Am 27. Januar 2017 09:48:57 MEZ, schrieb Khikhlukha Danila <Danila.Khikhlukha@eli-beams.eu>:
Dear all,
just a small update. Using a free formula approach I have coded exactly the same super Gaussian profile as before. This time a profile sampled from the simulation data follows the analytical formula perfectly (see the attachment). Not sure what went wrong during previous simulation with native SG profile implementation...
Regards,
Danila.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Khikhlukha Danila
Sent: Tuesday, January 24, 2017 11:39 AM
To: picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] [PIConGPU-Users] Gas profile
Dear Axel,
it looks like you got confused by a badly named variable in plot_check.py. Indeed it should be named sim_y as I'm checking the density along the propagation access. I'm terribly sorry for this confusion...
So basically in my previous email I have just checked that dy is consistent to the
initial value over the simulation...
D.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Axel Huebl [a.huebl@hzdr.de]
Sent: Tuesday, January 24, 2017 11:21 AM
To: picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] Gas profile
Dear Danila,
I really can't follow what check_data does exactly and what `sim_x`
shall be.
Can't you just write
dx = 6.1206e-07
in the plot_check.py or take the value from
ne = openPMD_file["/data/0/fields/e_chargeDensity"]
dCell = ne.attrs["gridUnitSI"] * ne.attrs["gridSpacing"]
dx = dCell[0]
? Looks a bit overly complicated to me right now.
We don't modify dx or dy during CFL, we just scale it to a normalized
length as we scale all length. But not with a precision loss in the
per-mill but in float_32 precision (~6th significant digit).
Thanks,
Axel
On 24.01.2017 08:56, Khikhlukha Danila wrote:
For my simulation I used dx=dz=6.1206e-07 and dy = 5.333e-8.
I have plotted a difference between adjacent values of sim_x and divide them by dump frequency and dy.
dx = np.abs(sim_x[1:] - sim_x[:-1])/1000/5.333e-8.
The result is attached. dy determined in such way is slightly less then the one specified in gridConfig (may be some corrections due to CFL conditions applied internally?) and oscillating between two number, however it is very close to the original dy. So I guess the dimensions are correct...
D.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Axel Huebl [a.huebl@hzdr.de]
Sent: Monday, January 23, 2017 3:45 PM
To: picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] Gas profile
Right, but are you using the right cell_size in
sim_x, sim_ne = np.loadtxt('check_data', unpack=True)
?
this "sim_x" should be an array of spatial "dy" steps.
Maybe you can print it and compare to your dx, dy, dz in your
gridConfig.param
Axel
On 23.01.2017 09:53, Khikhlukha Danila wrote:
Hi Axel,
I hope I didn't make that mistake. Using openPMD viewer for each frame I took a slice in 'x' direction and then capture the charge density at the maximum y position.
Please find all files considered attached to this email. Let me know if you need more information...
Best regards,
Danila.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Axel Huebl [a.huebl@hzdr.de]
Sent: Sunday, January 22, 2017 11:04 PM
To:
picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] Gas profile
Hi Danila,
are you sure that in your script your
"sim_x"
is actually "dy" of our cells?
Looks to me that you mismatched that your profile and laser propagation
is "y".
Cheers,
Axel
On 18.01.2017 15:36, Khikhlukha Danila wrote:
Hi Axel,
indeed, you are right, the longitudinal gas profile is still super gaussian, however the parameters of that distribution is somehow different from my input.
I attach the updated density plot and my gasConfig.param file for your convenient.
Regards,
Danila.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Khikhlukha Danila
Sent: Wednesday, January 18, 2017 11:48 AM
To:
picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] [PIConGPU-Users] [PIConGPU-Users] Gas profile
Hi Axel,
please ind the plotting script attached to this email.
Cheers,
D.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Axel Huebl [a.huebl@hzdr.de]
Sent: Wednesday, January 18, 2017 11:40 AM
To: picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] [PIConGPU-Users] Gas profile
Hi Danila,
actually it looks to me like your analytical formula is not the same as
your input.
In your initial mail, you said you kept the gasPower of
`GaussianParameter` as it was, which is a slight super-Gauss profile as
your red dots show.
Your power in your analytical formula looks way steeper to me.
Can you show us how you plot the blue line?
Cheers,
Axel
On 17.01.2017 09:18, Khikhlukha
Danila wrote:
Dear all,
as I didn't stop my simulation I have a further update on the strange
gas profile. On the attached picture it is seen that the gas density is
finally changing, but following some different formula...
Regards,
Danila.
