THREAD CLOSED!!! Post Your AMD RyZen Blender Benchmarks at 200 Samples!

[Recon-UK]

Welcome to Blend, unreliable crap benchmark, 3570k stock is faster than a stock Haswell i7 LMFAO.

 

[dyonoctis]

I'm guessing i'll just wait for the launch of Ryzen, and see how it bench then. According to what the blender benchmark is showing in this thread, Ryzen went from pretty good to " un coup d'épée dans l'eau" (meh).

 

[P4-630]

Ok, got it working again, the error was caused by my firewall, fixed.

1:08.83

 

[Elezorn]

v2.78
2:10 200
1:06 100
v2.77
2:12 200
1:06 100

i7-3630QM@3.2GHz

 

[Atnevon]

2:53:76 OsX Yosemite!?!?!?!?!?

Well I'm testing all my machines. So...well, lets mix it up some. Getting a new Mac for work soon so thought to test my old one first

 

[ShurikN]

Man this thread got f** up really hard. Different versions, different settings. 2 hours from now someone will notice something different form the zen render as well. There's really no point anymore. Lets wait for January or whatever.

 

[Ionut B]

I7 5820K @ 4.5 GHZ, memory is in dual channel. will do a quad channel next year.
Edit: I got the blender that amd linked on their website, I dunno anything about rendering stuff so I left the default settings. Win10x64 if it matters.

 

[P4-630]

 

[X828]

4770k@ 4.3Ghz

 

[P4-630]

/thread

 

[CAPSLOCKSTUCK]

unsubbed

 

[_JP_]

So, because this started to lose all kinds of coherency, I went ahead and did some more tests, just 2.78 since there's negligible difference between versions....can't really see the reasoning behind these values and what AMD tried to prove besides the usual PR move "Let's use this tool that not many know about and allows to run a specific function that clearly benefit our single point of view and/or product without actually explaining why it is better for that reached score".
I mean, I'm eager for Zen...but...well, salt it is then...until the engi versions start to leak with more popular tests.

To recap:
02:51:87 - 100 samples
05:38:00 - 200 samples
11:21:46 - 400 samples

Spoiler

 

[TRWOV]

3:43:78

 

[biowiet]

Sure people can pick wrong targets and whine about this and that, but the bottom line is, AMD is showing vast improvements on main fronts.

Let me clarify what I mean to say... I have built this 4th gen rig cca. 2 years ago when 5th gen was on the shelves already. Since then we have 6th gen running for a while and AMD is now closing up to that generation IF we believe that they were playing the benchmark show on a fair and honest way.
Now my mind goes like: what we've got as a tool to check their claims is a download link to a file they've created themselves and a download link to a software to use with the downloaded file. What - whatnot, quickly turns out they are running a different version of that software and the settings file they gave us is also different than the settings they run the test on their show. Actually makes people believe their rig is damn slow compare to Ryzen. My conclusion from my own experience that I've posted on pg1, the older version of the software gave me better results and matching the settings (at least in that 1 already known value 200 vs 100 samples) cut the rendering time in half. With that my config's performance went from 1m30s to 0m39s and that is just around their numbers. And don't get me wrong, I don't believe my 4790K on stock speed could match a 6900K so I can't help to think they have rigged the game so that means they have to make up things look better and that means their product can't do the thing they claim and it means what it means. Intel's new generation is coming, and if AMD has to play tricky to match up with previous gen, than that's it... I think the performance won't exceed 6700K, but you know what... I honestly hope AMD will contradict me and I can get back to them.

Long story short, I'm really sorry if I sounded like whining to your ears I did not mean to... I'm just sharing my results and my conclusion. Peace.

 

[XiGMAKiD]

Here's mine
Render sample 200

Render sample 100

 

[Senbonzakura]

Core i73930K 32.56 Render sample 100

 

[natr0n]

Strangers always crawl out the depths to post bench scores never to be heard from again !!!

 

[Ferrum Master]

Strangers always crawl out the depths to post bench scores never to be heard from again !!!

I call them zerglings usually... really fitting name.

 

[GoldenX]

Forgot to take a screenshot.

11:31:54 with current specs.
Also teste a (stock) Celeron Haswell, 4:25:31.

lol

 

[TheHunter]

where do i change sample rate to 100?


This is what I got at default 200?


Ok found 100;

 

[XiGMAKiD]

where do i change sample rate to 100?


This is what I got at default 200?
View attachment 82006

After opening the file before running the render, on the lower right under the submenu Sampling change the Render value

 

[Deleted member 50521]

Sorry kids the OP has given up. Too much hassle due to keeping up with all the different versions.

 

[Scrizz]

YOLO

 

[R0H1T]

Never in the history of AMD have so many enthusiast Intel cores been used to render an AMD CPU logo!

Just repeating someone else's kind words

 

[darksf]

Guys i believe i finally understood where the confusion with the blender bench is coming from.
First obviously on the Live stream they did the CPUs took around 36 seconds but on unknown version of Blender and unknown samples count!

The 100 samples count comes from the closed doors presentation that has been done a week earlier i believe and the videos from there.

actually on this video i swear i think it is 24.82 but can't be sure someone with bigger display to confirm.

According to the Videos they were runing Blender 2.77 at 100 samples and task was completed in around 25/26 seconds according to WCCFTECH below are the numbers but not sure from where since they were around 36 on the live one.

http://wccftech.com/amd-ryzen-zen-cpu-benchmarks-official/

AMD RYZEN ES (3.4 GHz) 25.57s
Intel Core i7-6900K (3.7 GHz) 26.01s

I believe on the live stream the bench was done with 150 samples atleast looks like this given the results if it can be set to 150 cause i don't have blender and i don't know.