Shadow of the Tomb Raider benchmark

[AMF]

What happened ? reran it twice first was 67 average fps 4k now this run is this.. ultra everything only thing disabled is motion blur

 

[QuietBob]

What happened ? reran it twice first was 67 average fps now this run is this.. ultra everything only thing disabled is motion blurView attachment 316393View attachment 316394View attachment 316397View attachment 316402

Do you mean the same average fps in 1080p and 1440p? That's a clear sign of a CPU/RAM bottleneck. 0% GPU bound means being entirely limited by the processor (and possibly the memory).

Look at the CPU Game and GPU stats. With a newer processor, you would see the average fps closer to the number in the last column at 1080/1440p. Zen 2 simply cannot keep up with the 6950XT.

 

[AMF]

Do you mean the same average fps in 1080p and 1440p? That's a clear sign of a CPU/RAM bottleneck. 0% GPU bound means being entirely limited by the processor (and possibly the memory).

Look at the CPU Game and GPU stats. With a newer processor, you would see the average fps closer to the number in the last column at 1080/1440p. Zen 2 simply cannot keep up with the 6950XT.

no i ran 4k bench earlier and was max 67 fps average same settings now im getting 122 fps avg 4k same settings... actually i did change some sub timings on memory would it make that big of a difference?

 

[radosuaf]

It should be called 1T 2T in the bios. He can also drop TRFC to 520

OK, I did some research on the memory and11th gen. As far as I understand, there's no point of going over 3600 in case of Rocket Lake, because over that Gear 2 kicks in and it's not worth it. So basically I'm limited to 3600 MHz CL16. I could swap current 4x8 3200 CL16 for 2x8 3600 CL16 at no cost and honestly 16GB should suffice for some time still, in case I need more, I'll throw in 2 additional sticks. Is it worth the hassle (of selling old and buying new, in general I should be able to sell the old 32 GB at the cost of new 16 GB). I'm swapping the CPU for 11700 non-K (got a good deal on used) as well.

 

[AMF]

Do you mean the same average fps in 1080p and 1440p? That's a clear sign of a CPU/RAM bottleneck. 0% GPU bound means being entirely limited by the processor (and possibly the memory).

Look at the CPU Game and GPU stats. With a newer processor, you would see the average fps closer to the number in the last column at 1080/1440p. Zen 2 simply cannot keep up with the 6950XT.

well it certainly appears to be pretty good this 3900x is screaming no doubt 5900x would be better though

 

[QuietBob]

well it certainly appears to be pretty good this 3900x is screaming no doubt 5900x would be better though

Sure, the 3900X is still a capable gaming CPU. It's just that newer games and high-end GPUs require more ST power to achieve optimal results. Zen 2 may not be a bottleneck in many other games, and you can lessen the burden on the processor by playing in higher resolutions with more graphical detail.

The 6950XT can do 4K maxed out at 60+ fps in most titles, so that's what you'll likely be playing anyway

 

[fevgatos]

OK, I did some research on the memory and11th gen. As far as I understand, there's no point of going over 3600 in case of Rocket Lake, because over that Gear 2 kicks in and it's not worth it. So basically I'm limited to 3600 MHz CL16. I could swap current 4x8 3200 CL16 for 2x8 3600 CL16 at no cost and honestly 16GB should suffice for some time still, in case I need more, I'll throw in 2 additional sticks. Is it worth the hassle (of selling old and buying new, in general I should be able to sell the old 32 GB at the cost of new 16 GB). I'm swapping the CPU for 11700 non-K (got a good deal on used) as well.

Some rocket lake can actually get up to 4000 in gear 1. My 11600k couldn't, I was running 3466c12

 

[radosuaf]

I ordered Corsair 2x8 GB 3600 CL16 - does not work on Gear 1. Threw 11700 in, benchmarked SotTR... 131 FPS up from 122 FPS, GPU Bound... 25%.

I ran 3DMark - after overclocking the VRAM I got the best result in the World for 11700 & 6700, 4th in the World if we add 11700K & 11700KF as well. So it seems everything is set probably. What's wrong with SotTR then?

 

[AMF]

Sure, the 3900X is still a capable gaming CPU. It's just that newer games and high-end GPUs require more ST power to achieve optimal results. Zen 2 may not be a bottleneck in many other games, and you can lessen the burden on the processor by playing in higher resolutions with more graphical detail.

