• 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.

Intel Core i7-11700K beats Ryzen 9 5950X by 8% in Geekbench 5 single-core benchmark

Soldato
Joined
14 Aug 2018
Posts
3,370
Geekbench has AVX 512, something Zen 3 doesn't so it will be slower.

It the only thing Intel can win in, its why Geekbench scores were "leaked" over and over and over and over and over and over and over...... again in the build up to Rocket Lake launch, or rather pre-orders before official benchmarks.

And people wonder why others are so cynical about Intel.
Thanks, though I did realise it did use AVX512.

So, for the vast majority of people both using Intel and AMD that don't have AVX512 then it makes for a reasonable comparison but if you are comparing it with Rocket Lake, Knights Mill etc and want to make a comparison not skewed by AVX512 support just in case it is not needed, then you can just use Geekbench 3 (or 4.0) which don't use AVX512.

So, not really trash then?
 
Last edited:
Soldato
Joined
27 Feb 2015
Posts
12,596
Where in Geekbench can you get a relative score for something that only uses say 8 threads? The last time I used version 3/4/5 it only gives a Single Core Score and Multi Core Score. In this regard it is not much better or worse than Cinebench.

Exactly why is that? If you could give a more qualitative reasoning then that would be more informative. i.e. does it favour one brand over the other, are the results skewed in an unfair way etc.

People with brand loyalty to AMD for some reason dont like geekbench.

The reason I suggested it is because it incorporates multiple different tests for different workflows, compression, encryption, encoding, decoding, playing music, app launching and so on, whilst cinebench is basically just a software encoder test.

If you wanted to limit geekbench to 8 threads, you could either try cpu affinity locking it, or maybe ask them to add such a test.

No harm in using cinebench but just I think that shouldnt be the sole performance metric of your research unless all you plan to do is use your cpu for encoding content.
 
Caporegime
Joined
17 Mar 2012
Posts
47,379
Location
ARC-L1, Stanton System
People with brand loyalty to AMD for some reason dont like geekbench.

The reason I suggested it is because it incorporates multiple different tests for different workflows, compression, encryption, encoding, decoding, playing music, app launching and so on, whilst cinebench is basically just a software encoder test.

If you wanted to limit geekbench to 8 threads, you could either try cpu affinity locking it, or maybe ask them to add such a test.

No harm in using cinebench but just I think that shouldnt be the sole performance metric of your research unless all you plan to do is use your cpu for encoding content.

Its not representative of real life, its limited to 8 threads, so an i3 scores the same as an i9, or an R3 the same as an R9.

It also uses AVX 512 which AMD doesn't support, yet, which would be fine but outside of a couple data-centre applications there isn't anything that supports AVX 512. Its completely synthetic at this point.

Its nothing to do with Brand Loyalty, at least not the AMD side, its a perfectly fine application to use, if you want, the problem comes when people mirror what Intel do and push it on people as if its the only true way to measure a CPU's performance, its the opposite of that.
 
Soldato
Joined
6 Feb 2019
Posts
17,464
People with brand loyalty to AMD for some reason dont like geekbench.

The reason I suggested it is because it incorporates multiple different tests for different workflows, compression, encryption, encoding, decoding, playing music, app launching and so on, whilst cinebench is basically just a software encoder test.

If you wanted to limit geekbench to 8 threads, you could either try cpu affinity locking it, or maybe ask them to add such a test.

No harm in using cinebench but just I think that shouldnt be the sole performance metric of your research unless all you plan to do is use your cpu for encoding content.

the reason I don't like geekbench is that it's very hard to compare results, too many versions of the software makes it hard to compare apples to apples, doesn't track system specs correctly, results are easily skewed (overclocking ram = higher geekbench score, yeah that's great for comparing cpu...not), difficult to look up results
 
Caporegime
Joined
17 Mar 2012
Posts
47,379
Location
ARC-L1, Stanton System
Also, Cinebench is a performance testing tool for a real application. Cinema 4D There are many others like it, Blender, 3ds Max, Maya, Softimage, Zbrush, Lightwave, Houdini, Daz3D, Mudbox.... to name a few. They are all used by 2D and 3D artists the world over.
 
Soldato
Joined
28 Aug 2017
Posts
2,755
Location
United Kingdom
i used to use intel heavily in the past going all the way to a i9 7900x on the x299 platform, intel back then had some truly great products but when amd came along with 3rd gen ryzen i took the plunge and got my first amd chip in like 10 years and i was very happy i did, my 3900x build i went for was signifigantly faster than my older i9 7900x and consumed half the power

For me the power saving were a big part of why i switched over, why pay more when you can litrely have your cake and eat it, thing s got even better when 5000 ryzen launched and i was lucky to get a 5950x and holy cow my pc eats everything and dosnt break a sweat, gaming is unreal too which is a massive plus.

