Yup, it really is dated, from looking around online it seems to date from around 2004\2005, they just added options to it over the years. Daft amount of time to keep it essentially the same.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Nvidia Has a Driver Overhead Problem, GeForce vs Radeon on Low-End CPUs
- Thread starter RavenXXX2
- Start date
Yup, it really is dated, from looking around online it seems to date from around 2004\2005, they just added options to it over the years. Daft amount of time to keep it essentially the same.
Funny you should mention that, i get the same on my 3090. The more taxing the game becomes the less gpu utilisation occurs, it can start off floating around 60-70% and drop all the way down to 30-40% when there's a lot of buildings, materials etc.
Would love to see an updated control panel, but i kinda see why they haven't innovated in that department... if it works don't fix it mentality.
Caporegime
- Joined
- 8 Jul 2003
- Posts
- 30,063
- Location
- In a house
Contradicts what W1zzard is saying:
https://www.techpowerup.com/review/gpu-test-system-update-march-2021/
https://www.techpowerup.com/review/gpu-test-system-update-march-2021/
Soldato
- Joined
- 28 May 2007
- Posts
- 10,222
That's in DX11. Hardware Unboxed were using dx12 I believe.
I remember Digital Foundry joked in one of their videos that AMD's DX11 driver was so bad it was like it was from the 90s when compared to Nvidia's modern and advanced driver. I wonder what they'll say about Nvidia's seemingly inferior DX12 driver. Let me guess: nothing.
That's what I thought. Now it's the other way around. How times change.
That raises an interesting question doesn't it - is Nvidia's CPU overhead still equally heavy in DX11, it's just that by comparison to AMD's poor performance on DX11 it looks good?
It's something that you can only actually really compare relative to the competition rather than quantifiably in isolation.
Soldato
- Joined
- 8 Apr 2008
- Posts
- 3,076
- Location
- a galaxy far far away
Surely using the cards for their intended purpose ie 4k ultra settings makes the whole video pointless?
Who uses a 3090 at 1080p?
Who uses a 3090 at 1080p?
Future games that are more demanding on the CPU (given the new consoles) may expose this bottleneck further. Also, there are sometimes CPU bound areas in games even on the highest end CPUs.
there's absolutely no doubt that 3000 series experiences heavy CPU bottlenecking on anything below latest gen i7/Ryzen 7, I've been banging that drum when people ask for system quotes since these GPUs launched. The thing is, in my mind, if it was specifically a driver issue then why does the 2080Ti not suffer the same behaviour? That suggests something architecture/bandwidth specific to me.
There's also an argument that if you are spending 1/4 of what you spend on a graphics card on your CPU, then is it really the GPU/Driver's fault?
Whilst I've not yet been able to explain the precise reason for the bottlenecking, I'm not sold on it being a driver issue per se. So far I know that DLSS isn't the issue and overall performance is heavily related to single core performance, as apparent from only minute differences in performance between 5800X/5900X and 5950X (extra cores don't seem to help - which if it was an overhead issue I'd expect to be the opposite).
I might be wrong but that's my view.
There's also an argument that if you are spending 1/4 of what you spend on a graphics card on your CPU, then is it really the GPU/Driver's fault?
Whilst I've not yet been able to explain the precise reason for the bottlenecking, I'm not sold on it being a driver issue per se. So far I know that DLSS isn't the issue and overall performance is heavily related to single core performance, as apparent from only minute differences in performance between 5800X/5900X and 5950X (extra cores don't seem to help - which if it was an overhead issue I'd expect to be the opposite).
I might be wrong but that's my view.
The 2080Ti did exhibit the same behaviour in the video
"8:54 - is ampere to blame" timestamp
Soldato
- Joined
- 22 Nov 2018
- Posts
- 2,841
Maybe in many years time people on a low budget will buy a used 3090 for 1080p. By then it will struggle with 1440p.
After all, people have always bought used 780ti, 980ti etc many years later.
Interesting, can't say we ever saw that but we rarely paired 2080Ti with first gen Ryzen 5. Again though, the performance of the i3 suggests that I'm right about single core vs multithreaded performance and that screams to me that it might not be drivers.
It's happening on a 3070 too. There's no way that a lowly 5600XT should ever be 20% faster than both a 3090 and a 3070.
Leave it Desktop Colour Settings page and it opens instantly, how often do you need to open it anyhow?
Instead of all these "Sky is Falling Threads" from "Experts" I just play my games on my 3090...
Soldato
- Joined
- 25 Nov 2011
- Posts
- 20,679
- Location
- The KOP
How the tides have turned. Remember this was one area a lot of nvidia fans would jump onto to claim amd had driver issues with overhead.