FRONTPAGE AMD Ryzen 7 2700X and Ryzen 5 2600X CPU Review

[Brando]

Wasn't exactly sure what the question was but here's one of the charts from the review. Shows the 2600X Vs the 8700K both 6c/12t. You can see in the first three the 2600X was in the lead, x26 Intel has a much better system in the CL chips compared to AMD and their KL or SL chipse for handling AVX. Their score in this benchmark skyrocketed on coffeelake. Then you have 7ZIP which loves low latency which is another Intel strong point.

View attachment 198485

i could be wrong but i thought i read that intel still wins in single core situations and amd pulls ahead when the software is able to use multiple cores. was wondering how big a difference there is between the 2 brands when measuring single thread vs multi thread scenarios and how many threads it takes for amd to pull ahead due to multi core efficiency (infinity fabric?).

 

[Johan45]

My take on it is that Intel and AMD are neck and neck on core performance when you include the HT/SMT AMD even pulls ahead slightly. So similar CPU core counts will perform nearly the same at the same speed. The problem at the moment with AMD is it runs out of gas at 4.2-4.3 GHz where Intel will chug along up to 5.0 GHz so they still have the speed going for them which helps in gaming, especially poorly coded games that run on a couple cores Vs 8-16 threads. So when you pit the 8700k Vs the slightly cheaper 2700x you get pretty stiff competition. Even at 5.0 GHz the 2700X @ 4.2 will out score it in benchmarks like CB15 or POVray but you might lose 10 fps in gaming because of the nearly 1 GHz speed differential. Doesn't make things unplayable it's just 70 FPS instead of 80 FPS like you might get on the 8700K @ 5.0 GHz

 

[EarthDog]

was wondering how big a difference there is between the 2 brands when measuring single thread vs multi thread

You can also see that in the review. Super Pi is a single threaded benchmark, and others like CB, WPrime are multi-threaded so you should be able to make an assessment based on that as well.

 

[LRG5]

This is a XFR2.0 overclock with a 102 bus on Air

 

[Woomack]

So far no luck on my 2600X. Can pass tests at ~4.25GHz but randomly freezes at anything past 4.1GHz. At auto runs up to 4.2GHz so standard boost... exactly 4.199 in HWinfo64
My 1920X has about the same max stable clock ... but at 12 cores.

 

[Kenrou]

"Maximum turbo core clock 4666mhz" ???

 

[Woomack]

It's theoretical ratio x bclk but it's wrong reading in system info.

 

[LRG5]

@Woomack

Set the Cpu Vcore offset Voltage to -0.0500. Taichi boards over power just a little. If you want you can check it with a meter (its about 5%)



Are you sure about that being theoretical ratio ?

 

[Johan45]

I'm pretty sure it is single core at 4200 true speed is much higher than what you're seeing there in GB4, at least mine is.

4.2 GHz 2600X


4.5 GHz 2600X

 

[Woomack]

Are you sure about that being theoretical ratio ?

Bclk is a little cheat. If conditions are right then 2700X should boost to 43 anyway. Bclk is controlled by separated chip so CPU will still think it runs at 100bclk even though is higher.

My 2600X boosts to 4200 at 100bclk and higher when I set higher bclk. However I have a bad chip and it stays at 4.2GHz only on single cores for short period of time. During longer work it's more like 4.0GHz. Under full load it's 3.7-3.8GHz on all cores. Manual max 4.1GHz but cooling can limit my OC.
Best is to run hwinfo or something else to check at what clock it really runs. I found CPUZ to sometimes show wrong clock.

@Shawn, your results are making me sad, my 2600X can't even boot at 4.3GHz and you are showing 4.5GHz in quite demanding benchmark

 

[Johan45]

That was with cold water though Woomack. 4.3 would be doable though

 

[LRG5]

@ Johan45

what did you paid for that Ram, Dam and in win7. Try in win10.
are you using multiplier or XFR to overclock ?
is the XFR2.0 lock on MSI x470 M7 board ?

@Woomack

do you want to try the XFR2.0 boost (only by e-mail and your eyes only). I have not try it on X470 or with that chip. but I think it should work.

 

[LRG5]

@ johan45

I see the bandwidth numbers with that DDR4 Ram, but what kind of Latency is it?

 

[Clutch_Head]

my question, is it really worth pumping all the extra voltage, added heat, for a couple of hundred Mhz?
seems like A LOT of folks are adding a lot of voltage for extremely diminishing returns.

 

[Woomack]

Auto settings and AMD specs are up to 1.45V so it's not really so much when you set 1.40-1.45V. Of course at standard settings CPU won't work at max voltage all the time and with ~1.45V will run at single cores only. What I want to say is that max OC is not really far from default settings (at least on average overclocking chips).

 

[LRG5]

@ Clutch_Head *

All power setting are at shock with a -5% offset. The chip controls all power it needs.
The real overclocking here is the Ram. I have sub timing very tight.

 

[Woomack]

~60ns is already good latency for Ryzen, can't compare it with Intel. I don't think it really matters in daily usage but some users make a lot of noise that ryzen needs CL14. It affects almost only some benchmarks but barely anyone can see it in games or anything else.
If someone has some time and keep it up to date then would be nice to see Ryzen AIDA64 thread with results from forum users (can be another benchmark). Even though there are a lot of Ryzen users around, I feel like not many are actually sharing their results and AMD section is not as much active as it could be.

I was testing X470 Taichi and I feel like it's overclocking about the same as X370 so max memory clock ~3600. It can run at tight timings etc and are no issues with stability or CPU overclocking. It's only that on X370 ITX or X399M I'm able to set 3866 without bigger issues, while on X470 Taichi max is 3600 and some kits that could run at 3600 will run at 3466. Dual rank kit works the same on all motherboards so max 3466.

Btw. I see that QVL on motherboard manufacturers websites are good source of info about used IC in memory kits. MSI is filling that pretty good, ASRock has a lot of data, sometimes with more detailed IC info. These lists used to have limited info about the IC itself.

 

[Johan45]

There's a thread in here for posting Ryzen overclock results but there was very little interest in it. There's only myself and Bassnut who posted anything. http://www.overclockers.com/forums/...D-Ryzen-Overclocks?highlight=ryzen+overclocks

 

[Woomack]

Thanks for reminding that. I guess that barely anyone knows about it or there would be more results.

I was testing one Micron-based memory kit during weekend and for some reason it overclocks higher than my Samsung kits on the X470 Taichi. Maybe not at tight timings but I was able to set 3733 CL20-23-23 1.35V so the same as on my last Microns on X399. Max on Samsung is 3600 while 3733 can't even boot.
Even at CL20 results in tests are good.

 

[trickson69]

Man you all so inspire me. I just love to see your systems and the absolute raw power you squeeze from them all. I just want to say..
Congratulations and KEEP UP THE AWESOME JOB!
I can post you some of my current offerings. and as not as good as you all I think not to bad for a moron.