****Blender Ryzen render battle benchmark****

What do you need? First the render software Blender in version 2.78a
http://mirror.cs.umn.edu/blender.org...-windows64.msi

And after render scene:
https://ulozto.cz/!v5ELMDZKE3k6/ryzengraphic-27-blend with 200 predefined samples
AMD official render scene with 150 samples

In default is render test set to 200 samples quality. AMD ?tested the CPU Ryzen ES with 150 samples and also at 100. AMD RyZen 3.4 GHz 8C/16T got the scores

100 samples = 25s
150 samples = 36s


Show your results+info about samples and CPU-Z. Enjoy guys :)

AMD updated the file on their site to include default 150 setting. Although i havent seen a 6900k@ def score to confirm.Maybe someone here has one.
UPDATED // Guy at hardocp confirmed zen presentation result, 150 setting broadwell working at 3.5ghz.
Also i updated scores with less clutter in background ;-)

i7 5820K 4.50Ghz 6C/12T
setting 150=38,33
setting 100=25.62
http://oi67.tinypic.com/21kwjzp.jpg

FX 9370 all stock, Cool & Quiet, etc 150 samples 1:56:59

http://3800z24.info/FX/150_RYZEN_FX9370_stock.JPG

Quote:

---

Originally Posted by charged3800z24

FX 9370 all stock, Cool & Quiet, etc 150 samples 1:56:59

http://3800z24.info/FX/150_RYZEN_FX9370_stock.JPG

---

Wow, this is awesome jump for AMD here....I thought, Vishera will be better. Interesting is as core i7 6c/12t is strong here compare to FX.

i7 4790k @4.4GHz Blender 150 samples = 59.98s (16GB RAM@2400 CL10).

From that score you can extrapolate that even with perfect scaling, 8c16t Haswell @3.4GHz would finish in over 38s.
At least in Blender Zen is really good and I'm very much looking forward to swapping my main rig back to AMD ASAP :)

Can anyone else run an FX on this test? Also, what are the settings everyone is using?

Quote:

---

Originally Posted by charged3800z24

Can anyone else run an FX on this test? Also, what are the settings everyone is using?

---

As of few days ago, you only need to download the benchmark file from AMD (it has been updated to 150 samples settings) .No other changes in blender are to be made.As for your FX i would set it on FLAT 4.4Ghz without CNQ and also would make sure it doesnt throttle.When i had 8320 ,cnq was working poorly (as in, it would not keep high clocks).

Phenom II X6 1055T @ 3.8Ghz --> 1:47:10 (150 samples)

http://i.imgur.com/uIKjqaY.jpg

@DAVEY

Huh, i was kinda expecting PII HEX to beat Vishera clock for clock, but this is even more pronounced than i thought ...
I Always knew thuban had more pure raw grunt than FX :-|

AMD Phenom II 1090T stock everything

8GB RAM (4x2GB DDR3 1333MHz)

sample 100: 1:39:65

sample 150: 2:25:53 on 1090T 8GB RAM (stock everything)

Im sorry, for comparison must to show Blue side :)

i7-5960X 3.5 GHz, 2400MHz DDR4 = 39.07s, 150 samples oficial AMD scene

https://c7.staticflickr.com/1/556/31...a2cbf746_z.jpg

Still 3s worse than Ryzen at 3400 MHz...At least, in Blender will be Zen really strong.

Quote:

---

Originally Posted by FlanK3r

Im sorry, for comparison must to show Blue side :)

i7-5960X 3.5 GHz, 2400MHz DDR4 = 39.07s, 150 samples oficial AMD scene


Still 3s worse than Ryzen at 3400 MHz...At least, in Blender will be Zen really strong.

---

It matches almost perfectly with calculations from my previous post. Blender scales very well with cores, so does Handbrake which means a lot will depend on Ryzen turbo as to how good ST performance will be.

http://www.xtremesystems.org/forums/...0&d=1482513689

Phenom II 965@4.5GHz 2:13 for 150 Samples.

https://i.imgsafe.org/07b9a479f7.png


1100T @ 4GHZ
32GB @ 1866 cas 10
Velociraptor 300GB
57 process running

1:44:57 / 150 samples

Dual Opteron6386@4,2GHz - 2 "Piledriver" cores glued head by head with 12MB L3 cache common, for each of the 2 CPU, 32c (16 modules) / 32t
32GB DDR3 1600 CL9
100 samples
22,5 sec.

Ryzen looks like its 75% better than a 1100T

36s/144s= 2.9*

20MB/8MB=2.5*

16T/6T=2.666666667*

3.4Ghz/4Ghz=17.6%

:shocked:

:clap:

:D

Fluid Mark and 3Dmark scores please.

blender-2.78a / Win7
i7-3770 @ 4GHz (100 x 40), 16GB DDR3 @ 1600 8-8-8-24-2T.
150 samples: 01:15:85

http://i.imgur.com/Gek01AU.png

x4 965 BE at 4500MHz?!:shocked:

Quote:

---

Originally Posted by FlanK3r

x4 965 BE at 4500MHz?!:shocked:

---

Yes, it is actually stable at 4.5GHz. Mostly I run it at 4.6 with two cores disabled though.

Impressive! If I find the courage or a way to run it (current version needs SSE2), I'll edit this post with a K6-III+ run.
I'm curious about my 965BE, as it was a last year model that does 4.0 at 1.40V for >12h p95. For quick runs at or bellow 1.5V it might be decent for >4200. Don't have the sig system at hand now.

Great thread guys!! I can't tell you how long I've waited for this day. My first intel chip was a Q6600. that's how long I've been waiting.

All this AMD talk has the juices flowing again.

Please AMD don't release these new cpu's with an expensive price!!!!

Any AMD FX-70 4x4 screen shots? :up:

Quote:

---

Originally Posted by Dr_Swizz

Yes, it is actually stable at 4.5GHz. Mostly I run it at 4.6 with two cores disabled though.

---

!!!this is for separate thread about your Deneb. Tell and show us more about this gold chip! I think the best Deneb what I heard.

Quote:

---

Originally Posted by FlanK3r

!!!this is for separate thread about your Deneb. Tell and show us more about this gold chip! I think the best Deneb what I heard.

---

Well, there is not much more to tell, except that for some reason CnQ works despite using a 22.5 multiplier.
I am sure there are other Deneb CPUs out there equally capable though, especially those of later production (I got this one just weeks before they completely disappeared from the market).