Shadow of the Tomb Raider benchmark

[jaybepr]

 

[skilzorz]

 

[Zyll Goliat]

Here is my result....what I found interesting is that this old R9 Fury having good result in 1080p but 1440p is even better considering the 4gb limitation and card age on 1440p I have 41FPS which is not that far from 1080p.....

 

[bissag]

1080Ti @ Stock 1440p @ Ultra

 

[Lew Zealand]

Almost dead thread needs moar love. Just like 92% GPU bound needs Moar Cores!

I present to you a modest upgrade from i5-8400 6c6t to i7-9700F (not K) 8c8t in an effort to get that pesky GPU-bound number to an overkill position. At 1080p. When I play at 1440p. I swear this makes sense *somewhere*...

Above upgrade gets from 92 to 99% GPU bound and fps from 102 to 109.

Woo.

Still way cheaper than a new GPU so I win.

 

[QuietBob]

3300X @ 4.5 all core, 6600XT @ stock + SAM:

Who said quad core gaming is dead?

 

[harm9963]

All CPU , 1080Ti vs 3080Ti.

 

[pyrotenax]

1920x1080 | 5900X | 3080

 

[AVATARAT]

1920x1080 | 5600x@PBO | 6700 XT

 

[zebra_hun]

Hi, original settings, i don't know SotTR. Just Benched for a friend. 10*5400MHz allcore, 5100MHz Uncore, 4266MHz RAM OC.

 

[Felix123BU]

and they said Windows 11 is worse for gaming

+11 FPS with the same hardware settings vs Windows 10

 

[masterdeejay]

Dx12 single gpu mode, dont know how to enable multi gpu mode.
If i enable Crossfire and fullscreen exclusive the pc reboots, but other games working fine on crossfire.
CPU is not for gaming (low freq 20 core xeon), and also the video card is not a gaming card but its playable.

 

[Lew Zealand]

Dx12 single gpu mode, dont know how to enable multi gpu mode.
If i enable Crossfire and fullscreen exclusive the pc reboots, but other games working fine on crossfire.
CPU is not for gaming (low freq 20 core xeon), and also the video card is not a gaming card but its playable.

Set it on High, or Med with some judicious High changes (or HUB's SotTR recommended settings) and you'll get much closer to 60fps minimum and a great gaming experience. On a GTX 1060, my Mins went up about 18% to close to 60FPS @1080p with the HUB settings.

 

[Ruru]

Everything on max and RT at medium

 

[phanbuey]

PSA: if you're comparing your scores, make sure to pay attention to GAME BUILD -- the DEMO does not get as high as the real game.... There are large variations across game versions as well.

 

[mrthanhnguyen]

Last run for AMD system. Im waiting for Alder Lake.

 

[professor dumb dumb]

Might want to hold onto that 5950x.
Tough to beat that 314. I've been working on it a bit - and still have a bit left i'd say but here's my current best:
Alder lake all p-core 5.1 e-cores disabled
cache / ring 49
DDR4-DR-4x8gb-3800mt/14,14,14,32,280-tight subs
Alltime is lowest settings, highest is as you'd expect.

I'll give it a go with 6900xt to see if there is a difference in the overhead (as cpu render is significantly faster with alder lake and there is copious optimizations in sotr for radeon) - but this is a reasonable result with geforce. As soon as I can test with ddr5 I'll update.

Edit:
Best average cpu render: 642
Best average cpu game: 312 (radeon)

 

[mrthanhnguyen]

Might want to hold onto that 5950x.
Tough to beat that 314. I've been working on it a bit - and still have a bit left i'd say but here's my current best:
Alder lake all p-core 5.1 e-cores disabled
cache / ring 49
DDR4-DR-4x8gb-3800mt/14,14,14,32,280-tight subs
Alltime is lowest settings, highest is as you'd expect.

I'll give it a go with 6900xt to see if there is a difference in the overhead (as cpu render is significantly faster with alder lake and there is copious optimizations in sotr for radeon) - but this is a reasonable result with geforce. As soon as I can test with ddr5 I'll update.

I saw from murmick with 4000c15 1t hit 354fps and carillo with 4300c14 1t hit 366. You should tweak more.

 

[professor dumb dumb]

I saw from murmick with 4000c15 1t hit 354fps and carillo with 4300c14 1t hit 366. You should tweak more.

