Shadow of the Tomb Raider benchmark

[Melvis]

SLI? I have AMD cards

Haha indeed you do! My last crossfire set up was 280X's, for the most part worked great but a string of 11updates broke it in Trine 2 (and maybe some others) and never got fixed I gave up and moved back to SLi. How are you finding your crossfire in todays games? I have had alot of crossfire cards over the yrs, starting back with 2x4870X2's

 

[Ruru]

Haha indeed you do! My last crossfire set up was 280X's, for the most part worked great but a string of 11updates broke it in Trine 2 (and maybe some others) and never got fixed I gave up and moved back to SLi. How are you finding your crossfire in todays games? I have had alot of crossfire cards over the yrs, starting back with 2x4870X2's

At least BF1 got the benefit of two cards and IT KICKS ASS!

I have a perverted idea of getting a 295X2 with these.

Fucking Quadfire.

 

[Melvis]

At least BF1 got the benefit of two cards and IT KICKS ASS!

I have a perverted idea of getting a 295X2 with these.

Fucking Quadfire.

Nice! I can only imagine! as I dont have the game haha, but I know from the past when it works it works very well.

O M G that is perverted but I like it! lol Do tell if you end up doing it, i want pics!

My Crossfire 4870X2 days, omg such power, such performance, such NOISE! lol still one of my all time favorite set ups.

 

[Athlonite]

well here we go actually ran better than I thought it would

and with a couple of things changed

 

[Ruru]

Nice! I can only imagine! as I dont have the game haha, but I know from the past when it works it works very well.

O M G that is perverted but I like it! lol Do tell if you end up doing it, i want pics!

There will definitely be pics if I end up with that, Tri-Fire at least

My Crossfire 4870X2 days, omg such power, such performance, such NOISE! lol still one of my all time favorite set ups.

Heh, exactly the same when I had a HD 3870 X2 10 years ago.. I'd had that in Tri-Fire with my old HD 3850 but I had a NForce motherboard so no Crossfire

 

[Melvis]

There will definitely be pics if I end up with that, Tri-Fire at least


Heh, exactly the same when I had a HD 3870 X2 10 years ago.. I'd had that in Tri-Fire with my old HD 3850 but I had a NForce motherboard so no Crossfire

Nice! I will be looking forward to it if you do

oh wow ok even older then my set up, nice! and yeah thats a bummer, the good old days haha

 

[augustben]

Finally saved up enough money to get myself a new PC. The game runs smoothly and amazingly! I had to disable DirectX 12 though since the game won’t launch with it enabled.

Edit: I managed to get SOTR to work with DirectX 12 again, and I got better frame rates than before!

 

[Pumper]

So is up with that GPU Bound stat? I'm running 4770K at 4.2GHz with 32GB 1600 DDR3 oc'ed to 1866. Upgraded my 970 to 1070Ti today, but the GPU Bound stat makes no sense - was 5% with 970 (46 avg. FPS on Ultra 1200p) and 23% with 1070Ti (67 avg. FPS).

So if lower % means bigger CPU bottleneck, shouldn't it we reverse?

 

[lsevald]

So is up with that GPU Bound stat? I'm running 4770K at 4.2GHz with 32GB 1600 DDR3 oc'ed to 1866. Upgraded my 970 to 1070Ti today, but the GPU Bound stat makes no sense - was 5% with 970 (46 avg. FPS on Ultra 1200p) and 23% with 1070Ti (67 avg. FPS).

So if lower % means bigger CPU bottleneck, shouldn't it we reverse?

I went from a 4770k@4.4GHz + 1080ti=116fps (43% gpu bound) to a 9900k@5.1GHz + 1080ti =143fps (87% gpu bound):

So I guess it means that, with a faster CPU, my GPU is more of a limiting factor than before. Maybe think of it as a scale, 50% would mean the game engine is waiting equally as much time for the CPU as the GPU to finish work. So in your case, yes, a faster GPU should mean less gpu bound? Are you sure you run exactly the same settings between the two, as that number will change depending on high or low quality settings and res, if I run it at 1280x720 and low settings I get like 1-2% GPU bound. Confusing stuff, don't listen to me

 

[Pumper]

Looks like DX11 is the issue here (I'm still on Win7), so the CPU usage gets murdered in populated areas (drops to 40-50% usage) and takes the GPU along for the ride.

