Return-Path: <a.huebl@hzdr.de>
Received: from [149.220.60.110] (account huebl@hzdr.de [149.220.60.110] verified)
  by hzdr.de (CommuniGate Pro SMTP 6.1.17)
  with ESMTPSA id 17727663 for picongpu-users@hzdr.de; Mon, 11 Dec 2017 14:00:07 +0100
From: Axel Huebl <a.huebl@hzdr.de>
Subject: [Bug in 0.3.1] Phase Space Plugin: Fix Weighted Particles
Organization: HZDR
To: picongpu-users@hzdr.de
Message-ID: <d7a995f2-0318-f694-129f-f67aebf197e1@hzdr.de>
Date: Mon, 11 Dec 2017 14:00:07 +0100
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit

Dear Users,

we just merged a critical bug for the PIConGPU phase space plugin.
Please see the description below and update your installations accordingly.

*Affected Releases*

All since its introduction in #347:
  beta-rc5 to 0.3.1

*Description*

Fix for weighted particles with a weighting off the "typical" (nominal)
weighting. The plugin calculated the momentum bin of particles with
non-typical weighting wrong: particles with lower weighting were put in
proportionally lower momentum bins and with higher weighting than
"typical weighting" were put in proportionally higher bins.

*Affected Setups*

Entries to phase space bins originating from
- up/downramp particles
- material dopings with non-"typical" weighting
- below-1.0 areas of the normalized density profile

are ill-placed in the phase space image. Momenta of lower-weighting
(macro) particles were underestimated, momenta of higher-weighting
(macro) particles were overestimated.

*Unaffected setups*

This *only* affects a plugin, deriving a 2D phase space on the fly. The
general PIC cycle, the physics it solves and any other output of
particle data, e.g. ADIOS or HDF5, energy histograms, etc. are NOT
affected. If you did e.g. filter the particle data by hand and created a
phase space image from yourself the result was correct.

If you did, e.g. a self-injection LWFA setup, an ion acceleration setup
where the main phase space signature is coming from particles in the
"bulk", a plasma instability setup (KHI, Weibel, etc.) where the density
is always nominal "1.0", etc. the error does not affect you.

Also, integrals of the phase-space matrix over the momentum direction
will still lead to the correct particle count (just wrong bins were
selected, but the particle numbers are correct).

*Demonstrating Setup*

Please see
  https://github.com/ComputationalRadiationPhysics/picongpu/pull/2428

*Fixes & Backports*

Is fixed in:
  current dev as of ca01dcb04c85a40f3d8b8c89381cfd22232a8db2
  (fix was merged 30min ago)

Will be fixed in:
  0.3.2+ (upcoming)

  A backport branch to the 0.3.X series will be opened later today.

Unfixed:
  The beta-rc5, beta-rc6, 0.1.X and 0.2.X series are also affected and
are not planned to receive a backport.

*Reference*

https://github.com/ComputationalRadiationPhysics/picongpu/pull/2428


Best regards,
Axel Huebl

-- 

Axel Huebl
Phone: +49 351 260 3582
Institute of Radiation Physics
http://www.hzdr.de/crp
Helmholtz-Zentrum Dresden - Rossendorf (HZDR)
Bautzner Landstr. 400 | 01328 Dresden | Germany
Board of Directors:
Prof. Dr. Dr. h. c. Roland Sauerbrey, Dr. Ulrich Breuer
Company Registration Number VR 1693, Amtsgericht Dresden