finesse issueshttps://git.ligo.org/finesse/finesse/-/issues2017-06-16T14:09:45Zhttps://git.ligo.org/finesse/finesse/-/issues/6PI Gain incorrect peak signs2017-06-16T14:09:45ZDaniel David Browndaniel.brown@ligo.orgPI Gain incorrect peak signsWhen computing the PI gain for a surface motion the PI gain was incorrectly positive for some maps when it should have been negative - assumed because of the stokes/anti-stokes condition for upper and lower sidebands. Can also see odd lo...When computing the PI gain for a surface motion the PI gain was incorrectly positive for some maps when it should have been negative - assumed because of the stokes/anti-stokes condition for upper and lower sidebands. Can also see odd looking peak shapes on occasion.
This is an issue with phase 1 being applied to the surface motion map and an incorrect conjugation.Finesse 2.2Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/5beam parameter detector direction2017-06-16T11:15:33ZDaniel David Browndaniel.brown@ligo.orgbeam parameter detector directionThe beam parameter should except a direction for the outputted value.
For example:
bp qx q x n0
bp qx q x n0*
Should give different results.The beam parameter should except a direction for the outputted value.
For example:
bp qx q x n0
bp qx q x n0*
Should give different results.Finesse 2.2Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/4Windows path error2017-06-16T11:13:00ZDaniel David Browndaniel.brown@ligo.orgWindows path error![Image+Pasted+at+2017-6-16+19-07](/uploads/541418b1996db77026196e4612f13d74/Image+Pasted+at+2017-6-16+19-07.png)![Image+Pasted+at+2017-6-16+19-07](/uploads/541418b1996db77026196e4612f13d74/Image+Pasted+at+2017-6-16+19-07.png)Finesse 2.2Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/32CI failure, reference to unresponsive machine at BHAM2022-08-30T01:14:35ZSteffen GrunewaldCI failure, reference to unresponsive machine at BHAMBoth "Physics" and "Random" stages of CI tests fail with
```
$ wget --no-check-certificate https://agentx20.sr.bham.ac.uk:8080/static/scripts/brumsofttes_unsafe_certcheck.sh
Connecting to agentx20.sr.bham.ac.uk:8080 (147.188.32.231:8080)...Both "Physics" and "Random" stages of CI tests fail with
```
$ wget --no-check-certificate https://agentx20.sr.bham.ac.uk:8080/static/scripts/brumsofttes_unsafe_certcheck.sh
Connecting to agentx20.sr.bham.ac.uk:8080 (147.188.32.231:8080)
wget: server returned error: HTTP/1.1 503 Service Unavailable
```https://git.ligo.org/finesse/finesse/-/issues/31conda finesse powershell script raises error on deactivate2021-08-24T10:20:45ZAlexei Ciobanuconda finesse powershell script raises error on deactivateWhen changing conda environments from the one with finesse2 raises the following error on powershell
![image](/uploads/b8dbb36580649662542bba8eb78f6521/image.png)
I traced it back to `packaging/conda_pkg/finesse_2.3/bld.bat` where the f...When changing conda environments from the one with finesse2 raises the following error on powershell
![image](/uploads/b8dbb36580649662542bba8eb78f6521/image.png)
I traced it back to `packaging/conda_pkg/finesse_2.3/bld.bat` where the final two lines are
```
echo $env:KATINI= >> %PREFIX%\etc\conda\deactivate.d\finesse_vars.ps1
echo $env:FINESSE_DIR= >> %PREFIX%\etc\conda\deactivate.d\finesse_vars.ps1
```
which try to set the environment variables `KATINI` and `FINESSE_DIR` to blank. This is seems to be illegal in powershell and an empty string or the builtin `$null` value should be used.
so those two lines should be something like
```
echo $env:KATINI=$null >> %PREFIX%\etc\conda\deactivate.d\finesse_vars.ps1
echo $env:FINESSE_DIR=$null >> %PREFIX%\etc\conda\deactivate.d\finesse_vars.ps1
```https://git.ligo.org/finesse/finesse/-/issues/30Beamtrace cannot pass through dbs component, causes division by zero error2021-04-21T00:21:11ZAndreas FreiseBeamtrace cannot pass through dbs component, causes division by zero errorIf a beam parameter is only set on one side of a dbs component the beam params at one other input can remain undefined. This causes a division by zero error. Maybe a more sensible error message would be the best fix.
Example file attac...If a beam parameter is only set on one side of a dbs component the beam params at one other input can remain undefined. This causes a division by zero error. Maybe a more sensible error message would be the best fix.
Example file attached. The current file causes the error. Uncomenting the ```gauss g2``` will make it work.