The 6950XT can do 4K maxed out at 60+ fps in most titles, so that's what you'll likely be playing anyway

solved my problem the 67 fps avg in 4 k i got yesterday was with ultra raytracing shadows enabled 122fps without raytrace enabled

 

[QuietBob]

What's wrong with SotTR then?

Judging from your result with an i5-11400F, your CPU was underperforming. Could you post your current scores?
Also, please fill out your hardware profile here on TPU.

 

[radosuaf]

There you go:

 

[QuietBob]

There you go:
View attachment 316485

CPU Game still seems too low. Are you running your RAM in Gear 1? Did you max out PL1 and PL2 in the BIOS? Is the CPU boosting properly?

 

[radosuaf]

Yes, it's Gear 1. Strange thing is my other benchmarks look OK - Cinebench 2024 is 740 for example, which is what you would expect from 11700...

 

[freeagent]

1080p low:

And 4K Ultra with SMAA:

5900X 1080p low:

5900X 4K max:

 

[Teex]

6700 should work much better with 11700 or 11400, just for comparison

RX 6700 + 5800x performance in stock Bios, stock RAM 2133 Mhz, stock GPU:

RX 6700 + 5800x with maximum tune, PBO +100, 3800 CL14R RAM and heavy GPUOC:

 

[radosuaf]

View attachment 316799

So you have 139 FPS and I have 133 FPS here (also stock GPU) - why are you then 70% GPU Bound and I am 25%?

 

[Teex]

You are not using full potential of your CPU.
Its hard to say how this game calculates GPU bottleneck, but as you can see, you are leaving much performance on the table.
I even tested this RX 6700 with older I7 - 8700 and I could achieve 152 FPS with 2666 Mhz CL14 RAM.

 

[radosuaf]

No idea what happens here... All the other benchmarks I run (3DMark, Cinebench) seem to be fine, I see that is on a different game version, can it be a reason?

Do you have other results for the 6700? Cyberpunk, RDR2, GTA V for example?

 

[Teex]

Try to post your 3DMark result:
https://www.3dmark.com/spy/41755454 - this is for my old 8700

 

[radosuaf]

https://www.3dmark.com/spy/42141953 - GPU OC
https://www.3dmark.com/spy/42141736 - GPU Stock

CPU score is 11628 (my 11700) vs. 7835 (your 8700)...

 

[Teex]

The GPU score is ok 11700 vs my heavy OCed 12350 that is normal and to be expected.

But you could squeeze 1000 points more from 11700, your score is average in comparison to other benchmarks with 11700.

Have you removed all powerlimits in BIOS? Only thing I can think off is, that your CPU can´t pull all the power it needs.
You have TUF B560, this motherboard is capable to deliver all the power 11700 needs.
So try to search on youtube for some guide

 

[radosuaf]

Another one... 800x600 Lowest - 34% GPU Bound?

2560x1440 Highest:

 

[natr0n]

1080p/dx12/highest

 

[mouacyk]

Fine-tuned 6400C28 FCLK=2200 VDD=1.5 5150MHz CPU:

 

[Gica]

I'm wasting my time with other comparisons.
Game set to Highest preset, RT Ultra, DLSS OFF (activation not required).

7800X3D mode = <80W = maximum consumption recorded by 7800X3D in multi-core applications (Cine R23, for example). In this way, with the same consumption, the 14700KF destroys the 7800X3D in the single/multi core benchmarks.

Default: you know, I think. PL1/2: 253W

By default, I use a middle path, named 7900X mode (<104W), which provides enough power for almost all applications, except rendering and encoding.
So, default versus 7800X3D mode, results and impact on processor consumption.

Note: 7900X, 7800X3D and 14700KF have about the same price, with small variations depending on the country. With us, 14700KF is the cheapest.
In the video, you must follow what is seen in the capture, especially in the third part of the benchmark.

Default

7800X3D Mode

.......................................................................

Same settings, almost maximum, CPU without HT and Turbo (20c/20t x 3.4GHz).
The same result obtained by the video card.
Obviously, it is not a viable option because it involves intervention in the BIOS, but it is a landmark about the potential of the processor in gaming.
In addition, at the same consumption with 7800X3D, the differences are smaller in Multi Core (only 20.5k in Cinebench 23 compared to 25K obtained with the <80W preset) and in single core you return in 2020, to the 10th series and ryzen 3000.
I repeat, it's just one of the tests to see if the processor, at 50% of its potential, can get everything out of the 3070Ti.