 

[EarthDog]

Looks like DX11 is the issue here (I'm still on Win7), so the CPU usage gets murdered in populated areas (drops to 40-50% usage) and takes the GPU along for the ride.

Doesn't DX12 use more CPU? Also, look above you.. DX12 shows the same drop in the same area, no?

 

[Pumper]

Doesn't DX12 use more CPU? Also, look above you.. DX12 shows the same drop in the same area, no?

It does, but in a good way as it does not bottleneck the GPU, unlike in DX11.

Here's a great DX11vsDX12 video with CPU/GPU usage shown: https://youtu.be/G9YNqbmIYYI

 

[EarthDog]

It does, but in a good way as it does not bottleneck the GPU, unlike in DX11.

I don't understand why it is any different in bottlenecking the GPU...when it uses more of the CPU in the first place. What speed is your processor at? Stock? Maybe create a system spec list so we can see what hardware you have going....

 

[anachron]

Here are my results with and without dx12 in 1080p highest settings. For reference, i have gsync and vsync on in nvidia control panel as i forgot to remove it. I did hit 144fps but only during a very short time.
1080p dx12 :


1080p dx11 :


It seems that dx12 does indeed use a lot more cpu (43% cpu bound in dx12 vs 27% in dx11), but the performances are still a lot better than in dx11. Whoops, bad reading skills.

My results at 1440p / Rtx medium / no DLSS :

 

[EarthDog]

It seems that dx12 does indeed use a lot more cpu (43% cpu bound in dx12 vs 27% in dx11), but the performances are still a lot better than in dx11.

That is GPU bound value, not CPU.

 

[anachron]

That is GPU bound value, not CPU.

Yep sorry, was editing while you replied. Guess i didn't get enough sleep last night

 

[Pumper]

I don't understand why it is any different in bottlenecking the GPU...when it uses more of the CPU in the first place. What speed is your processor at? Stock? Maybe create a system spec list so we can see what hardware you have going....

It's a software issue, not hardware (as posted in my first reply, I'm running 4770k at 4.2 with RAM at 1866), just like anachron showed in his DX11vDX12 tests. DX11 kills the CPU render performance for some reason and drags down GPU along with it, resulting it lower overall CPU/GPU usage % and thus lower FPS.

 

[EarthDog]

It doesn't do that with my system... that doesn't make sense with how its supposed to work, AFAIK.

 

[mtrai]

Here are mine so far on my Ref PowerColor 5700 XT posting since I am not seeing any. 1920 x 1080 highest settings. Redoing my 2560 x 1440 as I ran it at high not highest.

 

[Taraquin]

Hmm, I'm getting really low CPU-score with my i5 8400. Weird.

 

[anachron]

It would be interesting to know at which frequency your 8400 run during the bench and what memory you have. I got twice the performance with my i5 8600k @ 4.7ghz.

 

[phanbuey]

Hmm, I'm getting really low CPU-score with my i5 8400. Weird. View attachment 129105

Put your power settings to HIGH PERFORMANCE and try again. You're probably downclocking.

 

[Taraquin]

Hmm, it was my ram timings. Been running 2666@13-16-16-32\CR1\tRFC 350 for a while now, but seemed tovremember that I used to get higher scotes when I ran 14-16-16-32. Changed it and things made more sense

I change power plan after that and got 20% better with that thx for advice.

Actually, turns out it was not the ram, but a Throttlestop-setting called speedshift, checked in my BIOS and speedstep is enabled by default there. disabled speedshift in Throttlestop, selected high performance Power plan again with the odd ram-timings (13-16-16-32) and CPU render average went up to 176 so a slight improvement from CL14 Funny thing is that this setting has not interfered With any of my other games or benchmarks (Cinebench, Unigine heaven, For Honor etc), only SOTTR which seem to drop CPU-speed quite significantly.

 

[Taraquin]

i7 6700hq 4c/8h 3.1 all core turbo is awfully slow compared to my i5 8400. 87 vs 176 on CPU render avg.

 

[EarthDog]

My dude..... print screen that... screen caps are sooooooo 1990s.....

I'm also pretty sure that CPU render is influenced some by the GPU.