Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

General project and community related discussions and offtopic threads.
Post Reply
User avatar
Sharlybg
Donor
Posts: 2909
Joined: Mon Dec 04, 2017 10:11 pm
Location: Ivory Coast

Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by Sharlybg » Wed Jul 07, 2021 1:43 pm

Tomorrow I'll be publishing a video 2 years in the making: A render engine speed comparison! Cycles, Vray, Redshift, Luxcore, Arnold, Corona, Octane, E-Cycles & Cycles X, going head to head.

So place your bets: which do you think will rank the fastest, and which the slowest?
https://twitter.com/andrewpprice/status ... 6460720128

Let's see what happen. Lot of people are predicting that Arnold and Luxcore will be at the bottom but let's see what happen.
I hope it is not going to be as biased as this OLAV3D Video from today :

https://youtu.be/ZdiRjqp_YR4

Completelly misleading everyone with bad settings and information.
Support LuxCoreRender project with salts and bounties

Portfolio : https://www.behance.net/DRAVIA

zuljin3d
Posts: 72
Joined: Sun Apr 08, 2018 12:13 pm

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by zuljin3d » Wed Jul 07, 2021 4:56 pm

For some reason, people still think luxcore is only good for caustics... :?
And almost all of them do not know about the existence of autoportals and light cache(PGI).
I'm not even talking about the latest new features in luxcore
Actualy sorry for my google translate english :)

User avatar
TAO
Developer
Posts: 753
Joined: Sun Mar 24, 2019 4:49 pm
Location: France
Contact:

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by TAO » Wed Jul 07, 2021 8:11 pm

That's because we do not have good videos to introduces the LuxCore and all the features, and by video, I mean good, small, commercial-like video.
If you look at the Cycles-X introduction video or unreal engine 5 you will get what I'm talking about.
Even sometimes they exaggerate the abilities and the features, just look at the Cycles-X introduction video you will see strange things like 10x faster and more but when you try it yourself you will notice it can be almost 30% faster (1/3).
We need good advertising to sell our product (not exactly selling you know what I mean), I know we do not have enough resources but we already thinking of fixing that problem.
Omid Ghotbi (TAO)
Last update information
https://forums.luxcorerender.org/viewto ... 590#p30154

zuljin3d
Posts: 72
Joined: Sun Apr 08, 2018 12:13 pm

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by zuljin3d » Wed Jul 07, 2021 9:43 pm

TAO wrote:
Wed Jul 07, 2021 8:11 pm
That's because we do not have good videos to introduces the LuxCore and all the features, and by video, I mean good, small, commercial-like video.
If you look at the Cycles-X introduction video or unreal engine 5 you will get what I'm talking about.
Even sometimes they exaggerate the abilities and the features, just look at the Cycles-X introduction video you will see strange things like 10x faster and more but when you try it yourself you will notice it can be almost 30% faster (1/3).
We need good advertising to sell our product (not exactly selling you know what I mean), I know we do not have enough resources but we already thinking of fixing that problem.
E-Cycles as an example.
People pay $ 900 for marketing scam
Actualy sorry for my google translate english :)

User avatar
TAO
Developer
Posts: 753
Joined: Sun Mar 24, 2019 4:49 pm
Location: France
Contact:

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by TAO » Thu Jul 08, 2021 12:40 am

zuljin3d wrote:
Wed Jul 07, 2021 9:43 pm
TAO wrote:
Wed Jul 07, 2021 8:11 pm
That's because we do not have good videos to introduces the LuxCore and all the features, and by video, I mean good, small, commercial-like video.
If you look at the Cycles-X introduction video or unreal engine 5 you will get what I'm talking about.
Even sometimes they exaggerate the abilities and the features, just look at the Cycles-X introduction video you will see strange things like 10x faster and more but when you try it yourself you will notice it can be almost 30% faster (1/3).
We need good advertising to sell our product (not exactly selling you know what I mean), I know we do not have enough resources but we already thinking of fixing that problem.
E-Cycles as an example.
People pay $ 900 for marketing scam
yes, that's the truth.
Omid Ghotbi (TAO)
Last update information
https://forums.luxcorerender.org/viewto ... 590#p30154

