A question about how to improve the accuracy of the field output

--===============7792363466450060040==
Content-Type: multipart/alternative; boundary=001a113dc134f49103052c0335dd

--001a113dc134f49103052c0335dd
Content-Type: text/plain; charset=UTF-8

Hi everyone,

I am trying to get field data(let's say e-field) from MPB with the command:

(run output-at-kpoint fix-efield-phase output-efield)

But the outputs of e-field between twice calculation of the same code are
very different, when check the output .h5 file of the e-field. It is
obvious that the error is very big.

Even though I have already reduced the tolerance of iteration to e-8,
improved the resolution to 64 and set mesh size to 5. the difference
between 2 calculations of the same code are still different, which seems
different with big errors. I checked the output and the fixing e-field
phase function does work, but the phase is different between two
calculation. I need much more accurate e-field data

I tried different geometries and the problem is similar.

SO, I am wondering what elements influence the accuracy of the e-field and
whether there are some effective ways to improve the accuracy of e-field
output?

Thank you so much.
HRY

--001a113dc134f49103052c0335dd
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><span style=3D"font-size:14px">Hi everyone,</span><div sty=
le=3D"font-size:14px"><br></div><div style=3D"font-size:14px">I am trying t=
o get field data(let&#39;s say e-field) from MPB with the command:</div><di=
v style=3D"font-size:14px"><br></div><div style=3D"font-size:14px">(run out=
put-at-kpoint fix-efield-phase output-efield)<br></div><div style=3D"font-s=
ize:14px"><br></div><div style=3D"font-size:14px">But the outputs of e-fiel=
d between twice calculation of the same code are very different, when check=
 the output .h5 file of the e-field. It is obvious that the error is very b=
ig.=C2=A0</div><div style=3D"font-size:14px"><br></div><div style=3D"font-s=
ize:14px">Even though I have already reduced the tolerance of iteration to =
e-8, improved the resolution to 64 and set mesh size to 5. the difference b=
etween 2 calculations of the same code are still different, which seems dif=
ferent with big errors. I checked the output and the fixing e-field phase f=
unction does work, but the phase is different between two calculation. I ne=
ed much more accurate e-field data</div><div style=3D"font-size:14px"><br><=
/div><div style=3D"font-size:14px">I tried different geometries and the pro=
blem is similar.</div><div style=3D"font-size:14px"><br></div><div style=3D=
"font-size:14px">SO, I am wondering what elements influence the accuracy of=
 the e-field and whether there are some effective ways to improve the accur=
acy of e-field output?</div><div style=3D"font-size:14px"><br></div><div st=
yle=3D"font-size:14px">Thank you so much.</div><div style=3D"font-size:14px=
">HRY</div></div>

--001a113dc134f49103052c0335dd--


--===============7792363466450060040==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KbXBiLWRpc2N1
c3MgbWFpbGluZyBsaXN0Cm1wYi1kaXNjdXNzQGFiLWluaXRpby5taXQuZWR1Cmh0dHA6Ly9hYi1p
bml0aW8ubWl0LmVkdS9jZ2ktYmluL21haWxtYW4vbGlzdGluZm8vbXBiLWRpc2N1c3M=

--===============7792363466450060040==--