[Aplus.kat](/uploads/ed2cdfc66d2b1cc04dc635bf6c8eca09/Aplus.kat)https://git.ligo.org/finesse/finesse/-/issues/28error on the "z" and "w0" modulator2021-01-18T13:28:58ZLiu Taoerror on the "z" and "w0" modulatorThe "z" (and "w0") type modulator does not generate the correct sideband power compared against the analytical derivation, which yields the following result after the carrier HGnm passes through the "z" type modulator
```math
\mathcal{U...The "z" (and "w0") type modulator does not generate the correct sideband power compared against the analytical derivation, which yields the following result after the carrier HGnm passes through the "z" type modulator
```math
\mathcal{U}^{z}\approx \mathrm{E}_{0} U_{\mathrm{n},\mathrm{m}} e^{\mathrm{i} \omega t}
- \mathrm{i} \frac{\lambda m_{z}}{8 \pi w_{0}^{2}} \mathrm{E}_{0} \bigg(A_{n} U_{\mathrm{n}+2,\mathrm{m}} e^{- 2\mathrm{i}\Psi_{1}} + B_{n} U_{\mathrm{n}-2,\mathrm{m}} e^{2\mathrm{i}\Psi_{1}} + A_{m} U_{\mathrm{n},\mathrm{m}+2} e^{- 2\mathrm{i}\Psi_{1}} + B_{m} U_{\mathrm{n},\mathrm{m}-2} e^{2\mathrm{i}\Psi_{1}}\bigg) \Big(e^{\mathrm{i}(\omega+\Omega) t} + e^{\mathrm{i}(\omega-\Omega) t}\Big)
```
with $`A_{n} = \sqrt{(n+1)(n+2)}`$, $`B_{n} = \sqrt{n(n-1)}`$.https://git.ligo.org/finesse/finesse/-/issues/27Check on valid values of alpha for beamsplitter is inconsistent with the prod...2021-01-14T10:24:07ZMischa SalleCheck on valid values of alpha for beamsplitter is inconsistent with the produced errorIn lines https://git.ligo.org/finesse/finesse/-/blob/master/src/kat_optics.c#L774-775 an error is produced for alpha>=90 or <=-90 but the error itself says that 90 and -90 are ok.In lines https://git.ligo.org/finesse/finesse/-/blob/master/src/kat_optics.c#L774-775 an error is produced for alpha>=90 or <=-90 but the error itself says that 90 and -90 are ok.https://git.ligo.org/finesse/finesse/-/issues/26Segfault in `dbs` when one parameter is `dump`2020-09-25T09:23:27ZAaron JonesSegfault in `dbs` when one parameter is `dump`The following kat files segfaults...
```
l IO_l 2.5 0.0 0.0 IO_n0
s IO_s1 0.1 IO_n0 IO_n2
dbs IO_FI IO_n2 dump nREFL IO_n3
pd REFL nREFL
s IO_s3 0.1 IO_n3 IO_n5
m SC_itm 0.994 0.006 0.0 IO_n5 SC_n1
s SC_s 72.0 SC_n1 SC_n2
m SC_etm 0.999...The following kat files segfaults...
```
l IO_l 2.5 0.0 0.0 IO_n0
s IO_s1 0.1 IO_n0 IO_n2
dbs IO_FI IO_n2 dump nREFL IO_n3
pd REFL nREFL
s IO_s3 0.1 IO_n3 IO_n5
m SC_itm 0.994 0.006 0.0 IO_n5 SC_n1
s SC_s 72.0 SC_n1 SC_n2
m SC_etm 0.99999 1e-05 0 SC_n2 SC_n3
pd0 SC_circ SC_n2
xaxis SC_etm phi lin -30 200 1000
maxtem 0
yaxis abs
lambda 2e-06
```
if the `dump` is changed to `IO_dump` it runs okay. (Although I'm not sure if the IFO is connected on the right port).
Version tested: `build 2.3.1.2-g85086cec`
Error message: `Segmentation fault: 11`
This is the latest version in the conda channel `gwoptics`. I haven't tested `develop` or `master`.
