Ryzen benchmarking and overclocking results

[BiggieShady]

Memory, that AMD shipped with the CPU: doesn't work at 2666, bought two new kits

That's rough ... considering X370 boards from Gigabyte have memory compatibility list 4 pages long

 

[W1zzard]

I have this for CPU benchmarks. Probably gonna use wPrime 1024M though.

Any feedback?

Also will add game benchmarks after I got those done

 

[W1zzard]

Memory is here, wish me luck

 

[Jhelms]

I was aiming for the 1700 non x for a value beast when overclocked - legit reviews confirmed my suspicions. http://www.legitreviews.com/amd-ryzen-7-1700-overclocking-best-ryzen-processor_192191

Not bad for a day 1 processor. Little bummed on the voltage required for 4ghz stability / overall wattage under full load was higher than I would like to see but she was stable and easily outran an 1800x at stock clocks... Reminds me of back in the day with my 2500+ xp gunning down 3200+ chips

 

[Steevo]

I have this for CPU benchmarks. Probably gonna use wPrime 1024M though.

Any feedback?

Also will add game benchmarks after I got those done

Can we get runs on SuperPi and wPrime back to back compared to between reboots, and with set core affinity and thread priority sir?

 

[Fluffmeister]

I was aiming for the 1700 non x for a value beast when overclocked - legit reviews confirmed my suspicions. http://www.legitreviews.com/amd-ryzen-7-1700-overclocking-best-ryzen-processor_192191

Not bad for a day 1 processor. Little bummed on the voltage required for 4ghz stability / overall wattage under full load was higher than I would like to see but she was stable and easily outran an 1800x at stock clocks... Reminds me of back in the day with my 2500+ xp gunning down 3200+ chips

1700 seems to be the best choice by far considering they all clock much of a muchness. You'd have to be a bit of a loony to pay extra for the 1700X or 1800X.

 

[alucasa]

I personally think 1800k is not worth the extra 100 USD.

1700k for the most users probably.

For me, I am probably going to pick up 1700 once mobo situation is solved.

 

[the54thvoid]

@W1zzard, could you benchmark the games running overclocked memory?

http://www.eteknix.com/memory-speed-large-impact-ryzen-performance/

Seems that the slow kits and sent out are possibly hampering performance...

 

[cadaveca]

@W1zzard, could you benchmark the games running overclocked memory?

http://www.eteknix.com/memory-speed-large-impact-ryzen-performance/

Seems that the slow kits and sent out are possibly hampering performance...

"stock" is not hampering performance.

 

[MetalRacer]

Add RealBench to your benchmark list.

 

[cadaveca]

Add RealBench to your benchmark list.

It's an ASUS-branded benchmark that simply uses other tests in a wrapper. That's why I stopped using it; can get the same results from those actual benchmarks individually anyway.

 

[CAPSLOCKSTUCK]

CPUZ bench would be good

 

[W1zzard]

CPUZ bench would be good

Is it? Not sure myself

 

[CAPSLOCKSTUCK]

it would be good to see where they sit in comparison to those posted here.

https://www.techpowerup.com/forums/threads/share-your-cpuz-benchmarks.216765/page-30

 

[MetalRacer]

It's an ASUS-branded benchmark that simply uses other tests in a wrapper. That's why I stopped using it; can get the same results from those actual benchmarks individually anyway.

I know what it is. I think it should still be added for a good stress test of the OC.

 

[The Lighthouse]

It seems Ryzen is really picky regarding power states (has to be High Performance mode) and SMT... and not to mention Bios update as well.

And it seems using Windows 7 instead of Windows 10 will bring far more performance as well (which is quite ironic), with 17% increase in draw call performance.

 

[Ravenas]

I think communication between AMD and mobo vendors were horrid at best. They had so many intentional leaks to boost hype but they weren't talking to mobo vendors.

I highly doubt there was a lack of communication between AMD and it's MOBO vendors. I think it's more like the lack of competition AMD has provided over the last 2-3 years. Vendors are most likely telling them they don't have the resources to support two platforms... Example, look at the state of software updates and bios updates for 990fx... Any 990fx user knows exactly what I'm talking about.

Hopefully now with Ryzen, vendors will begin supporting a second platform once more with the newly garnished respect that is generated from Ryzen sales and performance alike.

