Quickie Rise of the Tomb Raider DX12 testing on Ryzen vs Kaby Lake

Hardware
1

7700k @4.5ghz all cores vs 1800x @4.1ghz all cores

DX12 with the Asus strix fury

68.1 min fps on 1800x vs 81.1 on 7700k
On spine of the mountain

14910748384592072302993.jpg2688x1512 609 KB

IMAG0044.jpg2688x1512 887 KB

So there was some talk of dx12 on Nvidia being weird. This may be but the 7700k seems slightly faster in testing still.

14910749463281920817160.jpg2688x1512 911 KB

This is unofficial I've more testing to do but thought I'd post.

Test Rigs:

i7-7700k @ 4.5ghz all cores
ASUS Strix Z270
GSkill Trident Z 14-14-14-34 DDR4 @ 3200mhz
ASUS Strix Radeon Fury @ ASUS defaults
OCZ RD400 NVMe

Ryzen 7 @ 4.1ghz All Cores
ASRock Taichi X370
GSkill Trident Z 14-14-14-34 DDR4 @ 3200mhz
OCZ RD400 NVMe
ASUS Strix Radeon Fury @ ASUS defaults

Ryzen DX 12 +Fury (I did DX11 also can show those if anyone cares)

GeoThermal Valley:
Min FPS: 68.5
Max FPS: 109.3
Average FPS: 86.8

Prophets Tomb:
Min FPS: 26.3
Max FPS: 128.5
Average FPS: 90.6

SpineOfTheMountains
Min FPS: 68.1
Max FPS: 189.0
Average FPS: 121.3

7700k DX 12 + Fury (same w/dx11)

GeoThermal Valley:
Min FPS: 69.7
Max FPS: 114.7
Average FPS: 88.5

Prophets Tomb:
Min FPS: 67.7
Max FPS: 105.1
Average FPS: 92.8

SpineOfTheMountain:
Min FPS: 81.1
Max FPS: 195.5
Average FPS: 126.1

Can anyone think of anything useful/worthwhile to do? I'm going to take a close look at the frame times but at first blush it doesn't look too bad. This is a canned benchmark too so I stared at it intently while benchmarking. When the i7 hitches, it hitches bad but they both had significant frame dips. Subjectively I can't say that either experience was bad at any point, though.

I am also testing same scenarios with the 1080 but same story prettymuch.

10 Likes
2 

#!/usr/bin/GOALKEEPERthoughts
----------

"

The Ryzen results make me happy on the inside.

"
2 Likes
3

Do you have the option to test crossfire DX12? According to AdoredTV the difference between the two CPUs narrows.

4

can you measure temps against ryzn + nv and intel + nv?

1 Like
5

temps of what?

1 Like
6

I only have the one fury I got from microcenter. hahaha sadface.

We gotta grow our subs and community and whatnot so I can widen the pool of hardware. :smiley:

Might buy some fury Xs on super sale.... hahahaha thats silly with vega right around the corner hmm

2 Likes
7

DeepCool Steam Castle

Oh god, I forgot that case existed.

8

Aww :frowning:

9

Well that dip in min fps on the 1800X is interesting i think.
On Average its looks good, but we still see the better ipc on Kabylake manage to keep the lower frame rates better i think.
Atleast at that particular part of the game.
That dip to 26.3fps is pretty low there i think, not sure what was going on there.

10

both cpu and gpu.

(fury-x's won't give you more performance, i'd consider unlocking your asus strix instead) It has more overclocking potential.

11

They said Ryzen with amd gpu and DX12 gave the biggest improvement at during the Soviet Installation during game play.

12

Older case but...if you look at FPS, it doesn't move, no matter how much it hitches. It should be zero occasionally as bad as it hitches.

Same gfx card model, zero hitching with AMD FX-8350. Lower FPS tho.

Seems like if you have Intel, stutter and hitching does not affect FPS at all. Which is BS.

13

Hi Wendell,

Could you please try running the fury and 1080 at PCIe x8 instead of x16 on the Ryzen?

I am interested to see if the mechanism that switches the x8 PCIe lanes from the 1st slot to the 2nd slot is negatively impacting performance if slot 1 is running at x16

14

These #s were done at pcie x8 with an RX480 installed in the secondary slot. There wasn't much difference that I could tell with the fury at x8 vs x16 on Ryzen. I haven't tested the 1080 at x8 yet though, unless someone got to it yesterday and I haven't found the files yet

1 Like
15

Thanks for the update. Was the Intel at x8 or x16?

Seeing if the Nvidia does something weird or not should be interesting.

16

I did intel at x8 as well with the fury just for giggles haha :smiley:

4 Likes
17

Should benchmarks be posted on 4/1 :slight_smile:

1 Like
18

It would be interesting to see if the gains (or losses) from going from DX11 to DX12, for any (or each) CPU, are similar with the 1080 and the Fury. I think that would allow you to confirm whether the GPU matters for such comparison, vs. everything being a scaled up or down version of the same results. I understand that re-running things switching APIs, GPUs, etc, can be very time-consuming, though.

19 

#!/usr/bin/GOALKEEPERthoughts
----------

"

Did anyone else notice lake is spelled wrong?

I am tempted to correct it.

"
20

Could you try running the benches with like the 1050ti or something because in this case I think that in some cases the cpu bottlenecks while sometines the cpu.
Directx12 was also addressed for low end systems.
Also in Geothermal valley, more cores gave high max fps but slow cores gave low minimum fps. Ideally no-one is going to run fury at 1080p