Intel seemed to plato at 10th gen what they did on 14nm is astonding in that generation but i'm afraid they cocked up with 11th gen, yes its around 8% better ipc but holy hell the power draw the chips demand is simply far too high, when intel get there act together and refine 7/10nm then yes they very well may be back on top once again, but for now AMD has the crown.

Unless amd do something very wrong with next gen chips it'll be hard for intel to supass but as above if tthey sort out there process node who knows they could do it and we have competition again, which is good for us the consumers :)
 
Last edited:
Soldato
Joined
27 Feb 2015
Posts
12,596
the reason I don't like geekbench is that it's very hard to compare results, too many versions of the software makes it hard to compare apples to apples, doesn't track system specs correctly, results are easily skewed (overclocking ram = higher geekbench score, yeah that's great for comparing cpu...not), difficult to look up results

That can be said for a variety of software really, there is always going to be multiple contributors to performance not just one component.

Ram i/o performance is part of cpu performance really, the problem is windows doesnt show it which gives a false impression, as an example here is a line from basic diagnostics available on linux.

Code:
%Cpu(s):  0.0 us,  0.3 sy,  0.0 ni, 99.7 id,  0.0 wa,  0.0 hi,  0.0 si,  0.0 st

On windows you can basically see only us, sy, id which is userland, system (kernel) and idle cpu time.
There is also interrupts, and io wait time. Faster ram will reduce io wait time, making more resources available to userland processing. So yeah geekbench is a kind of system performance metric, but that applies to many apps.

Ideally any comparisons been looked at should try to keep the variables as close as possible ram frequency/timings, gpu, storage speed etc.
 
Soldato
Joined
27 Feb 2015
Posts
12,596
They don't artificially paint Intel to be better than it really is?

This is why I said AMD fan boys dont like it, if an app makes one cpu look better or worse it doesnt mean its artificial. Are we next going to see claims geekbench is sponsored by intel lol.

What a strange way to rate a way of testing, "I will only like this benchmark if it makes my brand look good".

Also last I checked AMD is on top in geekbench on their 5000 series cpu's.

The correct answer to the question is, those are all content creator apps, as I said cinebech is fine if all you care about is content creation thats not offloaded to gpu.
 
Caporegime
Joined
17 Mar 2012
Posts
47,379
Location
ARC-L1, Stanton System
Use and real application, code compile, rendering, compression, decompression, encryption, encoding, 2D / 3D who comes out on top? AMD or Intel?

I don't care that Intel can open Windows Media Player 0.00246 Seconds faster, i do care that i can bake lighting in Unreal Engine in 6 hours vs 8 hours or even 4 if you're lucky enough to own a 5950X.
 
Soldato
Joined
27 Feb 2015
Posts
12,596
Use and real application, code compile, rendering, compression, decompression, encryption, encoding, 2D / 3D who comes out on top? AMD or Intel?

I don't care that Intel can open Windows Media Player 0.00246 Seconds faster, i do care that i can bake lighting in Unreal Engine in 6 hours vs 8 hours or even 4 if you're lucky enough to own a 5950X.

Now days AMD does, I think their latest cpus are great. Of course I have been saying this for weeks, it was only yesterday I bashed the i/o performance on my intel chip, but my opinion shifts based on the quality of the product, not on the brand name. I have seen people claim AMD 3000 series were better chips when they were not for most use cases, however finally AMD have got there with 5000 series.

This is what geekbench is correctly showing, AMD now it has the best product now does well in more than just content creation.
 
Soldato
Joined
31 May 2009
Posts
21,257
You could go full shill and use usertestbench test, although when i actually investigate the live reports from people suing it, virtually none manage to achieve the cpu score the shills think the processor is capable of.
They can't even fix their own test anymore.
 
Caporegime
Joined
17 Mar 2012
Posts
47,379
Location
ARC-L1, Stanton System
You could go full shill and use usertestbench test, although when i actually investigate the live reports from people suing it, virtually none manage to achieve the cpu score the shills think the processor is capable of.
They can't even fix their own test anymore.

These two applications are all but identical. i think the only difference is the UI.
 

V F

V F

Soldato
Joined
13 Aug 2003
Posts
21,184
Location
UK
Use and real application, code compile, rendering, compression, decompression, encryption, encoding, 2D / 3D who comes out on top? AMD or Intel?

I don't care that Intel can open Windows Media Player 0.00246 Seconds faster, i do care that i can bake lighting in Unreal Engine in 6 hours vs 8 hours or even 4 if you're lucky enough to own a 5950X.

Imagine it in 30 minutes.
 
Back
Top Bottom