Unigine Valley benchmark score - massive drop

[Slinky69]

I installed my new system on Weds:

3800X, X570 Tomahawk, 16GB 3600 C14 memory. RTX 2080.

I had taken a Unigine Valley benchmark on my old system (i5-6600k, still had the RTX 2080) before I stripped the bits out, and it was: 83.3 fps, score 3484 at 1440p.

One of the first things I did once Windows was installed on my new box was to run the same benchmark again, and was v happy to see 180.7 fps, score 7560 - Making me feel like my upgrade was a worthy one! As far as I remember, at that point I had the memory set to XMP only, and only the M2 drive was plugged in, not my 3 SATA drives.

Today, I did the following:
1) Upgraded the BIOS from 1.0 to 1.2
2) Tried changing some memory settings which resulted in better times in DRAM Calculator test

So I tried the Valley benchmark again to see if the tighter memory timings had helped. Awful score - Roughly the same as my old system!

So - I reverted to the XMP settings, no change. Other things I have tried with no luck:
Reverted to BIOS 1.0, a fresh scan in MSI Afterburner, unplugged the SATA drives,unplugged HDMI cable for VR headset.

So I am going slightly mad.

Any suggestions as to things to check, please? I am baffled as to how I am now getting so much worse a benchmark than I did on the same system 3 days ago

 

[sgoaty]

Maybe a daft question but have your valley settings stayed the same?

 

[Slinky69]

Maybe a daft question but have your valley settings stayed the same?

A perfectly valid question, for sure!

The short answer is that I can't be certain. The current score actually looks sort of correct, looking at other scores online, it's at the low end for a 2080, but not outrageously so.

And running Unigine Superpostion, which has a lot more recent scores to compare, my score is at the top end of a 2070s, low end 2080.

So, the short answer is that I probably messed up when I got that uber score, then spent yesterday going mad trying to recreate it

 

[Benass]

do you have a screen shot of thet uber score? is there a way to compare it to your new scores?

 

[WantoN]

A perfectly valid question, for sure!

The short answer is that I can't be certain. The current score actually looks sort of correct, looking at other scores online, it's at the low end for a 2080, but not outrageously so.

And running Unigine Superpostion, which has a lot more recent scores to compare, my score is at the top end of a 2070s, low end 2080.

So, the short answer is that I probably messed up when I got that uber score, then spent yesterday going mad trying to recreate it

I'd be inclined to ignore that stuff unless you're pushing for the best overclock.

Have you got any FPS numbers from favourite games you can compare?

 

[Slinky69]

do you have a screen shot of thet uber score? is there a way to compare it to your new scores?

Sadly no, I just wrote it down

 

[Slinky69]

I'd be inclined to ignore that stuff unless you're pushing for the best overclock.

Have you got any FPS numbers from favourite games you can compare?

I guess the thing is that, really, at 1440p I shouldn't expect much FPS change from a new CPU, which is why I was so blown away by that initial result. I am putting it down to me having done it wrong somehow and moving on, otherwise I could drive myself even madder

 

[WantoN]

I guess the thing is that, really, at 1440p I shouldn't expect much FPS change from a new CPU, which is why I was so blown away by that initial result. I am putting it down to me having done it wrong somehow and moving on, otherwise I could drive myself even madder

Benchmarking is a dangerous game mate! That's what I'd do. I tend to go for quite aggressive OCs, especially on my GPUs, but I avoid benchmarks for this very reason.

As ou eluded to, these days upgrading your CPU is more about multi-tasking, productivity and removing GPU bottlenecks, especially on higher resolutions.

Lovely spec mate, enjoy!

 

[Rom]

Have you ran it again and confirmed the low score?

Yesterday, I had a senior moment. I was testing my 3080, messing with different settings. I was in Unigine Heaven 4.0, used it for years and doesnt take forever (probably not great now I guess, i was just looking for gains and stability)

So I started getting worse and worse scores, even though I ran the same clock as a known good score etc. Turns out I had about 8 instances open, after each run, I didnt exit properly. Was just hitting go and doing other stuff (off pc). Surprise surprise, when I closed them all, my results were repeatable. I never mess with settings, as im not looking for best score, just something to measure loss or gain against.

On a side but related note.
I have a 6700k @4.5, did have a 1080ti FTW3. Now a 3080 FTW3 Ultra.
I am hugely cpu bottlenecked now! Like some games it will sit at 40% load. My fps has increased, but not as much as I hoped.

I would guess a 2080 is similar performance to a 1080ti? And your 6600k is even slower than my 6700k. So I would expect to see fps gains on the new cpu.

Im waiting for the 5900x, and im expecting gains for sure!

Edit...

I just ran Valley @ 1440p

6700k @ 4.5
32gb Corsair 3200
3080 Ultra

Score : 6941
FPS : 165.0
Min : 31.8
Max : 267.4