Imc on this 12900k sample is a bit weak - SR does a little bit better - but it won't finish a run through at 4k. Have some 4000c14 coming which is a bin better than the 4400c16 I've been using so we'll see if there is anything left. Some people have had success with new bios rev's on the strix (not me unfortunately) - but I won't have access to another 12900k until january to see if this is board (dirty dimm socket) or cpu (weak imc) - leaning more towards a weak imc as I can run gear2 at 5000+.

Edit: also I think win11 is slowing things down a bit - just tried the 6900xt and it didn't go as well as I'd hoped.
Avg Cpu render: 583
Avg Cpu game: 312

 

[Cowboystrekk]

Imc on this 12900k sample is a bit weak - SR does a little bit better - but it won't finish a run through at 4k. Have some 4000c14 coming which is a bin better than the 4400c16 I've been using so we'll see if there is anything left. Some people have had success with new bios rev's on the strix (not me unfortunately) - but I won't have access to another 12900k until january to see if this is board (dirty dimm socket) or cpu (weak imc) - leaning more towards a weak imc as I can run gear2 at 5000+.

Edit: also I think win11 is slowing things down a bit - just tried the 6900xt and it didn't go as well as I'd hoped.

Are you on the 449 beta of SOTTR? Performance seems best on thet one. Mumriken used that for hus 350+ run as I adviced him to

 

[Teex]

CPU: Ryzen 7 5800x -> PBO + CO optimized per core
RAM: G.SKILL 32 GB KIT DDR4 3200 MHz CL16 Ripjaws V CL16-18-18-38 1,35 V OCed to -> 3733 MHz - 16-19-14-32-48 1,45 V
GPU: Sapphire Pulse 5700 XT OC + UV - > Core Clock - 2032 MHz Memory Clock - 1864 MHz CV 1111 mV

 

[Teex]

My Old PC:
CPU: Intel i7 - 8700
RAM: Kingston FURY Beast 2 x 8 GB KIT DDR4 2666 MHz CL16 + Corsair Vengeance LPX 2 x 8 GB KIT DDR4 2666 MHz CL16 CL16-18-18-38 1,20 V OCed to -> 2666 MHz - 13-16-16-35 1,35 V
GPU: NVIDIA GeForce GTX 1070 G1 Gaming 8G - OC - > Core Clock + 135 MHz Memory Clock - 335 MHz

 

[Lew Zealand]

A new GPU snuck into my rig last night - 6600 XT. Turns out it plays well with others:

1080p, Highest TAA.

Sapphire Pulse 6600 XT, 2750 Cores, 2250 Memory, +20% Power

Only 23% faster than my GTX 1080 at 1080p here, but 37% faster at 1440p. 8-core 9700f hits 80-90% CPU use in some areas, so a good match, that's set to 130W in Throttlestop instead of it's 65W default, so no CPU speed throttling.

GPU may not be paying attention to Undervolting as reported in GPU-Z (still at 1.15v when I have it set to 1.10v) but Superposition scores have inched up each time I make a change there so maybe it's just bad reporting? You need to do this UVing for max OC on these GPUs as it bangs against the power limit even at it's default 2664 MHz Core clock.

It's fun to have a new toy ehm, serious computation device to optimize. Some things I've noticed so far compared to my Pulse 5600 XT is (1) it plays much better with OC tools and (2) the trash Radeon OGL implementation, most notable to me in Mineraft, seems to work much better on the 6600 XT than on the 5600 XT.

 

[QuietBob]

Since I grabbed the freebie full version from EGS I re-tested to see how it compares with the demo benchmark:

As some people here have noted, the full version seems to use a newer, more optimized engine and yields higher results. The score is full 11% better in my case, with the same exact setup as in my demo run (stock 6600XT and oc'd 3300X).

What's more interesting, the higher score is entirely down to higher CPU fps: 20% higher avg render and 27% higher avg game It appears that the revised engine has better CPU threading/utilization.

 

[walker15130]

All the gpu power I could buy in 2021.

That's where you were wrong, buddy, here's some fresh new tests with my two month old 3060Ti.
Stock CPU:

CPU at 4400MHz didn't help much:

Even though I play at 1440p this good ol' ryzen 3600 tends to be the bottleneck more often than I'd like to admit. Too bad every upgrade path is steeply priced. Maybe 5800X3D will make people ditch their 5600/5800x