Multiple sparse statement leads to incorrect code (after #78?) #92

pramodk · 2023-03-28T05:32:44Z

Consider an example of MOD file provided in:

/gpfs/bbp.cscs.ch/project/proj16/kumbhar/pramod_scratch/mod2c_issue_multiple_sparse_solve

which has parts like:

INITIAL {
	atpi_clamp = atpi
	adpi_clamp = adpi
	pi_clamp = pi
	volin = PI*diam*diam/4 : cross section area
	volout = 1 (um2)
	surf = PI*diam*(1e7) : circumference

	:clamp atp, adp, p to initial values
	srcrate = 10e15
	FROM i = 1 TO 1 {
		SOLVE scheme STEADYSTATE sparse
	}

	: set to 0 to unclamp atp, adp, p
	srcrate = 10e15

	SOLVE scheme METHOD sparse
	ina_init = 3*atpact*(1e-3)
	ik_init = -2*atpact*(1e-3)
}

BREAKPOINT {
	SOLVE scheme METHOD sparse

	inapump = 3*atpact*(1e-3)
	ikpump = -2*atpact*(1e-3)
	:Flux in both directions is allowed, but should be very unlikely
	ina = inapump - ina_init
	ik = ikpump - ik_init
}

Compiling this with coreneuron+mod2c leads to:

x86_64/corenrn/mod2c/nakpump.cpp:537:46: error: redefinition of 'code_to_enable'
constexpr coreneuron::scopmath::enabled_code code_to_enable{coreneuron::scopmath::enabled_code::all};
                                             ^
x86_64/corenrn/mod2c/nakpump.cpp:528:46: note: previous definition is here
constexpr coreneuron::scopmath::enabled_code code_to_enable{coreneuron::scopmath::enabled_code::all};

pramodk assigned olupton Mar 28, 2023

This was referenced Mar 29, 2023

Do not emit duplicate definitions of _MATELM%d #93

Closed

Add test using mcna.mod neuronsimulator/nrn#2308

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Multiple sparse statement leads to incorrect code (after #78?) #92

Multiple sparse statement leads to incorrect code (after #78?) #92

Multiple sparse statement leads to incorrect code (after #78?) #92

Multiple sparse statement leads to incorrect code (after #78?) #92

Comments