User avatar
Sharlybg
Donor
Posts: 2909
Joined: Mon Dec 04, 2017 10:11 pm
Location: Ivory Coast

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by Sharlybg » Thu Jul 08, 2021 8:17 am

Support LuxCoreRender project with salts and bounties

Portfolio : https://www.behance.net/DRAVIA

zuljin3d
Posts: 72
Joined: Sun Apr 08, 2018 12:13 pm

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by zuljin3d » Thu Jul 08, 2021 9:08 am

Image
:?
Actualy sorry for my google translate english :)

User avatar
Sharlybg
Donor
Posts: 2909
Joined: Mon Dec 04, 2017 10:11 pm
Location: Ivory Coast

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by Sharlybg » Thu Jul 08, 2021 9:21 am

Aside from this issue the problem is that without clamping Luxcore start to produce too much fireflies
and noise. It is something i have already noticed myself in the past even in simple scene. In my test It
Look like Cycles tend to have something that stop this to happen like an invisible clamping code operating
behind the scene. I'm not a dev so can't confirm this.

But here is a simple scene where you can see the behaviour :
Luxcore Vs Cycles.jpg

Just render first with cycles and use the render time of cycles
as Luxcore halt time value to compare noise level. In my test
Both took 113 secs on a i7 4930k

Blendfile
CvsLX.blend
(1.69 MiB) Downloaded 36 times
Luxcore is much noisier and you can see fireflies don't clear
fast enough.
Support LuxCoreRender project with salts and bounties

Portfolio : https://www.behance.net/DRAVIA

User avatar
Sharlybg
Donor
Posts: 2909
Joined: Mon Dec 04, 2017 10:11 pm
Location: Ivory Coast

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by Sharlybg » Thu Jul 08, 2021 9:50 am

To reach the same noise level as cycles you have to run
the render for at least 3X the cycles time.


Cycles 113 secs
Cycles.png
luxcore 348 secs
Lux 5.48.png
I have already try to show that in the past :

https://forums.luxcorerender.org/viewto ... it=porsche

https://forums.luxcorerender.org/viewto ... che#p25539

The Porsche test :

Luxcore
Image


Cycles
Image
Support LuxCoreRender project with salts and bounties

Portfolio : https://www.behance.net/DRAVIA

BlenderBrit
Posts: 9
Joined: Thu Apr 15, 2021 7:19 pm

Re: Giga Benchmark battle : Luxcore Redshift CyclesX ECycles Vray Corona

Post by BlenderBrit » Thu Jul 08, 2021 11:58 am

Hey, fellow LuxCore users.

I was the research assistant mentioned in the video and did all the testing on the various engines. I had a feeling there would be some disappointment from the LuxCore base so wanted to jump on here and explain why LuxCore didn't get it's caching options included.

Basically, it didn't work in our test scenes properly.

In the All-Rounder test for example, switching it on caused some visual issues and actually increased the render time. I have no doubt with some tweaks it could have been fixed and better times achieved but that would have fallen outside of the scope of the tests. If you start allowing a tweak on one engine, you have to allow tweaks elsewhere... then, where do you draw the line? It then becomes a competition between my knowledge of various engines rather than anything else.

I think it's fair to say that LuxCore is definitely more for an intermediate user and above, and I think it's fair to say it's not particularly fast unless you know specifically how to squeeze performance out of it so, I think the tests reflect that fairly.

I'm happy to discuss this further with anyone that wants to, if the video does well we'll certainly do a version 2 somewhere along the line and I'm happy to hear any feedback in how we can improve our testing methodology next time around. I would ask that you contact me on Twitter though (@BlenderBrit) as I tend to only visit these forums when I get stuck using LuxCore.

Post Reply