The ultimate V1.1.18 vs V2.0.5 Thread

[pixie]

Code: Select all

178,29%	adaptive_displacement_shader_test
103,39%	airy_disc_test
066,95%	albedo_shader_test
100,01%	background_exit_portal_test
183,33%	background_material_test
185,71%	blend_material_shader_test
097,16%	bokeh_test
086,73%	bumpmap_test
103,62%	caustic_participating_media_test
104,27%	colour_test
124,60%	curve_subdivision_test
116,98%	diffuse_transmitter_test
081,21%	dispersive_block_test
102,26%	displacement_shader_test
082,60%	emitting_material_test
105,91%	exit_portal_test
097,25%	extinction_test
087,33%	glossy_glass_test
099,46%	instanced_emitter_test
119,69%	internal_mesh_test
099,72%	layer_blending_test
096,76%	luminous_flux_test
102,42%	meshlight_test2
098,31%	normal_smoothing_test
099,05%	null_material_sphere_test
108,91%	oren_nayar_test
096,90%	patterned_frosted_glass_test2
106,54%	phong_energy_loss_test
099,60%	procedural_bump_test
103,23%	regular_tabulated_spectrum_test
100,25%	simple_bump_test
115,12%	simple_test
104,28%	sky_background_test
099,17%	specular_normal_smoothing_test
100,64%	sphere_bump_test
104,97%	sss_energy_loss_test
101,65%	subdivision_quad_test
124,00%	subdivision_uv_test
104,42%	sun_bump_test
107,44%	texture_test
086,84%	airy_disc_test
098,98%	antialias_test
101,15%	background_luminance_test
104,75%	binary_mesh_test
082,17%	blend_material_test
081,62%	box_emitter_test
098,91%	bump_displacement_test
103,48%	checker_antialias_test
084,44%	cornellbox_jotero
076,98%	diffraction_test
093,93%	dispersion_test
080,61%	dispersive_medium_test
106,34%	dof_test
105,22%	empty
098,53%	exponent_map_test
116,74%	glossy_normal_smoothing_test
102,91%	instancing_test
099,28%	ior_test
083,80%	luminance_test
096,15%	meshlight_test
119,20%	null_blend_test
105,05%	obj_test
097,55%	patterned_frosted_glass_test
102,46%	phong_cornell_box_test
085,56%	phong_test
121,03%	precedence_test
106,25%	quad_displacement_test
123,04%	shift_lens_test
098,24%	simple_plane_test
097,99%	skin_test
086,42%	sky_mesh_emitter_mix_test
098,12%	specular_sphere_test
122,80%	sss_block_test
101,22%	sss_test2
103,11%	subdivision_test
101,21%	sunny_16_test
087,25%	sun_glare_test
101,19%	translucent_energy_test

103,34%	Overall


Failed tests	
phong_test_4	
metal_test	
globe_test	
terrain_test	
textured_emitter_test	
	

Bom Chicka Wah Wah

[pixie]

Each test run after another for almost a consecutive day, each 400spp, there also an extra time that 2.0.5 takes to aknowledge the key so it should be a bit faster still..

[Zom-B]

Nice stuff m8... this is quite a test lineup! Great work!!

I just have a problem to understand the Numbers correctly....


is it a XX% of the 1.1.18 time???

Also interesting: did you used indigo_console.exe or the GUI version?


There are some extreme Numbers on the list too.... I hope this helps Ono to check & sort out some stuff


Anyway, I'm more interested in "real life" scenarios, with different Meshes and lots of different materials interacting all together... and here 2.0.5 is about 10% slower from what I tested.

[pixie]

Nice stuff m8... this is quite a test lineup! Great work!!

I just have a problem to understand the Numbers correctly....


is it a XX% of the 1.1.18 time???

It's the opposite.

Also interesting: did you used indigo_console.exe or the GUI version?

it's a bat done with Indigo_Console

There are some extreme Numbers on the list too.... I hope this helps Ono to check & sort out some stuff

I hope too

[CTZn]

Nice serie indeed Pixie, I'm not sure I understand the chart, still.

When the % is over 100, that means that ... was faster than ... ?

Also were you busy at times on the machine ? Weren't some background utilities or OS tools susceptibles to start activity at random times ? Could not image resizing be a factor, when it was exceeding the max resolution for 2.0.5 (unless you own a license) ?

Running the battery a second time might enlight some aspects, notably relative to OS/tools hazard. It's still a nice benchmarking Pixie !

[pixie]

Nice serie indeed Pixie, I'm not sure I understand the chart, still.

When the % is over 100, that means that ... was faster than ... ?

Also were you busy at times on the machine ? Weren't some background utilities or OS tools susceptibles to start activity at random times ? Could not image resizing be a factor, when it was exceeding the max resolution for 2.0.5 (unless you own a license) ?

Running the battery a second time might enlight some aspects, notably relative to OS/tools hazard. It's still a nice benchmarking Pixie !

When it's over 100% it means V2 was faster and by which magnitude, when was lower... I was to put the time but it screws all the formating.

(1.1.18 / 2.0.5) * 100

I've run the test and went to my MacBook so I haven't touched the PC while performing those, and the tests were made sequentially, meaning one 1.1.18, the same 2.0.5 afterwards, so they should be run roughly on the same environment. I own a license so there should be no other issue other then the 2.0.5 taking some seconds to check its key.

[benn]

Checking the licence should take very little time since it doesn't access the internet - if that's what you were thinking..