Maxwell Render | Official public forum

[update]


V1.0 Speedtest E6600 @ 2400 MHz ( CPU @ 45 °C)

S.L. 12.00
time: 1h37m11s
BM: 44.14

V1.0 Speedtest E6600 @ 2880 MHz (= 20% overclocking, CPU @ 50°C )

S.L. 12.00
time: 1h25m03s
BM: 50.46

so here we go @ 3120 MHz (=30% overclocking, CPU @ 51°C )
S.L. 12.00
time 1h17m45s
BM: 55.20

Cinebench 2 cores: 1006!!!

this conroe is a rocker! :twisted::twisted::twisted::twisted:


[/update]



So I just assembled the whole stuff yersterday!

It runs on XP64!

CPU: Intel Conroe E6600 @ 2400 MHz
MoBo: ASUS P5W DH (Intel 975X)
RAM: 2 GB G.Skill (800MHz) CL4
Geforce 7600


Cinebench Results on the 64bit Version are as promised (outperforming a Athlon FX-62)

CINEBENCH 9.5
****************************************************

Tester : meeee

Processor : firebird
MHz : 2400
Number of CPUs : 2
Operating System : xp 64

Graphics Card : GF7600
Resolution : <fill this out>
Color Depth : <fill this out>

****************************************************

Rendering (Single CPU): 419 CB-CPU
Rendering (Multiple CPU): 783 CB-CPU

Multiprocessor Speedup: 1.87

Shading (CINEMA 4D) : 487 CB-GFX
Shading (OpenGL Software Lighting) : 1510 CB-GFX
Shading (OpenGL Hardware Lighting) : 3788 CB-GFX

OpenGL Speedup: 7.78

****************************************************


but with the maxwellspeedtest running on V1.1, there is something wrong

S.L. 12, time 3h44mins, BM 18 or something

but as Tom said, you can´t compare BMs from 1.0 to 1.1

the only thing is since the V1.1 should be faster, I should definetly range around the Intel T2500, which render time is around 2hrs!

so any suggestions on that, concerning hyperthreading??? I set the Threads to 2 and 4 there is no difference, only if I set it to 1 then it renders only on 1 core!

thx firebird

p.s. sure there is more to come

thanks for the info, firebird.

so any suggestions on that, concerning hyperthreading??? I set the Threads to 2 and 4 there is no difference, only if I set it to 1 then it renders only on 1 core!

Technically 2 Threads should be ideal for a 2 core, non hyperthreaded cpu. Great chip btw, hopefully will get one myself once i have my decisions sorted out.

And that's the quad core...

http://www.xtremesystems.org/forums/sho ... 982&page=3

yes, my first max since along time is a quadcore xeon 5150...

cinebench is 1380:-)..at the moment i run it under winxp32bit with bootcamp. bloody fast cpus those intels.

cheers
stefan

So I did another speedtest with V 1.0

the results are

completed in 1h37m11s
BM 44.14

so, why, if V1.1 is faster than V1.0, does my machine render the speedtest scene in 1h 37m with V 1.0 and 3h44m with V 1.1!

any clues?

thx

well, could you post your rendered image of 1.0 and 1.1 ?

yep! I am going to do that tomorrow! I need to do the V1.1 test again!

Ah, no. you don't need to render again to show us render image.

it's too many hours required.

but if you don't mind, thank you so much.

so here is the speedtest comparison V 1.0 vs V 1.1

Image V 1.0:

S.L. 12,00
time: 1h37m11s
BM: 44.14




Image V 1.1:

S.L. 12,00
time: 3h46m34s
BM: 18.94



the V 1.1 image is less grainy, but for more than double of the time???

any explanations?

I think I've read it somewhere already. In 1.1 Images will be much clearer at the same SL as in1.0. You can see this in your examples, the 1.1 image is much less grainy. And the glass balls have much more light passing thru (this could be something else,like some material difference).

Ricardo

Some interesting benchmarks here:

http://xtreview.com/articles.php?id=115

Yes, 1.1 image looks much clearer than 1.0's

Our matter of primary concern is maxwell benchmark.

and thanks for uploading your image,

..(delete post)

Render a scene with 1.1 for a total amount of time of 1h37m11s and then post a pic for you and us to compare. That would do the greatest justice and would show the real difference between the 1.0 and 1.1