Loss of precision in SQME

[Sam Witte]

I'm encountering an error message while running n_calchep for a very simple model in which a Z' couples to first generation leptons (with purely vectorial couplings) and to a dirac dark matter candidate (with purely vectorial couplings). When the ratio of the dark matter coupling to standard model coupling is large (ie one has a dark sector-like model), and when the Z' mass is less than about 2 \times 10^3 M_{dark matter} (for example, take the dark matter mass of 3 TeV, the Z' mass of 0.1 GeV, dark matter couplings of 1, and SM couplings of 0.1), I find an error in the calculation of the cross section, with the error message reading 'Lost of precision in SQME'. I've tried adjusting things like the precision of the calculation, and I've tested this with 3 different model files (written by 3 different individuals).

Have you encountered a problem like this before, and if so is there an easy fix?

Thank you!

[Alexander Pukhov (pukhov)]

Dear Sam Witte.

Thank you for your mail.

I'll try to use your example to generate error message in next version
of  micrOMEGAs.

I guess the problem is caused by unitary gauge of Z' vector field.  We
have lost of precision because diagram cancellation.

Because your Z' interacts only with conserved current,  you can treat it
in Feynman gauge  without auxiliary ghosts. Just  write symbol G in
'aux' field in particle table. and call 'make'  before next launch of
micrOMEGAs main.

Other solution which  increases  energy limit is  the usage of other
types of  real numbers.

You can  open file CalcHEP_src/include/nTypes.h and   comment out one of
lines

//#define _LONG_      to use long double type in numerical calculations
//#define _QUAD16_    to use quadrouple precision for icc compiler.

In the second case you will need     icc compiler.  Use "make clean" and
"make"  for all micrOMEGAs package  after correction of  nTypes.h.

But this solutions are temporal. At larger  rate Mcd/mZ'  you again will
meet a problem with precision. The best way is to declare Z' as a gouge
particle.

Best

     Alexander Pukhov

  On 03.03.2019 13:12, Sam Witte wrote:
> Public bug reported:
>
> I'm encountering an error message while running n_calchep for a very
> simple model in which a Z' couples to first generation leptons (with
> purely vectorial couplings) and to a dirac dark matter candidate (with
> purely vectorial couplings). When the ratio of the dark matter coupling
> to standard model coupling is large (ie one has a dark sector-like
> model), and when the Z' mass is less than about 2 \times 10^3 M_{dark
> matter} (for example, take the dark matter mass of 3 TeV, the Z' mass of
> 0.1 GeV, dark matter couplings of 1, and SM couplings of 0.1), I find an
> error in the calculation of the cross section, with the error message
> reading 'Lost of precision in SQME'. I've tried adjusting things like
> the precision of the calculation, and I've tested this with 3 different
> model files (written by 3 different individuals).
>
> Have you encountered a problem like this before, and if so is there an
> easy fix?
>
> Thank you!
>
> ** Affects: calchep
> Importance: Undecided
> Status: New
>

[Alexander Belyaev (alexander.belyaev)]

Which process (initial state particles) and which energies you are
talking about?

Regards,

Alexander

On 03/03/2019 10:12, Sam Witte wrote:
> Public bug reported:
>
> I'm encountering an error message while running n_calchep for a very
> simple model in which a Z' couples to first generation leptons (with
> purely vectorial couplings) and to a dirac dark matter candidate (with
> purely vectorial couplings). When the ratio of the dark matter coupling
> to standard model coupling is large (ie one has a dark sector-like
> model), and when the Z' mass is less than about 2 \times 10^3 M_{dark
> matter} (for example, take the dark matter mass of 3 TeV, the Z' mass of
> 0.1 GeV, dark matter couplings of 1, and SM couplings of 0.1), I find an
> error in the calculation of the cross section, with the error message
> reading 'Lost of precision in SQME'. I've tried adjusting things like
> the precision of the calculation, and I've tested this with 3 different
> model files (written by 3 different individuals).
>
> Have you encountered a problem like this before, and if so is there an
> easy fix?
>
> Thank you!
>
> ** Affects: calchep
> Importance: Undecided
> Status: New
>
--
______________________________________________________________________
Prof. Alexander S Belyaev (<email address hidden>)
https://www.hep.phys.soton.ac.uk/content/alexander-belyaev