 

[cdawall]

Gigabyte board BIOS: worst I've ever seen.
Memory, that AMD shipped with the CPU: doesn't work at 2666, bought two new kits this morning with same-day.
Results: so far extremely unimpressed

Gaming 5? If so shoot me a pm if you want some of the tricks for it. The board is a hunk of junk that lacks a clockgen chip...so good luck

@cdawall
The obligatory thank you for starters and i mean it (as in, our disagreements in the previous thread aside)
Got two questions if you wouldn't mind:

- As i cannot do the math without knowing the Amperes it took, can you tell me how many Watts total we're talking about? For those 1.475 volts ?
- Mayhap random, mayhap self-evident, but am not really involved in all this jazz.. folks like me don't have 'samples' sent to them. What exactly is it you do in relation to this? Be it as a full or part time job?
edit: no innuendos, should have phrased it better but will not edit it for transparency purposes.

You are just fine I appreciate the call out. As for wattage I do not have the ability to test just CPU consumption but I was looking at 260w at the wall using 1.475v. These really jump up when you push the voltage on them. It looks like the only reason AMD was able to show a "95w" chip was by neutering the clockspeed even in turbo. If I had to estimate the chip based off of temps, personal experience with guessing I would say the chip under load was just shy of 180w.

My current job is in a repair center at a retail store while I am back in school. They had me build a demo unit for AMD. So I ended up with the sample a little early. I am hoping to take the connections I made here with me when I swap to my new job as lead instructor at a local school and hopefully will continue getting samples.

 

[The Lighthouse]

Continuing from my comment regarding Windows 7 vs Windows 10 performance, there is indeed huge performance difference in gaming for Ryzen. From The Stilt in Anandtech forum.

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/page-8#post-38775732

Very noticeable increase in min framerate. I really wish AMD delayed the release of the chips and let Microsoft to actually release the Windows 10 patch before launching.

 

[Iceni]

@W1zzard

A long shot here. Could you do a set of results including older benchmark chips like the 2500K . I'm seeing a lot of reviews that are only interested in current CPU's, and its getting frustrating for me at least to have 3-4 test pages open to try get a good overview. There is no need from my perspective to do a fast review, a good comprehensive one would be far more beneficial.

 

[HTC]

@cdawall and / or @W1zzard

Any chance some underclocking can be included as well?

Wondering if it's worth it to underclock the CPU a bit without losing too much performance!

 

[Caring1]

Continuing from my comment regarding Windows 7 vs Windows 10 performance, there is indeed huge performance difference in gaming for Ryzen. From The Stilt in Anandtech forum.

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/page-8#post-38775732

Very noticeable increase in min framerate. I really wish AMD delayed the release of the chips and let Microsoft to actually release the Windows 10 patch before launching.

I was interested in running W7 on Ryzen when I get it, just to see how it performs in comparison to my current CPU, although all listings I look at for Motherboards show W10 and newer is required.

 

[v12dock]

Not much I can do overclocking on the 1700X I did manage to get my ram from 2166 to 2800 huge different in ram speeds. I agree the whole platform is more a "beta" still a super impressive CPU

 

[Melvis]

I have this for CPU benchmarks. Probably gonna use wPrime 1024M though.

Any feedback?

Also will add game benchmarks after I got those done

Id personally Ditch SuperPi if your going to compare it to other CPU's. (intel)

Anything multi threaded would be nice and id like to see some DOOM benchmarks with Vulkan turned on, seems to love cores.

And it seems using Windows 7 instead of Windows 10 will bring far more performance as well (which is quite ironic), with 17% increase in draw call performance.

Do you have any other links that show tests done on Windows 7? gaming and alike, would be very helpful.

I was interested in running W7 on Ryzen when I get it, just to see how it performs in comparison to my current CPU, although all listings I look at for Motherboards show W10 and newer is required.

Gigabyte website shows they have Drivers for all the new chipsets for Windows 7.

 

[The Lighthouse]

Do you have any other links that show tests done on Windows 7? gaming and alike, would be very helpful.

As I posted above, it's done by The Stilt. This is Total war : War hammer.

https://forums.anandtech.com/threads/ryzen-strictly-technical.2500572/page-8#post-38775732

And it seems in Windows 7, enabling SMT improves performance, while enabling SMT hurts performance in Windows 10.


*Tries to click that Start button my Windows 10, realizes it is no longer working AGAIN* I guess this is why people call Windows 10 a 'beta OS.' The amount of bugs and stuffs sometimes astonish me....