This is an Arnold render benchmark to test your pc power! It is a pretty simple Maya render scene, you can download it and run one of the two files I created.
One is a faster test (around 5 minutes on a pretty decent cpu), the other test file is way much heavy and slow. You can use this one to test the stability of your system.
Take a look to the results at the bottom of the page to compare your cpu speed. You can also post your results in the form below.
To unlock the archive use the password antoniobosi. Enjoy!
Arnold render cpu benchmark (fast)
Once the download is finished, open the maya scene and be sure to set the render resolution in the maya render window to 100% (1170x659). Just take a look at the instructions below.
Results
These results comes from maya 2019, Arnold 3.1.2 for Maya, fast benchmark 1170x659px, latest results update on 28 Jan 2019
Use the form below to submit your Arnold Render Maya Benchmark results!
23 thoughts on “Arnold render CPU speed benchmark (outdated)”
Render Time: 2min 39sec (Fast Version)
AMD Ryzen 5950X with 128GB (3200Mhz DDR4, CL16)
Maya 2020
Arnold 6.0.3.0
MtoA 4.0.3
Hi Reza, thank you for submitting your results. You can find the new version of the benchmark here:
https://www.antoniobosi.com/maya-render-tests-reviews-comparisons/arnold-render-cpu-gpu-speed-benchmark/
It is for CPU and GPU, I created two files to test CPU and also GPU. Looking forward to see the speed of your new AMD Ryzen 5950!
Render Time: 1min 24sec (Fast Version)
AMD Threadripper 3970X with 32GB (3200Mhz DDR, CL16)
Maya 2020.2
Arnold 6.0.3.0
MtoA 4.0.3
I just ran the test using a 3950x
imgur.com/XmAEpNu
System:
AMD Ryzen 9 3950x 16 core processor
RAM: 64GB CAS18 3200MHZ
GPU: NVIDIA GeForce 1070
Autodesk Maya 2020
Result: CPU FAST: 2:52
System:
AMD: Ryzen Threadripper 2990WX 32-Core Processor 3.00GHz
RAM: 64GB
GPU: NVIDIA GeForce RTX2080 Ti
Autodesk Maya 2020
Result:
CPU FAST: 02:00
CPU SLOW: 07:28
GPU: 00:07 (Camera AA 5)
GPU: 00:52 (Camera AA 15)
Ryzen Threadripper 3970x
128GB Ram
Result: 1m 23s
Would be nice to have a public Arnold GPU test now that Arnold 6 has been released.
Impressive result!
You are right, I’ll make another file for Arnold GPU benchmark soon.
MacBookPro (150inch, 2018)
CPU 2,9GHz i9
32 GB RAM
macOS Mojave 10.14.3
MAC Maya 2018
Frame:1 Memory: 1127Mb Sampling: 5/5/2/2/2
Render Time 10:10
I read about the cpu throttling problem on new apple macpros with i9. How is your experience?
Hello Antonio, thats great that you have this Arnold CPU benchmark chart. But did you ever tested it with some Xeon processors?
Hello Viacheslav,
I did not had any chance to put my hands on those processors. Generally speaking, Xeons are not faster than i7 or i9. They are built to be operative 24/7, their cost is because they are reliable not because they are fast. Anyway I think any Intel cpu is very reliable, if you want performance I would go for I9s they are just great.
P.s: say I to Moskow for me, I visit that city some months ago and I just loved it! Hope to come back soon.
Antonio
I did on my old(2014) xeon 2630 v2(2 of them) 6c/12t each 2.6ghz and got 9:19. Spec: 64GB RAM 1600, 1080ti, W10pro, ssd m500 from crucial.
Hi all, I’ve just uploaded a new Arnold render benchmark scene for Maya 2019. You can find update results above, fell free to send me your results!
Antonio
I rendered both using Maya 2018. I got the following results:
FAST: 1:16
SLOW: 40:44
My Specs:
CPU: AMD Ryzen Threadripper 1920X 12 Core
Motherboard: ASUS Prime X399-A
Memory: HyperX FURY Black 32 GB Kit 2133 MHz
Video Card: Zotac GeForce GTX 1060 6GB
CPU Cooler: Corsair H100i V2 Liquid Cooler
SSD: Samsung 850 EVO 250GB
HDD: Seagate Barracuda 2TB SATA 6Gb/s 64MB Cache
Power Supply: CORSAIR TX850M 850W
Case: Corsair Crystal Series 570X RGB
Hi Mishra, thanks for your results. What version of Maya and Arnold are you using? Thx
As I mentioned in my previous comment, I used Maya 2018.
.
Also Please share the benchmark graph for render times of SLOW file, the way you have shared for FAST file. It will also help me and others.
.
Thanks.
The Arnold Core version I am using is 5.0.1.0 and MToA 2.0.1
The reason because I dont have a complete set of “slow” results is because I got those numbers from a friend of mine that works with computer custom building and he kindly make my benchmark runs on his client’s computers to test them out and to make me a favor 😆 . I’m not sure he will run the slow test of every machine, because it is very time consuming! Maybe I can arrange this test to just few cpu (maybe the fastest one) and ask him, do you have any preference?
Ohh, I didn’t know that the “fast” benchmark results are from your friend. I thought it was from clients (like me) who sent you their results. Thanks for the clarification. The way I sent you my both results, if the same way if other send you in larger amount, then I think it would be easy to draw a benchmark graph for “slow” files. Otherwise self bench marking with own custom builds will take lots of time.
Anyhow nice files for testing. I hope graph for “slow” will come.
I saw that self made tests are more robust, because we are sure the test is done in the right way and there are not crucial differences in software and hardware (like overclocking). I’ll update the scores on January, so I could also put the slow test for a smaller selection of cpu, thanms for the feedback!
Hey man,
are you using the threadripper at base frequency? ( 3.5 ghz) or you pushed it at 4.2ghz?
Because I have a similar configuration of yours, but with faster memory, at 3200 mhz ( working at 3200 mhz)
and a 1070
and with Maya 2019 I’ve got 1:30.
Hey – thanks for this benchmark test, would you mind if I used it on my YouTube channel in addition to other benchmarks that I do? It’s a neat little test.
By the way, it could be due to several factors i.e. I’m using Maya 2018.4 now… but my 8700K finished the fast test at 1100×620 in 2:38. I have delidded the CPU and used liquid metal on the die, but the test was performed at stock frequencies. But it’s possible that whoever submitted the 8700K result of 3:20 to you was possibly using a terrible cooler which could have throttled the frequencies right down. Anyway, thanks again for the files, great job
Hi Neil, of coarse you can use these results on your youtube channel.. a mention/link would be much appreciated 😆 Regarding the difference in our results I think the major factor of incidence is the version of the software. An update in Arnold or Maya can massively change the render time. I think those tests are valid just in a short period of time, if we talk in absolute values. Anyway they are still consistent in relative values because the difference in percentage from one CPU to another should be always the same. If you have other tests using different cpus we can compare them and see if there’s something to adjust to my data. Have a nice day!