Shadow of the Tomb Raider benchmark

[Det0x]

5950x vs "CPU Game"

Used this benchmark as a purely cpu+memory test, hence the "medium" settings, but it dont matter for CPU game score which i was after
3090 is running stock bios and everyday clocks.

 

[mouacyk]

Think this is the fastest AMD "CPU Game" numbers in this thread (?)

Used this benchmark as a purely cpu+memory test, hence the "medium" settings, but it dont matter for CPU game score which i was after
3090 is running stock bios and everyday clocks.

Would be more meaningful comparing apples to apples.

 

[Lew Zealand]

Awright in honor of Nvidia re-releasing the GTX 1050 Ti so you only need to pay $150 over market value for a scalped GPU, I give you this barn burner:

Intel NUC8i5 Core i5-8259U @3.6GHz ACT 4c8t, 2x8GB 2400 RAM, eGPU w/ASUS Phoenix 1050 Ti (75W slot power only) OC +165cores, +200memory

Obviously useless at these settings, so I tried at Low (not Lowest- no shadows there and looks awful) settings, which look... OK:

That's probably playable...ish though 40fps lows are a little rough.

I'll pop it in the 8400 to see if full PCI bandwidth and latency will help for more normal folks out there.

 

[QuietBob]

Just a little update. I've noticed that AA wasn't enabled for my previous test, so here's the same config with TAA:



The HD7970 is soundly beating the 1050Ti, with the latter starting at EUR 180 / USD 220 new here

 

[Det0x]

New CPU+memory highscore
261 fps average "CPU game"

 

[mrthanhnguyen]

so this thread comes back to everyone custom setting again.

 

[Det0x]

LOL i was benching the wrong version of the game.. Free version on steam was v505 (patch 15) which was missing 4 performance updates
Version 298 = patch 19

This is more like it :innocent:
CPU Game = Average 282 fps

 

[thesmokingman]

All these custom settings are pointless.

 

[Taraquin]

LOL i was benching the wrong version of the game.. Free version on steam was v505 (patch 15) which was missing 4 performance updates
Version 298 = patch 19

This is more like it :innocent:
CPU Game = Average 282 fps
View attachment 189970
View attachment 189971

Could you please rerun at 1080p@highest to get an apple to apple with the others?

 

[Det0x]

Could you please rerun at 1080p@highest to get an apple to apple with the others?

I use this as a memory+cpu benchmark..
Dont matter if you run 1080p@medium or 4k@highest, you get the same "CPU Game" score either way (within run to run variance)

Reference for a highly optimized intel system 1440p vs 2160: https://www.techpowerup.com/forums/...-raider-benchmark.250511/page-11#post-4450321
1080p vs 5k: https://www.techpowerup.com/forums/...-raider-benchmark.250511/page-11#post-4457340
1080p vs 1440p vs 2160p: https://www.techpowerup.com/forums/threads/shadow-of-the-tomb-raider-benchmark.250511/page-10

The reason for me using exactly 1080 medium is because it was the settings for this thread: https://www.diskusjon.no/topic/1888561-shadow-of-the-tomb-raider-benchmark-tråden/

But like i said above, it makes no difference for "CPU Game" score.

The only interesting part of this benchmark is the memory (subsystem) scaling, as few other games/benchmarks show it as clearly as this one (and its a fair comparison for both amd and intel cpus)
If i wanted to bench something 90%+ GPU bound, i would benchmark my 3090 in 3dmark, not Shadow of the tomb raider lol
But it seems like some people didn't like the new CPU Game highscores. (?) "Tough luck but its you guys that dont understand how to use this benchmark and are doing it wrong"

 

[Felix123BU]

LOL i was benching the wrong version of the game.. Free version on steam was v505 (patch 15) which was missing 4 performance updates
Version 298 = patch 19

This is more like it :innocent:
CPU Game = Average 282 fps
View attachment 189970
View attachment 189971

I was like goood damn, 282, but then I saw Medium settings

 

[Taraquin]

I use this as a memory+cpu benchmark..
Dont matter if you run 1080p@medium or 4k@highest, you get the same "CPU Game" score either way (within run to run variance)

Reference for a optimized intel system: https://www.techpowerup.com/forums/...-raider-benchmark.250511/page-11#post-4450321
1080p vs 5k etc: https://www.techpowerup.com/forums/...-raider-benchmark.250511/page-11#post-4457340
1080p vs 1440p vs 2160p: https://www.techpowerup.com/forums/threads/shadow-of-the-tomb-raider-benchmark.250511/page-10

The reason for me using exactly 1080 medium is because it was the settings for this thread: https://www.diskusjon.no/topic/1888561-shadow-of-the-tomb-raider-benchmark-tråden/

But like i said above, it makes no difference for "CPU Game" score.

