Shadow Of The Tomb Raider - CPU Performance and general game benchmark discussions

[Block10]

Hi,
Here's mine, Stock. 12600k, 3070ti, DDR5 32GB RAM running 5600 mhz.

 

[Det0x]

Nothing earthshattering, but decent enough i guess

7950x
Memory @ 6100MT/s 28-36-36-28 1T
FCLK @ 2200mhz

1080p lowest
CPU game average fps = 385

Screen with hwinfo open and rest of memory settings shown


Getting pretty much same numbers @ 1080p highest..
And ~fully GPU bound with a watercooled 3090 without using resolution modifier.

1080å highest
CPU game average fps = 383

*edit*
And a CS benchmark score since there is no thread for that..

1080p highest = 991 FPS

 

[Det0x]

With new higher performing ASUS bios:

7950x is the second cpu after 5800x3d to break 400 avenge cpu fps!

Memory @ 6100MT/s 28-36-36-28 2T
FCLK @ 2200mhz

 

[BetrayerX]

5800X and 5700XT 16GB DDR4 3200

 

[Franz]

5800X and 5700XT 16GB DDR4 3200

You need to run in lowest settings

 

[BetrayerX]

You need to run in lowest settings

Ooopps, my bad. Fixed! ^_^ Thanks.

 

[Franz]

Ooopps, my bad. Fixed! ^_^ Thanks.

Its overclocked?

 

[Colddecked]

5800x3d (with -30 offset), 3080 with undervolt, 32gb 3200mhz ram clocked at 3800 @ 16-20-18-36 timing

 

[BetrayerX]

Its overclocked?

Yes! LLC on 4 on BIOS, the rest via Ryzen Master. CC was done per core.

 

[Det0x]

Me and a few other guys over @ the overlock.net forum seems to have reach the same conclusion.. The more GPU bound you are, the higher CPU Game Average numbers you can get:

Can share some legit number with my decently clocked 3090 @ 520w powerlimit:

1080p high = 412 fps average cpu fps @ 98% gpu limited


1080p lowest= 408 fps average cpu fps @ 28% gpu limited


720p lowest = 402 fps average cpu fps @ 1% gpu limited


But watch what happens with the average cpu fps numbers when i enforce a low powerlimit for the graphic card:
441 fps average cpu fps@ 100% gpu limited

Conclusion is that the more GPU limited you are, the more time the CPU have to push up the average cpu game numbers..
Going forward i suggest that either we run with resolution modifier at minimum to limit the GPU bottleneck as much as possible, or we run the benchmark @ 720p lowest like the Russians do..
Anything above 10% GPU limited should not count in my book.

(all this started with a guy showing too good to be true numbers, with a 4090 which was ~90% GPU limited @ 1080p lowest)

 

[cRs]

i3 12100F with rtx 2080 16Gb DDR4 3200Mhz all stock

 

[harm9963]

PowerColor Red Dragon RX 6800 , nice upgrade from my 1080Ti for 5 years , the 1080Ti will go into my second rig , to replace 290X CFX .​

​

 

[Det0x]

Seems like AMD Radeon GFX driver is alittle faster than Nvidia in this game..

1080p lowest = 409 average cpu game average

Hardware and settings:


720p lowest = 413 average cpu game average

 

[harm9963]

PowerColor Red Dragon RX 6800 , nice upgrade from my 1080Ti for 5 years , the 1080Ti will go into my second rig , to replace 290X CFX .​

​

View attachment 277610View attachment 277657View attachment 277662​

Went back to Micro Center , exchange the 6800 for PNY XLR8 4070Ti !

 

[Athlonite]

Went back to Micro Center , exchange the 6800 for PNY XLR8 4070Ti !

Why

 

[QuietBob]

Reference 7900XTX with stock clocks, RAM @ 3800 CL14 tweaked:

 

[Fizban]

Me and a few other guys over @ the overlock.net forum seems to have reach the same conclusion.. The more GPU bound you are, the higher CPU Game Average numbers you can get:

Can share some legit number with my decently clocked 3090 @ 520w powerlimit:

1080p high = 412 fps average cpu fps @ 98% gpu limited
View attachment 268658

1080p lowest= 408 fps average cpu fps @ 28% gpu limited
View attachment 268659

720p lowest = 402 fps average cpu fps @ 1% gpu limited
View attachment 268660

But watch what happens with the average cpu fps numbers when i enforce a low powerlimit for the graphic card:
441 fps average cpu fps@ 100% gpu limited
View attachment 268661
View attachment 268662
Conclusion is that the more GPU limited you are, the more time the CPU have to push up the average cpu game numbers..
Going forward i suggest that either we run with resolution modifier at minimum to limit the GPU bottleneck as much as possible, or we run the benchmark @ 720p lowest like the Russians do..
Anything above 10% GPU limited should not count in my book.

(all this started with a guy showing too good to be true numbers, with a 4090 which was ~90% GPU limited @ 1080p lowest)

I'm confused how you're 98% gpu-bound at 1080P with a 3090. My laptop is 15% gpu-bound at 1080P highest, with hardware I'd expect to perform similarly to yours.

Guessing my CPU (i9-13980HX) must only come close to 7950X in certain workloads, as I get way worse fps in this game. 215 fps at 1080P highest, 185 at 1440P highest, 99 at 4K highest.

It's probably because while it can perform similarly in something like Cinebench, that's with it pulling 200+W, whereas while when gaming it can't pull that kind of wattage while the gpu is also loaded.

 

[BetrayerX]

 

[Det0x]

First initial 7950x3d run for me

Not sure how good these numbers are compared to other Zen4X3D riggs, but atleast low's are pretty good i think.

• CPU Game min = 370 fps
• CPU Game average = 462 fps
• CPU Game 95%s = 400 fps (look at the flat line in graf )

Windows 10

 

[68Olds]

I think I set up the bench correctly.
5800X3D & RX 6800 XT @ default.

 

[QuietBob]

I think I set up the bench correctly.
5800X3D & RX 6800 XT @ default.

View attachment 287280

Wow, your CPU scores are higher than mine even though I was using 3800 CL14 with tight timings Guess I'll have to re-bench with my current dual rank setup.

 

[Lew Zealand]

R5 5600 +PBO, 3200MHz CL18, RX 6800 XT, 1080p Lowest, no AA

Specifically wanted to do this since 68Olds has the same GPU, plus the logical gaming CPU upgrade from this already good gaming CPU.

 

[Duvar]

This bench is really a good one for the 3D CPUs, but i also saw a nice 7950X result here.

WIN 11: 12700k 5.4/4.8 (E/HT off) 4100CL15

Difference vs stock is brutal tho XD

WIN 11: 12700k stock 3200CL16

 

[masterdeejay]

Cpu at 4.8ghz all core, ram 4000mhz

 

[DemonicRyzen666]

There is a problem with this benchmark

The RTX 4070 Ti CPU Bottleneck - How Much? Compared to the RTX 3090. - YouTube

How does it get the same average fps & have diffence of 27% gpu bottleneck ?