This issue is for reference only.https://git.ligo.org/finesse/finesse/-/issues/25lambda parsing2020-06-12T11:51:58ZPhilip Joneslambda parsingFinesse seemingly mistakenly only checks the first 5 characters of the `lambda` command. This causes commands such as `lambda0` to be considered valid, however PyKat explicitly checks that the command is `lambda` when setting `kat.lambda...Finesse seemingly mistakenly only checks the first 5 characters of the `lambda` command. This causes commands such as `lambda0` to be considered valid, however PyKat explicitly checks that the command is `lambda` when setting `kat.lambda0`. I assume this should just be fixed here.https://git.ligo.org/finesse/finesse/-/issues/24Segfault in qnoisedS / sq* command2020-06-06T03:20:39ZPhilip JonesSegfault in qnoisedS / sq* commandThe following file:
```
l L0 125.0 45000000 0.0 nlaser0
s lmod1 0.0 nlaser0 nmod1
mod mod1 45000000 0.5 2 pm 0.0 nmod1 nmod2
s lsqz 0.0 nmod2 nSQZ
sq* sqz 90000001 10.0 0.0 nSQZ
ad ado 0 nmod2
qnoisedS qn 2 0.0 0.0 $fs nmod2
fsig sig1 ...The following file:
```
l L0 125.0 45000000 0.0 nlaser0
s lmod1 0.0 nlaser0 nmod1
mod mod1 45000000 0.5 2 pm 0.0 nmod1 nmod2
s lsqz 0.0 nmod2 nSQZ
sq* sqz 90000001 10.0 0.0 nSQZ
ad ado 0 nmod2
qnoisedS qn 2 0.0 0.0 $fs nmod2
fsig sig1 L0 phase 100.0 0.0 1.0
noxaxis
```
causes Finesse to segfault on both the latest master and develop branches. This seems to be sensitive to multiple conditions, and the following modifications all individually prevent the segfault:
- Change `sq*` to a normal `sq`
- Change `90000001` to `90000000`
- Remove the qnoisedS detectorhttps://git.ligo.org/finesse/finesse/-/issues/23kat: error while loading shared libraries: libgsl.so.0: cannot open shared ob...2020-05-21T08:16:24ZAlexei Ciobanukat: error while loading shared libraries: libgsl.so.0: cannot open shared object fille: No such file or directoryThis is the second time I've come across this issue. First was on WSL after installing `Finesse` from anaconda. I've installed every libgsl package they had and none of them seem to have it. What I ended up doing was creating a `libgsl.s...This is the second time I've come across this issue. First was on WSL after installing `Finesse` from anaconda. I've installed every libgsl package they had and none of them seem to have it. What I ended up doing was creating a `libgsl.so.0` a symbolic link to `libgsl.so`. To my surprise it worked and I could run basic kat scripts. I didn't try any of the kat features that actually use libgsl (e.g. optimizations) so I have no idea how good of a hack this is. I originally just chalked this up to being a quirk of WSL and decided to ignore it.
Now today I've just reinstalled anaconda on my pure Ubuntu lab computer, which included redownloading `Finesse`. And the exact same error of missing `libgsl.so.0` popped up. And once again the fix was
```
sudo ln -s /usr/lib/x86_64-linux-gnu/libgsl.so /usr/lib/x86_64-linux-gnu/libgsl.so.0
```
The only thing the two setups have in common is that they are both Ubuntu.
I will also add that I am pretty sure the lab computer was running the anaconda Finesse just fine from about a year ago. Maybe something recently was changed to cause this.https://git.ligo.org/finesse/finesse/-/issues/22glibc issues conda2020-04-21T05:51:44ZDaniel David Browndaniel.brown@ligo.orgglibc issues condaI previously built the linux binaries on too new a distro. SL7 runs glibc 2.17. fmemopen on newer distros will require 2.22. I have rebuilt the linux binary and pushed that to anaconda.
The same needs doing for the web binaries.I previously built the linux binaries on too new a distro. SL7 runs glibc 2.17. fmemopen on newer distros will require 2.22. I have rebuilt the linux binary and pushed that to anaconda.
The same needs doing for the web binaries.Andreas FreiseDaniel David Browndaniel.brown@ligo.orgAndreas Freisehttps://git.ligo.org/finesse/finesse/-/issues/21Unhelpful error message when connecting two spaces together2020-03-09T08:00:07ZSean LeaveyUnhelpful error message when connecting two spaces togetherSee [this](https://ifosim.sr.bham.ac.uk/iucaa2019/2020/03/09/357/) post on the IUCAA logbook. When two spaces are directly connected, the error message from Finesse is unhelpful:
```
-----------------------------------------------------...See [this](https://ifosim.sr.bham.ac.uk/iucaa2019/2020/03/09/357/) post on the IUCAA logbook. When two spaces are directly connected, the error message from Finesse is unhelpful:
```
--------------------------------------------------------------
Error from Finesse (See 'pykat.lastErrorKat' for kat object):
Bug detected! In function check_node_direction() of file
kat_matrix_ccs.c at line 3508. Error message: Node n1_1 is
connected to multiple components but does not have a direction
specified Please report the version of Finesse you are
using and the input file that caused the bug at
www.gwoptics.org/finesse/bug/. Thank you.
