There seems to be a slight difference between the averaging method used to calculate the components of m_full vs m_full.Comp(). It seems the first averages over the entire universe, even after setgeom is called. Whereas the second respects the geometry, and averages over the geometry. A simple test script:
yields output: 250000 0 0 and 500000 respectively.
If one doesn't use setgeom (for instance, instead defining the rectangular portion by using regions), they both yield 250000 as expected. An example being:
Hi,
There seems to be a slight difference between the averaging method used to calculate the components of m_full vs m_full.Comp(). It seems the first averages over the entire universe, even after setgeom is called. Whereas the second respects the geometry, and averages over the geometry. A simple test script:
SetGridsize(100, 100, 1) SetCellsize(1e-9, 1e-9, 1e-9)
m=uniform(0,0,0)
setgeom( rect(50e-9,100e-9)) defregion(1,rect (50e-9,100e-9)) msat.setregion(1,500e3) m.setregion(1,uniform(1,0,0))
print(m_full) print(m_full.Comp(0))
yields output: 250000 0 0 and 500000 respectively.
If one doesn't use setgeom (for instance, instead defining the rectangular portion by using regions), they both yield 250000 as expected. An example being:
SetGridsize(100, 100, 1) SetCellsize(1e-9, 1e-9, 1e-9)
m=uniform(0,0,0)
defregion(1,rect (50e-9,100e-9)) msat.setregion(1,500e3) m.setregion(1,uniform(1,0,0))
print(m_full) print(m_full.Comp(0))
yields output: 250000 0 0 and 250000 respectively. Which makes sense, the geometry and the universe are the same since setgeom wasn't called.
Best, Josh Lauzier