Userbenchmark alternative

[Mqisty]

I don't know all that much about recent hardware and I'm looking for a website that can compare parts fairly accurately. I remember seeing that userbenchmark is not accurate at all and highly biased so I'm obviously not going to use that. I saw people mention technical.city in the Discord, I tried comparing a 9060 XT with my GTX 1080 and it said that the XT is 29.3% better than my 1080, this doesn't make sense to me at all. I feel like the percentage should be a lot higher than that or am I wrong?

[YoungBlade]

Try Techpowerup's Relative Performance chart on their GPU spec pages: https://www.techpowerup.com/gpu-specs/radeon-rx-9060-xt-16-gb.c4293

 

Also, there's Tom's Hardware's GPU Hierarchy chart: https://www.tomshardware.com/reviews/gpu-hierarchy,4388.html

[emosun]

2 minutes ago, Mqisty said:

this doesn't make sense to me at all. I feel like the percentage should be a lot higher than that or am I wrong?

it would depend on what task it's doing

but in general i just use passmarks site

[porina]

If you're looking for GPUs in mainly gaming performance, I like TechPowerUp. Look up one of the GPUs (preferably the newer) and then you can see how others are around it. Better are the reviews although they can have fewer GPUs.

[Shimejii]

3D mark for GPU and cpu. Cinnebench for CPU for programs.

[GuiltySpark_]

Typically experienced users don't use anything like those websites. We use reviews from (multiple) independent sources such as Hardware Unboxed, Gamers Nexus, Techpowerup, etc.. and come to conclusions on our own. Though you're unlikely to find very old cards in these tests. For basic quick and dirty "sanity checks", you can use TechPowerUp's relative performance chart. 

 

Pull up any GPU in TPU (usually one of the first hits Googling a cards name). 

 

[manikyath]

if you use userbenchfart at all (i do occasionally for some reference material), never EVER use their "overall percentages", they're heavily editorializing and weighing the math in favour of the products they prefer.

 

the cpu scores are a hilareous example of this, where intel chips will get marginally higher single core performance, where the AMD chips will get HUGELY higher multicore scores.. but somehow that still ends up with a major victory for the intel chip.. why? because their fudge sliders "sentiment and value" scores 'show' that the intel chip is clearly better.

 

especially for older parts (because userbenchmark relies on user data, which tends to go away when you alienate your users) userbenchmark can be a fairly decent point of reference for specifics, but dredging trough the editorialized mess that it has become is honestly not worth it, especially for cpu scores where an equally valuable database exists over at passmark.

 

it's a shame really.. because the concept of the website made a lot of sense, and they didnt even sell out to enshittify it, it's just all ego.