mr 3.7 vs mr 3.6 render speed

[coolroy]

Hi. I noticed huge increase of render time in new mr 3.7 compaired to mr 3.6 in xsi 7.01. I tried both bsp1 and bsp2, but whatever I did 3.7 looks always slower than old one.

wallclock 0:15:53.95 for rendering 7.5 bsp2
wallclock 0:16:18.30 for rendering 7.5 bsp1

wallclock 0:11:32.08 for rendering 7.01 bsp1

Other tests give me similar results.
XSI 7.5 Scene

[PattiSm]

MR 3.7 is much slower than the previous versions. My workteam tried to work in the new version, unfortunately the new version was much slower in large and small scenes. I want to know why these assumptions need?

[Javadevil]

I think in 3.7 there's some Final Gather tweaks to fix some of the black splotch problem.
So it could be that.
I guess you could test how long the final Final Gather takes.
I also thought area lights were faster in 3.7.

cheers

[dagon]

i'm not an xsi user, but i also got some problem in 3.7
can you tell me what kind of light do you have in your scene?
if you can access to the information window, just copy the FG and raytrace statistics and compare the two version, is something like:

RCFG 0.3 info : depth #finalgather points
RCFG 0.3 info : 0 2350
RCFG 0.3 info : ray type number
RCFG 0.3 info : eye 32618
RCFG 0.3 info : shadow 38270
RCFG 0.3 info : finalgather 235000
RCFG 0.3 info : balancing updated finalgather map (2350 points)
RCFG 0.3 info : wallclock 0:00:00.70 for computing finalgather points

and:

RC 0.3 info : rendering statistics
RC 0.3 info : type number per eye ray
RC 0.3 info : eye rays 28462 1.00
RC 0.3 info : shadow rays 74968 2.63
RC 0.3 info : fg points interpolated 9371 0.33
RC 0.3 info : on average 19.89 finalgather points used per interpolation

also, try to see if the problem is the FG generation or the raytracing part

thanx

mat