AMD's Ryzen 5 3600 delivers incredible Geekbench performance

thats ~1000 points more single core and ~15000 more multithread when compared to my 4790 at 4.5 oc.
Hopefully i'll be able to upgrade in a year or so :)
 
only problem is that geekbench is pretty useless.


not going into depth here but linus torwald and others wrote about it.
you can find it on reddit.



only one thing.. geekbench is meant as cross platform benchmark.

yet my TR has double the performance under linux (compared to windows).


most of the test are said to fit into the L3 cache.


there is so much wrong about geekbench i would not even know where to start.
 
Last edited:
Even still it's a worthy comparison to previous Ryzen. Even if it scores odd comparing between two similar architectures is still valid
 
thats ~1000 points more single core and ~15000 more multithread when compared to my 4790 at 4.5 oc.
Hopefully i'll be able to upgrade in a year or so :)

Did you run the 32-bit tryout because the 32-bit tryout scores considerably less than the 64-bit test?

Here's a comparison with a Haswell mobile i7-4980HQ
2rhwlt5.png


I'm meaning to write a user review for this BGA to LGA 4980HQ CPU that costs about half the Ryzen CPU mentioned above but enthusiasm has taken it's toll over the last five years to the point of even considering becoming a computerless person :eek:

I still remember running my old SNB 2500K at ~5GHz and now 8 years later we are still at 5GHz. Other improvements are all well and good but clock frequency advancement is very sad.
 
The main point here is that Ryzen 3rd Gen is overcoming the performance limiting factors that have held back the architecture's Geekbench performance. There appears to be better cache performance with AMD's newest processors, which is great news for a number of applications.

I agree that Geekbench is not the best benchmark on its own, but it is useful for highlighting some things.
 
You won't go in depth because you don't really understand it. Lol. Just like linus. Sure he created linux. But it seems he stopped evolving and can't keep up with the times.
And he said that years ago about geekbench3. Don't be a sheep that blindly follows and spouts the same unfounded claims just so you would look smart on the interwebz

By your linux having a better gb score than your windows score just means linux used the hardware better or it has better performance/programming/interpreter than windows. Cross platform means it can run on multiple os without changes to its code...just like java.

Just as long as it is not programmed for a specific brand then it should be no problem.

only problem is that geekbench is pretty useless.


not going into depth here but linus torwald and others wrote about it.
you can find it on reddit.



only one thing.. geekbench is meant as cross platform benchmark.

yet my TR has double the performance under linux (compared to windows).


most of the test are said to fit into the L3 cache.


there is so much wrong about geekbench i would not even know where to start.
 
Back
Top