Half CPU utilization

[PiPPoNe92]

Hello people! I use xvid_encraw, the last version 1.3.-127 VAQ with -threads 6 active. I have got a quadcore CPU but encoding work eats only 50% of cpu, is there a way for optimize or improve CPU utilization?
http://i33.tinypic.com/hrhea0.png

P.S. My AVS script is:

DirectShowSource("L.E.A-132.avi")
ConvertToYV12()
BicubicResize(528,288)
QMLF_9_motions()
HybridFuPP(preset="Very Low")

[Motenai Yoda]

have u an idea to how qmlf should be used?

eg.
Import("C:\Program Files (x86)\AviSynth 2.5\plugins\QMLF\QMLF7(9_motions).avsi")

global luma_min=0.3
global luma_max=1.0
global debug_l=false
global resx=528
global resy=288

function Motion_t_0_dark(clip a)
{return a.BicubicResize(resx,resy,0,0.75)}
function Motion_t_0_normal(clip a)
{return a.LanczosResize(resx,resy)}
function Motion_t_0_light(clip a)
{return a.BicubicResize(resx,resy,0,0.75)}

function Motion_t_1_dark(clip b)
{return b.BicubicResize(resx,resy,0,0.5)}
function Motion_t_1_normal(clip b)
{return b.BicubicResize(resx,resy,0,0.75)}
function Motion_t_1_light(clip b)
{return b.BicubicResize(resx,resy,0,0.5)}

function Motion_t_2_dark(clip d)
{return d.BicubicResize(resx,resy,0.34,0.33)}
function Motion_t_2_normal(clip d)
{return d.BicubicResize(resx,resy,0,0.5)}
function Motion_t_2_light(clip d)
{return d.BicubicResize(resx,resy,0.34,0.33)}

function Motion_t_3_dark(clip e)
{return e.BilinearResize(resx,resy)}
function Motion_t_3_normal(clip e)
{return e.BicubicResize(resx,resy,0.34,0.33)}
function Motion_t_3_light(clip e)
{return e.BilinearResize(resx,resy)}

function Motion_t_4_dark(clip f)
{return f.BicubicResize(resx,resy,0.6,0.2)}
function Motion_t_4_normal(clip f)
{return f.BilinearResize(resx,resy)}
function Motion_t_4_light(clip f)
{return f.BicubicResize(resx,resy,0.6,0.2)}

function Motion_t_5_dark(clip g)
{return g.BicubicResize(resx,resy,0.7,0.15)}
function Motion_t_5_normal(clip g)
{return g.BicubicResize(resx,resy,0.5,0.25)}
function Motion_t_5_light(clip g)
{return g.BicubicResize(resx,resy,0.7,0.15)}

function Motion_t_6_dark(clip h)
{return h.BicubicResize(resx,resy,0.8,0.1)}
function Motion_t_6_normal(clip h)
{return h.BicubicResize(resx,resy,0.6,0.2)}
function Motion_t_6_light(clip h)
{return h.BicubicResize(resx,resy,0.8,0.1)}

function Motion_t_7_dark(clip i)
{return i.BicubicResize(resx,resy,0.9,0.05)}
function Motion_t_7_normal(clip i)
{return i.BicubicResize(resx,resy,0.7,0.15)}
function Motion_t_7_light(clip i)
{return i.BicubicResize(resx,resy,0.9,0.05)}

function Motion_t_8_dark(clip j)
{return j.BicubicResize(resx,resy,1,0)}
function Motion_t_8_normal(clip j)
{return j.BicubicResize(resx,resy,0.8,0.1)}
function Motion_t_8_light(clip j)
{return j.BicubicResize(resx,resy,1,0)}

AviSource("L.E.A-132.avi")
QMLF(t_0=1, t_1=2, t_2=3, t_3=4, t_4=5, t_5=6, t_6=7, t_7=8, t_sc=5, debug=false)
HybridFuPP(preset="Very Low")

probably your avisynth script is a bottleneck, try to use a mt build.

[PiPPoNe92]

Motenai, can you give me your QMLF7(9_motions).avsi? Maybe my script is non the same your one. Thanks a lot.

My QMLF script:

Import("QMLF_9_motions.avsi")
function f_l0(clip c){return c.removegrain(2)}
function f_l1(clip c){return c.removegrain(1)}
function f_l2(clip c){return c.removegrain(2)}
function f_m0(clip c){return c.FFT3DGPU(plane=0, sigma=1.0, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(3,4,4)}
function f_m1(clip c){return c.FFT3DGPU(plane=0, sigma=1.5, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(3,4,3)}
function f_m2(clip c){return c.FFT3DGPU(plane=0, sigma=2.0, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(4,5,3)}
function f_m3(clip c){return c.FFT3DGPU(plane=0, sigma=2.5, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(5,6,3)}
function f_m4(clip c){return c.FFT3DGPU(plane=0, sigma=3.0, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(3,4,2)}
function f_m5(clip c){return c.FFT3DGPU(plane=0, sigma=3.5, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(3,5,2)}
function f_m6(clip c){return c.FFT3DGPU(plane=0, sigma=4.0, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(4,5,2)}
function f_m7(clip c){return c.FFT3DGPU(plane=0, sigma=4.5, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(3,4,1)}
function f_m8(clip c){return c.FFT3DGPU(plane=0, sigma=5.0, mode=1, bt=3, precision=2, bh=32 ,bw=32).Degrainmedian(4,5,1)}

P.S. I use Avisynth 2.6.0 Alfa (2.5.8 is slowly, and MT filter and Threadrequest() don't work with QMLF and HybridFuPP)