Chaos Corona for 3ds Max > [Max] General Discussion

Corona 1.3 Benchmark

<< < (51/55) > >>

Juraj:
To be honest, even the latest Cinebench R23 gets some numbers and windows version all wrong. Might just be hard to get accurate reports from unusual platforms.

absolute-fx:
Modeling workstation:
i9 13900 ks, msi z690 ace, 128Go DDR5 5600MHz (4000MHz because xmp does not work with 4 sticks... another story...) , rtx 4090 suprim X, windows 11
It renders until the end but crashes when it populates results...
-> 44s,  11 185 291 rays

Notebook workstation (MSI RAIDER GE77HX)
i9-12900HX, 64Go DDR5 4800MHz, RTX 3080TI, win 11 (also crash at the end)
-> 1.13m, 6 308 106 rays

Rendering station (5 years +)
Dual xeon E5 2696 V4, 128 Go DDR3 2133MHz, GTX 1080 Rog Strix, Z10 PE-DW WS, windows 10
-> 48s, 10 710 948 rays

The result in cinebench R23 tells a different story... +38.000 points for I9 13900ks and +28.000 points for the xeons (quite huge as difference...)
It makes me questioning. Does bucket rendering takes more advantage of the processor power? Does corona can take full advantage of 12 or 13th gen proc architecture?
If we use blind maths the I9 should have been rendered in about 32s.

Just trying to understand... Any guess?

absolute-fx:
I've just run cinebench r23 on the notebook and the results seems to be +/- on the good ratio with 19.000 points.
So I just wondering why this ratio is so different with a 13th gen proc.

Juraj:
I would not pay too much heed to any result from current benchmark. Cinebench R23 is far newer, and Corona will eventually prepare new benchmark :- )

What I do to compare pure "multi-threaded" speed (but I rarely do that nowadays since I value smoothness of workflow more than brute speed) is using production scene, and only paying attention to Ray/s.

Don't worry about memory speed for both 13th gen Intel and Zen3+ platforms for work. Neither can support more than 4600 MT/s +/- when memory reaches total of 8-ranks. Which for DDR5 memory usually means 4 DIMMS. It doesn't affect almost any workstation task at all, I've tested this with dual-channel 64GB 5600/6400 on my HX12800 & HX 13800 laptops vs 128GB at 4000 on my Dell&MSI laptops. Since with lower MT/s you can also lower the timings, the latency is identical, you only lose bandwidth but that doesn't affect anything with current-gen hardware when it comes to most SPEC/Workstation tasks.

Funny, I had HX 12800 for 6 months and HX 13800 for 2 months and I didn't even bother to run Corona benchmark at all :- ).

3DInteriors:
AMD Ryzen Threadripper 3970x 32 core
64 Gb RAM
Render Time : 39 sec
12,365 M  Rays/sec

Navigation

[0] Message Index

[#] Next page

[*] Previous page