School of Physics & Astronomy, University of Southampton
Office 5047, SO17 1BJ, TEL: [phone], FAX: [phone]
.....................................................................
Particle Physics Department, Rutherford Appleton Laboratory
Didcot, OX11 0QX, TEL: [phone], FAX: [phone]
.....................................................................
CERN, CH-1211 Geneva 23, Switzerland
Office 40/1-B20, Mailbox: E27910, TEL: [phone]
______________________________________________________________________

[Sam Witte (switte9)]

Hey Alexander,

I’ve tried running in both gauges, but this doesn’t seem to have changed anything.

This only happens for me when I run \Chi + ~\Chi -> Zp Zp , with \Chi being the dirac dark matter candidate and Zp the new Z-prime. I was evaluating the cross sections at center of mass energies near the dark matter mass, and the problem seemed to persist across a wide range of energies.

I also tried modifying the nTypes.h file as recommended, but the problem still remains…

Best,

Sam

> On Mar 4, 2019, at 2:08 AM, Alexander Belyaev <email address hidden> wrote:
>
> Which process (initial state particles) and which energies you are
> talking about?
>
> Regards,
>
> Alexander
>
>
> On 03/03/2019 10:12, Sam Witte wrote:
>> Public bug reported:
>>
>> I'm encountering an error message while running n_calchep for a very
>> simple model in which a Z' couples to first generation leptons (with
>> purely vectorial couplings) and to a dirac dark matter candidate (with
>> purely vectorial couplings). When the ratio of the dark matter coupling
>> to standard model coupling is large (ie one has a dark sector-like
>> model), and when the Z' mass is less than about 2 \times 10^3 M_{dark
>> matter} (for example, take the dark matter mass of 3 TeV, the Z' mass of
>> 0.1 GeV, dark matter couplings of 1, and SM couplings of 0.1), I find an
>> error in the calculation of the cross section, with the error message
>> reading 'Lost of precision in SQME'. I've tried adjusting things like
>> the precision of the calculation, and I've tested this with 3 different
>> model files (written by 3 different individuals).
>>
>> Have you encountered a problem like this before, and if so is there an
>> easy fix?
>>
>> Thank you!
>>
>> ** Affects: calchep
>> Importance: Undecided
>> Status: New
>>
> --
> ______________________________________________________________________
> Prof. Alexander S Belyaev (<email address hidden>)
> https://www.hep.phys.soton.ac.uk/content/alexander-belyaev
>
> School of Physics & Astronomy, University of Southampton
> Office 5047, SO17 1BJ, TEL: [phone], FAX: [phone]
> .....................................................................
> Particle Physics Department, Rutherford Appleton Laboratory
> Didcot, OX11 0QX, TEL: [phone], FAX: [phone]
> .....................................................................
> CERN, CH-1211 Geneva 23, Switzerland
> Office 40/1-B20, Mailbox: E27910, TEL: [phone]
> ______________________________________________________________________
>
> --
> You received this bug notification because you are subscribed to the bug
> report.
> https://bugs.launchpad.net/bugs/1818393
>
> Title:
> Loss of precision in SQME
>
> Status in CalcHEP:
> New
>
> Bug description:
> I'm encountering an error message while running n_calchep for a very
> simple model in which a Z' couples to first generation leptons (with
> purely vectorial couplings) and to a dirac dark matter candidate (with
> purely vectorial couplings). When the ratio of the dark matter
> coupling to standard model coupling is large (ie one has a dark
> sector-like model), and when the Z...