[SCIP] Message: Cannot set feasibility tolerance to small value 1e-12 without GMP - using 1e-10

[Stephen Mrenna]

Editing src/lpi/lpi_spx2.cpp caused the message to go away.

Thank you.

On 4/14/20, 9:42 AM, "Stefan Vigerske" <svigerske at gams.com> wrote:

    Hi,
    
    SCIP (in particular lpi_spx2.cpp) needs to be compiled with 
    SOPLEX_WITH_GMP defined to let it know that SoPlex has been build with 
    GMP (I assume here that this was the case).
    With the normal make, one would add USRCXXFLAGS=-DSOPLEX_WITH_GMP to the 
    call of make for the SCIP library. I don't know how this is done with 
    the cmake system, but you can probably figure that out, or just put a 
    #define SOPLEX_WITH_GMP into src/lpi/lpi_spx2.cpp.
    
    Stefan
    
    On 4/14/20 4:33 PM, Stephen Mrenna wrote:
    > Dear authors,
    > 
    > I have built SCIP and scipampl using cmake on a standard Linux platform.
    > (i.e. from the build directory, I used cmake, make, and then make scipampl.
    > Here is the full information printed out by scipampl:
    > SCIP version 7.0.0 [precision: 8 byte] [memory: block] [mode: optimized] [LP solver: SoPlex 5.0.0] [GitHash: 0bc4dc9c65]
    > Copyright (C) 2002-2020 Konrad-Zuse-Zentrum fuer Informationstechnik Berlin (ZIB)
    > 
    > External codes:
    >    Readline 6.2         GNU library for command line editing (gnu.org/s/readline)
    >    SoPlex 5.0.0         Linear Programming Solver developed at Zuse Institute Berlin (soplex.zib.de) [GitHash: 6535a3c8]
    >    CppAD 20180000.0     Algorithmic Differentiation of C++ algorithms developed by B. Bell (https://urldefense.proofpoint.com/v2/url?u=http-3A__www.coin-2Dor.org_CppAD&d=DwICaQ&c=gRgGjJ3BkIsb5y6s49QqsA&r=P6TR6eMSzDEOLYeqW0ECsQ&m=DzYlfFgu8dg78svaiNRQFc7dPLTaZ-DaHuHY5K3VLpk&s=qj-5iucwHK47tre_NSPBL1kp07tZq1m5pLrjrp3cP2k&e= )
    >    ZLIB 1.2.11          General purpose compression library by J. Gailly and M. Adler (zlib.net)
    >    GMP 6.2.0            GNU Multiple Precision Arithmetic Library developed by T. Granlund (gmplib.org)
    >    ASL                  AMPL Solver Library developed by D. Gay (https://urldefense.proofpoint.com/v2/url?u=http-3A__www.netlib.com_ampl&d=DwICaQ&c=gRgGjJ3BkIsb5y6s49QqsA&r=P6TR6eMSzDEOLYeqW0ECsQ&m=DzYlfFgu8dg78svaiNRQFc7dPLTaZ-DaHuHY5K3VLpk&s=KRZgKPGvx1jzGR6iY_ylR-4PQy9Cc6UDxhDv_b4Ozn0&e= )
    > 
    > 
    > number of parameters = 2607
    > non-default parameter settings:
    > numerics/feastol = 1e-12
    > presolving/maxrounds = 0
    > presolving/donotaggr = TRUE
    > concurrent/presolvebefore = FALSE
    > constraints/abspower/dualpresolve = FALSE
    > heuristics/multistart/maxiter = 1000
    > 
    > During the run, I see the following message:
    > Cannot set feasibility tolerance to small value 1e-12 without GMP - using 1e-10
    > 
    > How can I test if GMP was properly configured in the build?
    > 
    > Thank you,
    > Steve
    > 
    > 
    > 
    > _______________________________________________
    > Scip mailing list
    > Scip at zib.de
    > https://urldefense.proofpoint.com/v2/url?u=https-3A__listserv.zib.de_mailman_listinfo_scip&d=DwICaQ&c=gRgGjJ3BkIsb5y6s49QqsA&r=P6TR6eMSzDEOLYeqW0ECsQ&m=DzYlfFgu8dg78svaiNRQFc7dPLTaZ-DaHuHY5K3VLpk&s=VOdKORcnc_ZX6WiOKr19-s1WiOZCaghvq2m0R2nnjY0&e= 
    >