The only interesting part of this benchmark is the memory (subsystem) scaling, as few other games/benchmarks show it as clearly as this one (and its a fair comparison for both amd and intel cpus)
If i wanted to bench something 90%+ GPU bound, i would benchmark my 3090 in 3dmark, not Shadow of the tomb raider lol
But it seems like some people didn't like the new CPU Game highscores. (?) "Tough luck but its you guys that dont understand how to use this benchmark and are doing it wrong"

Yes, but this tread is not about CPU only, it is about GPU aswell Many here want to see GPU-scores. Could you rerun so we can see what actual fps you get with highest setting?

 

[Det0x]

Yes, but this tread is not about CPU only, it is about GPU aswell Many here want to see GPU-scores. Could you rerun so we can see what actual fps you get with highest setting?

Judging from previous comments in this thread, you seem to one of the few who actually know what they are talking about, and since you ask me fairly, i will do a run with graphics on highest later today or tomorrow

 

[Vendor]

not bad for my 750 ti but remember it's 1080p low but i still think my gpu did fairly well

 

[Ruru]

not bad for my 750 ti but remember it's 1080p low but i still think my gpu did fairly well
View attachment 190078

750 Ti is still surprisingly fine.

 

[Vendor]

750 Ti is still surprisingly fine.

yeah true but it was actually 1080p lowest* and now low, i made a slight mistake there

 

[Ruru]

yeah true but it was actually 1080p lowest* and now low, i made a slight mistake there

I suppose that it's playable at lowest with 720p so it still manages.

 

[Vendor]

I suppose that it's playable at lowest with 720p so it still manages.

40fps is more than enough for me and i even fancy stable 30fps because of my gpu

 

[Ruru]

40fps is more than enough for me and i even fancy stable 30fps because of my gpu

I understand, it's not so fast-paced game so it's playable.

 

[Det0x]

Scrubs

*edit*
Forgot to include "framecounter", so other run

 

[Taraquin]

Scrubs
View attachment 190106
*edit*
Forgot to include "framecounter", so other run
View attachment 190107

Impressive both CPU and GPU-wise. Any OC on GPU? Interesting to see that you are almost 50% GPU-bound with the fastest gaming CPU available!

 

[Det0x]

Impressive both CPU and GPU-wise. Any OC on GPU? Interesting to see that you are almost 50% GPU-bound with the fastest gaming CPU available!

GPU is MSI Suprim X on stock cooling and stock bios (max450w) running a OC @ +135 core and +1250 mem.

But like i said earlier, the interesting thing about this bench is the CPU+ memory scaling, not GPU performance..

Oh and where have all these naysayers gone ? They seemed very interested in fruit, especially apples ? Did they learn anything in regards "CPU Game" ?

Dont matter if you run 1080p@medium or 4k@highest, you get the same "CPU Game" score either way (within run to run variance)

 

[Lindatje]

With a small adjustment to the memory, the CPU game goes up. tFAW from 60 to 24. tRC from 87 to 50. tRFC from 350 to 304. tWR from 26 to 14.

And the GPU-Bound is strange, from 67% to 98%....

 

[Det0x]

With a small adjustment to the memory, the CPU game goes up. tFAW from 60 to 24. tRC from 87 to 50. tRFC from 350 to 304. tWR from 26 to 14.

And the GPU-Bound is strange, from 67% to 98%....

Not really, faster CPU Game = you are more GPU bound.
The CPU have to wait for the GPU before it can draw the next frame, så the GPU becomes more and more the bottleneck with a faster CPU and/or higher resolution.

 

[Taraquin]

GPU is MSI Suprim X on stock cooling and stock bios (max450w) running a OC @ +135 core and +1250 mem.

But like i said earlier, the interesting thing about this bench is the CPU+ memory scaling, not GPU performance..

Oh and where have all these naysayers gone ? They seemed very interested in fruit, especially apples ? Did they learn anything in regards "CPU Game" ?

I totally agree that cpu+mem-scaling is interesting, however, many here also wanna compare gpu performance and wants input on weather cpu is limiting perf or not Hence is why someone sa7d that we should test at 1080p highest to compare perf.