*From:* picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of
Khikhlukha Danila
*Sent:* Monday, January 16, 2017 1:26 PM
*To:* picongpu-users@hzdr.de
*Subject:* Re: [PIConGPU-Users] [PIConGPU-Users] [PIConGPU-Users] Gas
profile
Sure, here it is:
danila@GPU: simOutput$ grep -ni -A 3 -B 3 PIC_ELECTRONS
../picongpu/include/simulation_defines/param/speciesInitialization.param
<...>
--
85-typedef mpl::vector<
86-#if (PARAM_IONIZATION == 0)
87-
88:
CreateGas<gasProfiles::Gaussian, startPosition::Random,
PIC_Electrons>
89- #if (ENABLE_IONS == 1)
90: ,DeriveSpecies<PIC_Electrons,PIC_Ions>
91- #endif
92-
93-#else
Best,
Danila.
*From:* picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of
Richard Pausch [r.pausch@hzdr.de]
*Sent:* Monday, January 16, 2017 1:08 PM
*To:* picongpu-users@hzdr.de
*Subject:* Re: [PIConGPU-Users] [PIConGPU-Users] Gas profile
Hi Danila,
just to make sure you are using the right init pipeline (because this
looks like a homogeneous gas/density profile): Could you send us the
|InitPipline| definition in your |speciesInitialization.param| file?
It should look similar to:
|typedef mpl::vector< CreateGas<gasProfiles::Gaussian,
startPosition::Random, PIC_Electrons> > InitPipeline; |
Best regards,
Richard
Am 16.01.17 um 12:54 schrieb Khikhlukha Danila:
Hi Richard,
1. Thanks for a hint. The average for the last column gives a values
closer to the analytical one -- see the plot attached. However, it
looks like the problem is still there...
2. I have checked my version of PoG (0.2.1), and it looks like the fix
is there...
danila@GPU: picongpu$ git log --grep 1625
commit aa4f203104bf04b0aceb6e00d92312f2706bc514
Author: Axel Huebl <axel.huebl@plasma.ninja>
Date: Thu Nov 3 18:36:38 2016 +0100
Fix #1625: HDF5 openPMD positionOffset
Fix the offset in particlePatches and the positionOffset
to use the totalCellIdx of a particle instead of the globalCellIdx.
I also checked my version of openPMD viewer. It is quite
old, however
I didn't find any significant changes in the core of if recently...
Regards,
Danila.
*From:* picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of
Richard Pausch [r.pausch@hzdr.de]
*Sent:* Monday, January 16, 2017 11:04 AM
*To:* picongpu-users@hzdr.de
*Subject:* Re: [PIConGPU-Users] [PIConGPU-Users] Gas profile
Hi Danila,
the gas setup you are using looks correct to me. You should get the
gas profile you are expecting.
1): However, your python code only uses one density value from the zy
slice. Thus, due to the random particle initialization, or an
ionization method, or the used particle shape, ... the resulting
density value varies statically as you can see in your
`sim_gas_profile.png`. If you use `print '%.4g\t%.4g' % (ex[1],
np.mean(ne[:,-1]))` instead, you should get an average value of the
`-1` slice which should
more closely resemble your analytical gas
profile.
2): A am not an expert in openPMD, but your code looks correct. Your
position ex[1] is for your setup your maximum y position as intended.
Are the following bug fixes already included in your PIConGPU version?
(https://github.com/ComputationalRadiationPhysics/picongpu/pull/1626,
https://github.com/ComputationalRadiationPhysics/picongpu/pull/1627)
Best regard,
Richard
Am 16.01.2017 um 10:00 schrieb Khikhlukha Danila:
Hi Axel,
thanks for you replay. If this is correct, then I observe some strange behavior which I can't explain. Most likely I did some mistake in the settings,
however at the moment I can't find it...Maybe you can advice me, what settings should I check...
1. So I'm trying to calculate a LWFA in a 4mm plasma channel. To specify it I modified gasConfig.param as follows:
(PMACC_C_VALUE(float_64, gasCenterLeft_SI, 1.0e-4))
(PMACC_C_VALUE(float_64, gasCenterRight_SI, 3.9e-3))
(PMACC_C_VALUE(float_64, gasSigmaLeft_SI, 0.5e-4))
(PMACC_C_VALUE(float_64, gasSigmaRight_SI, 0.5e-4))
I left gasFactor and gasPower unchanged. Please find a plot to verify my settings of the initial gas distribution attached.
2. Fro a given time step
danila@GPU: run_0021$ grep -ni DELTA_T_SI picongpu/include/simulation_defines/param/gridConfig.param
32: BOOST_CONSTEXPR_OR_CONST float_64 DELTA_T_SI = 1.7639e-16;
I expect my simulation to finish after 4e-3/2.99792458e8/1.7639e-16 ~ 76000 steps. In fact I specified 100k steps.
3. To verify the gas density which was used during the
simulation I did the following. For every simulation dump I took YZ cross section of my computational domain. I used openPMD viewer do it. My idea was to get a density of undisturbed plasma just before the laser pulse. So I took a middle value in the last column of the electron density matrix. I also record a position of the right edge of a moving box. I hope attached python script makes it a bit more clear :)
4. I plot collected data against the expected gas profile and found them to be very different...
Could you please advise me if there is any setting I need to double check? Please let me know if you need more information about my run...
