Cyberpunk 2077: PC graphics benchmark analysis review

For me also, night and day difference on Ryzen 2600. Even though I was GPU bottlenecked all the time it still bumped my fps when driving and - no more stutters when looking around while driving in outside view. Sweet

 

Amazing . I can't belive they left such a bug . All reviews then have wrong scores with AMD platfrom .

 

Okay finally got a chance to really play the game last night. (Got to the point where I meet thy Shug Night looking MF) My settings for my 1080ti are high (with AF set to 16) @ 1440p staying around 60FPS (ultra was netting me 45ish) this game looks not really that different from ultra to high preset. It’s beautiful and the animations are next level. I have has a couple small bugs (like Jackie clipping through a cabinet) but overall that game has been great for me so far.

BTW if anyone is interested my 2060 laptop with the 9750h and 32GB of 2666Mhz dual channel RAM is doing good at 1080p with the high preset and DLSS on balanced. Getting around 70-80 FPS.

Both systems get hot. The 1080ti is hitting 80°C :0

 

I recommend using AMD sharpening , works really well with this game , set it 20-40% ( to your taste) .

Hilbert gona retest with fixed exe ? Game behaves totaly different when uses all threads. it is SMOOTH AS BUTTER on my system . 3300X(4.5ghz tho). + R6800

 

Yeah, I'm starting to see a pattern with the last 3 cards I've owned from Nvidia.
You either run out of VRAM even though the card can still hold itself, or the card inexplicably falls behind in performance. How convenient for Nvidia.

 

I have 60-75 FPS with 4 cores on 100% and others are sleeping. GPU usage is at 50% and when I change settings it only changes GPU usage. So, this game is acting like I have a bottleneck on my CPU (5950x and RTX 3090 combo here).

 

Like I said earlier my 1080ti is more around. 2070-2070ti (where it actually should be) I don’t know what’s going on with Hilbert’s Pascal results but they seem a little low.

 

The user who made the post of the workaround fix said that this workaround will not work with AMD cpus that use multi CCXs such as the 5900x and the 5950x.

 

That is a lie . Today i tested with my friends 1600 AF that uses 2 CCX , he has 3060ti , doubled his FPS.

 

You cant have cpu bottleneck with 5950X .... check Hilbert review.

 

People are saying that it's because architecture was not build for DX12 or that game requirements changed and suit Turing/Ampere better. But still, seems convenient that it happens every generation like clockwork.

 

What I meant to say way it may not work in the sense of that it won't do anything in some cases. But in your friend's case it did increase the performance.

 

IT only matters on high end gpus and city center with alot of AI.

 

I don't think it's just Hilbert, GN has shown similar numbers as well. It just seems like Pascal is getting assed hard in this game for some reason.

 

CPU use after fix

 

Now its a cpu bottleneck. Having all 8threads 100% will only bring you stuttering.

 

Not necessarily?

 

It's not that bad , far from bad , in other games i would say 100% yes , but here it doesnt stutter much it feels very smooth frametime is also quite good i can't complain , game sucked me in hard , dont remmber game that sucked me this hard since ... i dont know .

It is cpu bottleneck , just not as big as it was , it was almost unplayable in the city with only 6 threads.

 

I imagine its better having usage across all 8threads than before but all cores pegged at 100% is not good and it will indeed stutter when you need more.