Everything related to Maxwell Render and general stuff that doesn't fit in other categories.
By Aleksey Borisov
#396186
Hi, maybe someone already asked but I can't find information about properties meaning. What is that?
Benchmark 41.91
Material version 2
Software version Maxwell Render v3
Performance level 0.1

I can guess that material version can increase in case it it was updated. Software version is quiet obvious too. But what is Benchmark number and Performance level?
By luis.hijarrubia
#396190
Most of those statistics should be internal.

Material version is the version of that material in particular. I mean, if original material uploaded by someone is v1, and then Tom or somebody from here has upgrade it in some fashion, it's v2.

Software version is the version we have tested and done the preview image.

Benchmark is the same as in maxwell render, is some fashion of samples computed per time. So it depends on machine performance, and scene (and material) complexity.

As the marker above is machine dependand, I tried to give one marker machine independant. So we render every material on simball scene on similar machines, and then classified them in percentiles. It was suppoused to give some "heavy, medium, light" info. So 0.1 it's that the material is inside the 10% slower materials, it's a heavy one.
By Aleksey Borisov
#396191
Performance level is more or less clear. Lower is slower. But it's not clear about benchmark. Let's imagine I have to explain to someone what is that. So, I should say that material A with benchmark 100 is... than material B with benchmark 1000. What "is" in this case?
By luis.hijarrubia
#396192
Aleksey Borisov wrote:
Wed Dec 27, 2017 12:15 pm
Performance level is more or less clear. Lower is slower. But it's not clear about benchmark. Let's imagine I have to explain to someone what is that. So, I should say that material A with benchmark 100 is... than material B with benchmark 1000. What "is" in this case?
...is Slower. But it's not only about material. The same material can have 100 of performance in one machine, and 1000 in other one. That's why I don't find appropiate to use benchmark for material performances, unless is guaranteed to be done on the same machine.

You can think on benchmark as number of samples / time. It's more complex, but similar to that. So, the higher the number the better.
By Aleksey Borisov
#396193
I see. Thanks. Strange that that kind of variable parameter is in use. It's confusing everyone but still in description like it's can be used for some valuable purpose.
Sketchup 2024 Released

I would like to add my voice to this annual reques[…]