Ryzen 2700x same performance as 3700x on x470...

[Keirox]

After updated to a 3700x to maybe get that 10%-15% more performance in games, i just tested a lot of games, GTA V, The Division 2, Metro Exodus, CSGO, and some others. Switched back to 2700x to verify the same scenes in game to see the difference.

 

And i'm really disappointed to see that both processor give same FPS in the same scenes. Thats crazy. The only place that 3700x win is in cinebench r15 and 3dmark spy.

When i see all the benchmark videos on youtube, i start thinking its just for marketing. God damnit.

 

I didnt wanted to go intel and buy a new motherboard because already had x470, and now i feel like i'm just gonna return the 3700x and buy an Intel 9700K.

 

So for the guys who wonder how performance will go if you upgrade from 2700x to 3700x for gaming and keep an x470 motherboard, its just useless.

 

OC used for the tests : 2700x at 4,2 GHz 1,4-1,5 voltage (basic precision boost overdrive) 3700x at 4,325 GHz, 1,3875 voltage

 

What a big mistake i did ! If someone have any reason that could explain that shit i would be glad to understand why. I did test on same system, same settings and ram profile, just swapped processor without touch anything else beside OC.

[Damascus]

What was your testing methodology, and what GPU are you using? A difference of 10-15% can be basically invisible to the untrained eye, especially when you're expecting more than what it actually entails

[Keirox]

2 hours ago, Damascus said:

What was your testing methodology, and what GPU are you using? A difference of 10-15% can be basically invisible to the untrained eye, especially when you're expecting more than what it actually entails

I have a 2080, but playing under 1080p. I cannot be GPU bottlenecked, it never run more than 60% in most games.

 

My tests are perfectly equal. As i said as i did make test with 3700x, in the same places and conditions, and then switch to 2700x, and do it again...

[Keirox]

I didnt even saw a 5 fps improvement in any of the games. Its or the same or even sometime 2700x give some more...

[Damascus]

Just now, Keirox said:

I have a 2080, but playing under 1080p. I cannot be GPU bottlenecked, it never run more than 60% in most games.

 

My tests are perfectly equal. As i said as i did make test with 3700x, in the same places and conditions, and then switch to 2700x, and do it again...

How did you measure the difference?  Not to mention, you're running a 2070 at 1080p so there's basically no extra performance to easily rustle up 

[Keirox]

What do you mean ? As long as my GPU is not working at 100% the processor can bring more performance. I'm not GPU bottlenecked.

 

I measure the difference by staying on the exact same places in the game, looking at my feets, so it give the most stable results.

[Damascus]

4 minutes ago, Keirox said:

What do you mean ? As long as my GPU is not working at 100% the processor can bring more performance. I'm not GPU bottlenecked.

 

I measure the difference by staying on the exact same places in the game, looking at my feets, so it give the most stable results.

That's... the worst possible method I can think of.  Get some actual benchmarks, like the final fantasy benchmark 

http://benchmark.finalfantasyxv.com/na/

 

The rainbow 6 siege built in benchmark, the free time spy and firestrike benchmarks, and LITERALLY ANYTHING ELSE THAT'S A PROLONGED BENCHMARK will get you meaningful results 

[Keirox]

3D Mark Spy did gave more points on 3700x. 

 

https://www.3dmark.com/spy/8350158 3700x

 

https://www.3dmark.com/spy/8350679 2700x

 

 

But if i dont see any difference in my own games even if i look rly closely at my frames, then those results doesnt matter at all.

 

 

[Damascus]

2 minutes ago, Keirox said:

3D Mark Spy did gave more points on 3700x. 

 

https://www.3dmark.com/spy/8350158 3700x

 

https://www.3dmark.com/spy/8350679 2700x

 

 

But if i dont see any difference in my own games even if i look rly closely at my frames, then those results doesnt matter at all.

 

 

Play 10 games of CSGO on the same map with each cpu, and then compare the averages.  I garuntee theres a difference, it's just that your measurement is completely irrelevant 

[Keirox]

I tested the whole day on all my games and a couple of time, at different scenes. I know what i say.

[Keirox]

Pretty sure if i had bough a i9700k the difference would be huge

[Guest]

 

Your method seems.... unstable.

[Keirox]

Hmm all i did is switch processor and start time spy again

[Guest]

That isn't doing much to control variables, therefore your testing is not useful. You need to figure out why your GPU score went down before you can say the blame is on the CPU. So like @Damascus said, your testing methods are at fault, not the chip.

[Damascus]

4 minutes ago, Keirox said:

Hmm all i did is switch processor and start time spy again

In the GPU testing benchmarks, the 2700x was 8% better because the 2080 was running at a higher frequency

 

The cpu, when tested on it's own, nets you a whopping 14.4%

[Keirox]

It doesnt matter. Whats should be important is the CPU score difference. And yes it give 14.4% more for the 3700x, but i'm still blaming my 3700x for not see any improvement in games. You guys would be pissed off if you was in my situation aswell. I think its due to x470 not completely supported...

[Damascus]

Just now, Keirox said:

It doesnt matter. Whats should be important is the CPU score difference. And yes it give 14.4% more for the 3700x, but i'm still blaming my 3700x for not see any improvement in games. You guys would be pissed off if you was in my situation aswell. I think its due to x470 not completely supported...

No, it's because your GPU was just straight up higher clocked in one of the comparisons.  I've had situations like this in the past l, like when I moved from my 6820k to an overclocked 690l50x, with seemingly no changes in performance.  But because I was willing to accept that there might be other issues, I spent the time and effort required to diagnose the issue (higher heat output in my loop led the the GPU losing about 25mhz, leading to a 7% decrease in performance.  Solved by adjusting fan curves)

[Keirox]

GPU mhz is something changing frequently. Mine is OC stock i never touched it

[Damascus]

1 hour ago, Keirox said:

GPU mhz is something changing frequently. Mine is OC stock i never touched it

And that's why you'll see a variance between runs.  Auto OC changes based on heat, so the extra temperature added by the beefier CPU could be reducing that GPU performance, leading to the lackluster gaming performance 

[Damascus]

@Keirox wait, are you running 1080p 60hz?  Because you'll never see a difference between cpus at that res when running a 2080

[Keirox]

9 hours ago, Damascus said:

@Keirox wait, are you running 1080p 60hz?  Because you'll never see a difference between cpus at that res when running a 2080

Lol ^^ 1080p 144 hz off course :)

 

I think i fixed my problem i reinstalled Windows and everything and turn off SMT mode and it seems to work fine in games, i see the boost properly now.

[Damascus]

On 8/31/2019 at 7:29 AM, Keirox said:

Lol ^^ 1080p 144 hz off course

 

I think i fixed my problem i reinstalled Windows and everything and turn off SMT mode and it seems to work fine in games, i see the boost properly now.

Did you have smt off with the 2700x?

[C17H23NO2]

Maybe don't look so close at the FPS-Counter and play games!
Game is running absolutely smooth on your desired graphic settings ? Awesome. What else do you want ^^ 
The 3700x is an awesome CPU. No Discussion needed. You're good to go man, enjoy your awesome rig!