honestly.....

Sorry, I worded that incorrectly.

When I changed the fan profile, the fans would not speed up any more than how they are at 40 C. I think the card is throttling down to keep from heating up more, not sure. On the first 3 passes, you could easily hear them spin up, now they won't any more than maybe half.

 

[i6pwr]

I actually ran 6 tests to confirm some type of uniformity.

I tried to increase the fan speed after the 6th test, I wanted it to ramp up sooner as the temps increased, that's when the results were about 1000 points lower for the subsequent tests.

Here's the first 6 tests, after this, the subsequent tests were about the same with a 1K point drop. I only have 2 screenshots of the tests after the fan adjustment, however for the 5-6 tests after messing with the fan profile, the results were the same. During the lower performing tests, I could not hear the fan ramp up near 100%. I see that the temp was around 60 C, but I could clearly see an average FPS drop around 10-15 FPS at times on the GPU tests. Any thoughts on why this could happen?

 

[EarthDog]

Not with the info we currently have. Have a log (msi afterburner) running and check clock speeds and ppwer limits is where I would start.

 

[i6pwr]

I will try to get more info. Just tried WOT, usually run around 110-130 FPS at this resolution, last game was around 70-85 FPS. Something is surely amiss.

 

[i6pwr]

I heard the Aorus software was temperamental, I re-checked the driver for the 1080Ti and I want to say it was 300.00...

I reloaded the current 384.76 driver and ran the test again. Seems about where it was before I started messing with it. Let me know what you want to see with Afterburner, I will try to get a log later today.


Recent test after reloading GPU driver.

 

[EarthDog]

The latest driver is 461.09.

I already listed the important things to see in MSI AB/GPUz. What are the clocks when that slow down happens? What does the power limit value say when the slow downs happen?

 

[i6pwr]

The latest driver is 461.09.

Thank you, I will look into that. I was going from the Gigabyte site.

I went ahead and purchased the software... Have some things to do but will re-visit later today or tomorrow.

I included two graphs, the one on the right was the test that was slowed down, looks to be about a 300 Mhz drop.
I also ran a test @1080P, included that as well. All appears to be working ok for now.

Curious about the test options in the red box, should I leave that alone or what is the consensus on settings? Forgive me as this is my first time with this software. Thanks again, I will get the Afterburner data later as well.

The top graph is the difference on GPU clock you asked for from the earlier tests.

 

[EarthDog]

I don't see any significant clock drops (though 3DMark isn't MSI AB/GPUz and less granular).... but, the score here is normal(?). Again, what we would want to see is when this thing DOESN"T score normally, what happens then............

As far as test options, leave it alone. You want to test the same way each time.

 

[i6pwr]

In regards to the GPU clocks, on the top image, the GPU clock on the left fluctuates between 1900-2000 Mhz, on the right it's a steady 1500 ish Mhz, with no deviations or throttling.

 

[EarthDog]

Let us know when you have the GPUz/MSI AB chart posted and the benchmark pops low.

 

[i6pwr]

I will, thanks. I understand it's hard to make a call on what's going on without the logs.

There is a current GPU driver, 464.40, I loaded that, all is working well. I did run another test with the AB, score increased slightly for Time Spy @ 3440x1440 to 9677.

For the record, in case you see anything out of whack, I attached the log graph.

MSI log stats were:

Max Mem usage: 2016

Max Core Clock: 1949

Max Mem clock: 5508

Max Power: 103


I appreciate the help, will let you know if it drops way off again.