Ryzen 5 3600 vs higher priced process

[Baldbloke]

Hi,

I have a ryzen 5 3600 cpu.
Out of curiosity I've been looking at prices of newer ryzen cpus.

Now, I admit I've not done any proper research on what's a good upgrade but what little I've done on Userbenchmark suggests the improvent on the ryzen 5 3600 against a chip twice the price is somewhere between 3% and 11%.

Eg compare the 3600 (6 core, 12 thread) with a ryzen 9 3900xt (12 core, 24 thread) , the improvement in performance is roughly 3 percent. Yet in retail shops the price of the 3900xt i(circa £410) s more than double the 3600 (circa £189).

I'm missing something here, ain't I?

How useful, even ball park accurate, is Userbenchmark?

 

[Vince]

Hi,

I have a ryzen 5 3600 cpu.
Out of curiosity I've been looking at prices of newer ryzen cpus.

Now, I admit I've not done any proper research on what's a good upgrade but what little I've done on Userbenchmark suggests the improvent on the ryzen 5 3600 against a chip twice the price is somewhere between 3% and 11%.

Eg compare the 3600 (6 core, 12 thread) with a ryzen 9 3900xt (12 core, 24 thread) , the improvement in performance is roughly 3 percent. Yet in retail shops the price of the 3900xt i(circa £410) s more than double the 3600 (circa £189).

I'm missing something here, ain't I?

How useful, even ball park accurate, is Userbenchmark?

Userbench... thats the problem.. a 5600 is about 20% faster than 3600. 3900 is a good 10% (more multithreaded) than 3600. Basically just ignore UserBench.

For example a 3600 scores about 9000 points multi cinebench r23, my 3960x scores about 31100 or a little over 3 times the performance in multithreaded workloads.

 

[dave28]

Worthwhile upgrade would be 5800X or am i missing something here. Two more cores and much better single thread performance

 

[Vince]

Worthwhile upgrade would be 5800X or am i missing something here. Two more cores and much better single thread performance

Yea basically this...its about 20% faster single and then what almost double in multi?

 

[andy_mk3]

How useful, even ball park accurate, is Userbenchmark

They're utter trash Intel fan boys.

 

[Baldbloke]

Thank you. folks. It seems then that Userbenchmark is not a reliable gauge.

Since first asking the question I've had a look at some web sites and YouTube that show performance stats when pairing 3600 with 3080.
My 3600 will probably be quite acceptable with my recently acquired 3080 ti.
I don't expect I'll need that upgrade yet anyway.

 

[Baldbloke]

Double post

 

[rumple9]

depends on screen rez. At 1080p you need a high end cpu. When you get to 4k, the cpu doesnt matter so much and Ive seen some benchmarks at 4k where the 3600 beats the 5800.

As for synthetic benchmarks ignore them, only real world gaming benchmarks matter

 

[Ross Thomson]

looking at reviews there is very little real world difference in gaming at 1080p on my 5700XT with a 2600 (4.0ghz) and a 5600x. I wouldn't swap out my CPU purely on gaming results. I do render videos so I would like more CPU hp.

 

[Tetras]

Thank you. folks. It seems then that Userbenchmark is not a reliable gauge.

Since first asking the question I've had a look at some web sites and YouTube that show performance stats when pairing 3600 with 3080.
My 3600 will probably be quite acceptable with my recently acquired 3080 ti.
I don't expect I'll need that upgrade yet anyway.

If you were doing something that used the extra cores then the 3900XT would destroy the 3600. It could complete some workstation tasks in half the time, for example. That is well worth the money if it's your job.

In games that don't use the extra cores, then they are effectively the same CPU, apart from clock speed and cache differences, so you would expect marginal gains.

Zen 3 (5600, 5800, 5900) is different because the cores are faster, but it still only applies if there is a CPU bottleneck (apart from the lows), so at higher resolutions, you'll see little difference between a 3600 and 5600, unless it's a very high-end graphics card that has removed the GPU bottleneck.

 

[StephenJ]

Thank you. folks. It seems then that Userbenchmark is not a reliable gauge.

Since first asking the question I've had a look at some web sites and YouTube that show performance stats when pairing 3600 with 3080.
My 3600 will probably be quite acceptable with my recently acquired 3080 ti.
I don't expect I'll need that upgrade yet anyway.

benchmark your system then you can look at reviews to see if your system is getting the performance it should for the 3080ti

 

[Joxeon]

I went from a 3600 to 5800X and gaming at 1440p with an RTX3080 the upgrade was very underwhelming, the largest difference was 10% but many games showed no improvement at all and the ones which did were older games which were already running over 150fps.

tomb raider 3600

tomb raider 5800X

Then in watchdogs legion 3600

5800X

 

[jigger]

I’ve got a 3800X and 3080 and the 3080 is very much hamstrung. The performance is decent enough but a 5800X would be a much better match. I’d imagine a 3080ti would be held back a lot more.

 

[StephenJ]

do you have benchmarks to show the slowdown ?

 

[rumple9]

https://www.youtube.com/watch?v=DFqob9pSZhk&t=359s