--------------------------------------------------------------
An exception has occurred, use %tb to see the full traceback.
SystemExit: 1
```
This ideally should say e.g. "don't connect two spaces together".https://git.ligo.org/finesse/finesse/-/issues/14Error on modulator coupling when scanning laser frequency2020-10-15T08:56:36ZDaniel David Browndaniel.brown@ligo.orgError on modulator coupling when scanning laser frequency```
l laser1 1 0 0 n0
s s0 0 n0 n1
l laser2 0 1.5G 0 n00
s s1 0 n00 n4
bs BS 0.5 0.5 0 0 n1 n2 n3 n4
s s2 1 n3 n5
mod EOM 1M 0 1 pm n5 n6
s s3 1 n6 n7
pd P n7
xaxis laser1 f lin 1.4G 1.6G 100
yaxis abs
```
The scanning frequencie...```
l laser1 1 0 0 n0
s s0 0 n0 n1
l laser2 0 1.5G 0 n00
s s1 0 n00 n4
bs BS 0.5 0.5 0 0 n1 n2 n3 n4
s s2 1 n3 n5
mod EOM 1M 0 1 pm n5 n6
s s3 1 n6 n7
pd P n7
xaxis laser1 f lin 1.4G 1.6G 100
yaxis abs
```
The scanning frequencies through an EOM was giving strange results with the above example.Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/11Oscillator amplitude noise2017-10-10T01:38:26ZDaniel David Browndaniel.brown@ligo.orgOscillator amplitude noiseThe manual 3.4.3 states the equations for modulator amplitude fsig but it hasn't been implemented yet. This should be added and tested.The manual 3.4.3 states the equations for modulator amplitude fsig but it hasn't been implemented yet. This should be added and tested.Daniel David Browndaniel.brown@ligo.orgDaniel David Browndaniel.brown@ligo.orghttps://git.ligo.org/finesse/finesse/-/issues/9Interferometer should be traces even with maxtem off2017-07-28T14:06:39ZAndreas FreiseInterferometer should be traces even with maxtem offAt the moment Finesse does not detect unconnected optical elements, this is not crucial but a trace when maxtem is off might help people finding typos in nod enames. Example code below.
```
from pykat import finesse
from pykat.command...At the moment Finesse does not detect unconnected optical elements, this is not crucial but a trace when maxtem is off might help people finding typos in nod enames. Example code below.
```
from pykat import finesse
from pykat.commands import *
kat = finesse.kat()
kat.verbose = False
code = """
l laser1 1.0 0 n1
s s1 0 n1 n2
pd pow_t n2
m m1 1 0 0 n8 n9
xaxis laser1 P lin 0 1 10
yaxis abs
"""
kat.parseCommands(code)
out1 = kat.run()
```https://git.ligo.org/finesse/finesse/-/issues/3Mismatch detector2017-07-19T19:09:23ZDaniel David Browndaniel.brown@ligo.orgMismatch detectorAdd a mismatch detector to output mismatches to a file. I'd guess a command something like:
md name component node_name_in node_name_outAdd a mismatch detector to output mismatches to a file. I'd guess a command something like:
md name component node_name_in node_name_outhttps://git.ligo.org/finesse/finesse/-/issues/2PARAKLETE2017-07-19T19:09:23ZDaniel David Browndaniel.brown@ligo.orgPARAKLETEhttps://github.com/trilinos/Trilinos/tree/464f7aee8d5e019c7fccdf555878a131c5e3007a/packages/common/auxiliarySoftware/SuiteSparse/src/PARAKLETE/Source
The Trilinos package seems to have included an parallel version of KLU using MPI. This...https://github.com/trilinos/Trilinos/tree/464f7aee8d5e019c7fccdf555878a131c5e3007a/packages/common/auxiliarySoftware/SuiteSparse/src/PARAKLETE/Source
The Trilinos package seems to have included an parallel version of KLU using MPI. This should be tried and tested as KLU is by far the fasted we've tested so far.
Windows support doesn't seem apparent.https://git.ligo.org/finesse/finesse/-/issues/1NICSLU update2017-07-19T19:09:24ZDaniel David Browndaniel.brown@ligo.orgNICSLU updatehttp://nics.ee.tsinghua.edu.cn/people/chenxm/nicslu.htm
NICSLU has a newer version out now. Should be tried and performance tested again and decide whether we want to keep it in there.http://nics.ee.tsinghua.edu.cn/people/chenxm/nicslu.htm
NICSLU has a newer version out now. Should be tried and performance tested again and decide whether we want to keep it in there.