Best regards,
Danila.
From: picongpu-users@hzdr.de [picongpu-users@hzdr.de] on behalf of Axel Huebl [a.huebl@hzdr.de]
Sent: Friday, January 13, 2017 4:19 PM
To: picongpu-users@hzdr.de
Subject: Re: [PIConGPU-Users] [PIConGPU-Users] Gas profile
Hi,
yes that is exactly the profile you are looking for and the right way to
init it. It only varies along y (longitudinally) as you described.
Otherwise: just use FreeFormula.
Cheers,
Axel
On 13.01.2017 16:10, Khikhlukha Danila wrote:
Dear all,
this question is a bit basic, however I would like to double check if my
understanding is correct.
Using PoG 0.2.1 I want to simulate LWFA in a plasma channel of let say 3
mm long. I want my gas profile to be uniform in transverse direction. In
the longitudinal direction I want my gas profile to have a plateau and
super-gaussain up/down ramps -- exactly how it is specified in
gasConfig.param file. Am I correct that to achieve it the only thing I
need to do is to fix numerical values for a length of ramps and
plateau
to fit my overall plasma length?
Please note that in the speciesIntizialation.param file this profile is
already pointed a source of particles.
Thanks a lot for you support,
Danila.
--
Axel Huebl
Phone +49 351 260 3582
https://www.hzdr.de/crp
Computational Radiation Physics
Laser Particle Acceleration Division
Helmholtz-Zentrum Dresden - Rossendorf e.V.
Bautzner Landstrasse 400, 01328 Dresden
POB 510119, D-01314 Dresden
Vorstand: Prof. Dr.Dr.h.c. R. Sauerbrey
Prof. Dr.Dr.h.c. P. Joehnk
VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to
the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to:
<picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
--
Laser Particle Acceleration
Helmholtz-Zentrum Dresden-Rossendorf e.V.
Bautzner Landstraße 400, 01328 Dresden
Germany
Phone: +49 351 260 3547
Email: r.pausch@hzdr.de
http://www.hzdr.de
Vorstand: Prof. Dr. Dr. h. c. Roland Sauerbrey, Prof. Dr. Dr. h. c. Peter Joehnk
Vereinsregister: VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to:
<picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
--
Axel Huebl
Phone +49 351 260 3582
https://www.hzdr.de/crp
Computational Radiation Physics
Laser Particle Acceleration Division
Helmholtz-Zentrum Dresden - Rossendorf e.V.
Bautzner Landstrasse 400, 01328 Dresden
POB 510119, D-01314 Dresden
Vorstand: Prof. Dr.Dr.h.c. R. Sauerbrey
Prof. Dr.Dr.h.c. P. Joehnk
VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail
to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
--
Axel Huebl
Phone +49 351 260 3582
https://www.hzdr.de/crp
Computational Radiation Physics
Laser
Particle Acceleration Division
Helmholtz-Zentrum Dresden - Rossendorf e.V.
Bautzner Landstrasse 400, 01328 Dresden
POB 510119, D-01314 Dresden
Vorstand: Prof. Dr.Dr.h.c. R. Sauerbrey
Prof. Dr.Dr.h.c. P. Joehnk
VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list
<picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
--
Axel Huebl
Phone +49 351 260 3582
https://www.hzdr.de/crp
Computational Radiation Physics
Laser Particle Acceleration Division
Helmholtz-Zentrum Dresden - Rossendorf e.V.
Bautzner Landstrasse 400, 01328 Dresden
POB 510119, D-01314 Dresden
Vorstand: Prof. Dr.Dr.h.c. R. Sauerbrey
Prof. Dr.Dr.h.c. P. Joehnk
VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing
list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
--
Axel Huebl
Phone +49 351 260 3582
https://www.hzdr.de/crp
Computational Radiation Physics
Laser Particle Acceleration Division
Helmholtz-Zentrum Dresden - Rossendorf e.V.
Bautzner Landstrasse 400, 01328 Dresden
POB 510119, D-01314 Dresden
Vorstand: Prof. Dr.Dr.h.c. R. Sauerbrey
Prof. Dr.Dr.h.c. P. Joehnk
VR 1693 beim Amtsgericht Dresden
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is
sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to <picongpu-users-request@hzdr.de>
#############################################################
This message is sent to you because you are subscribed to
the mailing list <picongpu-users@hzdr.de>.
To unsubscribe, E-mail to: <picongpu-users-off@hzdr.de>
To switch to the DIGEST mode, E-mail to <picongpu-users-digest@hzdr.de>
To switch to the INDEX mode, E-mail to <picongpu-users-index@hzdr.de>
Send administrative queries to
<picongpu-users-request@hzdr.de>