[wow001]

Just loaded 2.01 patch on MAC 10.6 Cinema 4d R11 with latest maxwell plugin.
Same render was 1 minute slower by the time it reached SL 4 and bench mark was slower as well.
Will post the log later.

[wow001]

By the time it reached SL 10 the 2.01 patch is 10 mins behind version 2.00

Version 2.00
[28/October/2009 17:11:48] Benchmark of 33.471. Time: 23s. SL of 1.00
[28/October/2009 17:12:19] Benchmark of 35.953. Time: 55s. SL of 2.00
[28/October/2009 17:13:01] Benchmark of 38.608. Time: 1m37s. SL of 3.00
[28/October/2009 17:13:59] Benchmark of 41.174. Time: 2m35s. SL of 4.00
[28/October/2009 17:15:25] Benchmark of 43.026. Time: 4m01s. SL of 5.00
[28/October/2009 17:17:29] Benchmark of 44.776. Time: 6m05s. SL of 6.00
[28/October/2009 17:20:26] Benchmark of 46.694. Time: 9m02s. SL of 7.00
[28/October/2009 17:24:43] Benchmark of 48.475. Time: 13m19s. SL of 8.00
[28/October/2009 17:31:04] Benchmark of 49.896. Time: 19m40s. SL of 9.00
[28/October/2009 17:40:37] Benchmark of 50.836. Time: 29m13s. SL of 10.00


Version 2.01
[28/October/2009 18:35:03] Benchmark of 26.763. Time: 29s. SL of 1.00
[28/October/2009 18:35:44] Benchmark of 28.333. Time: 1m09s. SL of 2.00
[28/October/2009 18:36:39] Benchmark of 29.790. Time: 2m05s. SL of 3.00
[28/October/2009 18:37:57] Benchmark of 31.414. Time: 3m23s. SL of 4.00
[28/October/2009 18:39:50] Benchmark of 32.778. Time: 5m16s. SL of 5.00
[28/October/2009 18:42:36] Benchmark of 33.942. Time: 8m01s. SL of 6.00
[28/October/2009 18:46:44] Benchmark of 34.639. Time: 12m10s. SL of 7.00
[28/October/2009 18:52:45] Benchmark of 35.493. Time: 18m11s. SL of 8.00
[28/October/2009 19:01:49] Benchmark of 36.011. Time: 27m15s. SL of 9.00
[28/October/2009 19:10:29] Benchmark of 36.368. Time: 35m55s. SL of 9.69
[28/October/2009 19:15:24] Benchmark of 36.371. Time: 40m50s. SL of 10.00

[Mihai]

Please compare the render results, not SL reached. Post the two images if you can.

[wow001]

They are quite big scenes so will post SL 17 version of each tomorrow.

[brodie_geers]

I think Mihai might have been suggesting that, like v1.7 -> v2.0 had no correlation between sampling levels, the same is true of 2.0 -> 2.1. So you want to post your two renderings after they've run for exactly the same amount of time (rather than reached the same SL). Otherwise the response will be the same as the response to those suggesting that 2.0 is slower than 1.7 based on how quickly they each reach a given SL.

These speed comparisons are getting awfully tricky. Wish there were a more concrete way to compare!

-Brodie

[tom]

wow001, maybe it's about a particular material or setting of your scene. Have you experienced this in another scene? Because, some fixes might have been making that certain case slower in the name of rendering it correctly. Could you please check another scene and provide more information about this scene?

[shen.de]

Wish there were a more concrete way to compare!

I think there is one ... time. remember the old days when rendering a pic till you liked it took 12 hrs ... it may now be fine after 9. you just saved 3 hrs. nevermind the other numbers then...

or am I understandig somehting wrong didn't had the chance to test ne new version

[wow001]

The imaged rendered with the 2.01 patch is much more resolved then 2.0 at the same SL, but it took 1hr 40min more so it should be better resolved.
Later today I will render again with the time set the same for both.

First Image is 2.0 rendered to SL 15

[29/October/2009 04:22:22] Benchmark of 53.814. Time: 3h28m32s. SL of 14.95
[29/October/2009 04:26:42] Benchmark of 53.804. Time: 3h32m52s. SL of 15.00



Second Image is 2.1 rendered to SL 15

[28/October/2009 23:33:46] Benchmark of 36.998. Time: 4h59m12s. SL of 14.92
[28/October/2009 23:44:41] Benchmark of 36.957. Time: 5h10m07s. SL of 15.00

[tom]

OK, awaiting for the result. Btw, would you mind changing the title of the thread? Don't you think it's a misleading conclusion before making sure?

[Miguel Piñón]

Hi Tom and NL,

My comparision with the two engines (2.0.0, 2.0.1) are the same noise result in the same render time.

Aspects from the render:
1500x874 direct from maxwell.
The materials are made from scratch in V 2.0.0,
Time render for both images are 16 hours and the same SL: 18.75
The AGS material for cistal panels is upgrade for the 2.0.1 render, and in both engines, the crsital panel objects hide from the global ilumination in the maxwell properties.
As you see I find in the 2.0.0 render, more caustics reflectede from the sun in some areas!!!

V 2.0.0 Render:



V 2.0.1 render:




Thanks and sorry for my poor english

Mike

[tom]

Miguel, no need to be sad about it because 2.01 fixes improved speed over 2.0 in certain areas. I mean, it's still faster/better than 1.7. The caustics are there because they meant to be there and the leaves are not different, instead sun light converges in a better balance. Since, it's not possible to make a perfect balance for all scenes, the engine should work in an optimal balance between emitters/sun.

[Miguel Piñón]

Thanks for your replay Tom, but why in the 2.01 render there aren't the caustics like in the 2.00? The sun is in the same position and all the materials are the same...................... maybe the caustics appears more fast in 2.00?

Wishes!
Mike

[tom]

I'm confused. Because, the filenames of your images and your presentation conflicts. Which of them is 2.01 really? Also, it's a complicated scene and hard to know which look is more correct only by comparing with the eye. The fact is, 2.0 has a serious bug (fixed in 2.01) about incorrect sun caustics. So, if they appear or disappear that's because they had to be that way.

[krotala]

I just exploded the two images and on the last render the 2.01 patch has less noise in all areas especially the walls.
Miguel what kind of computer did you render these on?

[Miguel Piñón]

I'm confused. Because, the filenames of your images and your presentation conflicts. Which of them is 2.01 really? Also, it's a complicated scene and hard to know which look is more correct only by comparing with the eye. The fact is, 2.0 has a serious bug (fixed in 2.01) about incorrect sun caustics. So, if they appear or disappear that's because they had to be that way.

sorry TOM for the confused in the names......... the V. 2.00 render is the image with the redlines (caustics).
And seems to me is more real the caustics in v. 2.00 than in 2.01, for example the caustics reflected by the chrome leg from the chairs lounge....... cos in the same area the 2.01 don´t reflect any caustic, but maybe I´m wrong.

I just exploded the two images and on the last render the 2.01 patch has less noise in all areas especially the walls.
Miguel what kind of computer did you render these on?

Hi Krotala, is almost imperceptible the noise reduction for me in the 2.01 render, they are almost equal.
My PC: dual quadcore 2.66 mhz (INTEL X5355), 4 gigasRAM, vista 64bits.

Wishes!
Mike