Ryzen 3600 Cinebench Scores

[MaXxBoulton]

The information is all there in my post. I can't get the fabric clock beyond 1700MHz and have it run stably.

If it will not run at the rates speed of 1800 as said by amd should you not return it.

 

[Orange Nexus]

Th

Thanks for clarifying. I seem to have IF steady at 1900MHz now with SoC at 1.15 and RAM voltage at 1.38. Would you suggest tightening timings at the expense of higher ram speed and or IF?

What is your RAM speed set at?

 

[Phil Smith]

What is your RAM speed set at?

3333

 

[Orange Nexus]

3333

All this time i thought the IF should be half the speed of the RAM. Thanks.

 

[Phil Smith]

All this time i thought the IF should be half the speed of the RAM. Thanks.

That used to be the case but with Zen 2 they don't have to be locked into a 1:1 ratio

 

[Orange Nexus]

That used to be the case but with Zen 2 they don't have to be locked into a 1:1 ratio

I see. Thanks. But does it change the mem latency value (lower) when you check using Aida? I will be getting my 3600 soon and will pair it with G.Skill FlareX 3200. Hoping to get at least 3466MHZ Cl14.

 

[Phil Smith]

I see. Thanks. But does it change the mem latency value (lower) when you check using Aida? I will be getting my 3600 soon and will pair it with G.Skill FlareX 3200. Hoping to get at least 3466MHZ Cl14.

It seems to help yes

 

[humbug]

Have you

run a stability test using all of your ram with those timings?

10 minutes in AIDA nothing special but i don't foresee a problem on a 2 hour run, if it doesn't fail in the first 10 minutes it should be fine, i will do some proper stress testing at some point tho.

 

[Phil Smith]

10 minutes in AIDA nothing special but i don't foresee a problem on a 2 hour run, if it doesn't fail in the first 10 minutes it should be fine, i will do some proper stress testing at some point tho.

Lovely stuff. Does this see an improvement on your cinebench scores?

 

[humbug]

Lovely stuff. Does this see an improvement on your cinebench scores?

I doubt it, i'll have a look later, RAM speed is more about gaming performance.

 

[gerardfraser]

I do not agree with Humbug on Ram speed importance in gaming even at 1080p is maybe 5% with bad timings say from 2400Mhz -4000Mhz for 98% of games. Applications only 3% within same range.Higher resolution is even less %.
Anyway off topic for a second just because humbug repeating that memory performance is important in gaming,i have nothing against humbug as I am a AMD fanboy also.

Spoiler

https://www.techpowerup.com/review/amd-zen-2-memory-performance-scaling-benchmark/2.html

Back on topic 3600X 4350Mhz all core

 

[humbug]

I do not agree with Humbug on Ram speed importance in gaming even at 1080p is maybe 5% with bad timings say from 2400Mhz -4000Mhz for 98% of games. Applications only 3% within same range.Higher resolution is even less %.
Anyway off topic for a second just because humbug repeating that memory performance is important in gaming,i have nothing against humbug as I am a AMD fanboy also.

Spoiler

https://www.techpowerup.com/review/amd-zen-2-memory-performance-scaling-benchmark/2.html

Back on topic 3600X 4350Mhz all core

I'm sorry but these days TPU are a bit mediocre when it comes to reviews, they don't really do anything that you or i couldn't do in 10 minutes.

Steve From HUB looked beyond "just setting XMP and be done with it"

It also depends on the game, for example XMP 3000Mhz CL14 (not CL16) to Manual timings 3800Mhz saw gaming gains of 39% Minimums And 21% averages in ASO, huge numbers.

FC:ND 10%
Rainbow Six 5%
WWZ 9%

As you can see ASO is an outlier but where there is one there are more, two of the other 3 still got 10% higher frame rates, i mean in a world where a 3600 is +/- 5% an 8700K that 10% really matters when people say "but the 8700K overclocks" so do these, just in a different way.

 

[gerardfraser]

I am not going to argue with you.I do my own testing,I believe my own testing.
Guess I should add my own testing

Spoiler

 

[humbug]

I am not going to argue with you.I do my own testing,I believe my own testing.
Guess I should add my own testing

Spoiler

What GPU?

 

[gerardfraser]

On those test RTX 2080 ,3600X CPU .

 

[humbug]

On those test RTX 2080 ,3600X CPU .

1440P Ultra.

The whole point of testing CPU performance in games is that the CPU, not the GPU, is the bottleneck, there are two things in your testing which make the CPU less likely in being the bottleneck, you're testing the games at 1440P (a heavy GPU slanted load) and it's a 2080, not the faster 2080TI, not much you can do about that but to me your graphs look mostly GPU limited, 1080P would be better.

Not a criticism, just an opinion, take it or leave it

 

[gerardfraser]

Come on dude get real.Who buys a RTX 2080/ti to play 1080p/720p in the real world.

EDIT:Sorry off topic OP ,I will not do it again.

 

[R3v4n]

My 3600X is at stock, ram is XMP at 3200MHz and I'm getting -

Multicore fluctuates 3600-3715ish no idea why its so up and down maybe background processes/firefox etc are effecting it, my single core last night was 492 I believe, only ran it the once though.

I found even with auto OC/PBO turned on neither setting really makes any real difference in scores, and in gaming there is no difference at all lol. Seems ocing this chip at the moment is a little pointless I get no extra performance out of it really, hopefully once some more BIOS releases come out it will OC a bit better. I'm still running v17 BIOS (most up to date is v19) but it's 100$ stable on my B450 Mortar, after all the booting/posting issues with the MSI BIOS's atm I'm just going to hang on for a few more releases to be sure it's definetly stable before I flash another one.