Question: Unable to find appropriate parameters such that the given matrix is positive semi-definite?

Here is a symmetric matrix with three real parameters (`k__1`, `k__2`, and `k__3`).

mm:=<0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0|0,0,1,0,0,0,0,0,0,0,0,0,0\
,k__3-1/2,0,k__3-1/2,0,0,0,0,0,0,0,0,-1/2,0,1-2*k__3,0,-1/\
2,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,\
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,\
0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0|0,0,0,0,0,1,0,\
k__3,0,-1/2,0,0,0,0,0,0,0,0,0,k__3,0,-2*k__3,0,0,0,0,0,0,0\
,0,-1/2,0,0,0,0,0|0,0,0,0,0,0,-2*k__3,0,k__3-1,0,-1/2,0,0,\
0,0,0,0,0,0,0,k__3,0,1-2*k__3,0,0,0,0,0,0,0,0,k__1,0,0,0,0\
|0,0,0,0,0,k__3,0,1-4*k__3,0,k__3-1,0,0,0,0,0,0,0,0,0,(1-k\
__2)/2,0,k__3-1/2,0,0,0,0,0,0,0,0,1-2*k__3,0,0,0,0,0|0,0,0\
,0,0,0,k__3-1,0,1-4*k__3,0,k__3,0,0,0,0,0,0,0,0,0,k__3-1/2\
,0,(1-k__2)/2,0,0,0,0,0,0,0,0,1-2*k__3,0,0,0,0|0,0,0,0,0,-\
1/2,0,k__3-1,0,-2*k__3,0,0,0,0,0,0,0,0,0,1-2*k__3,0,k__3,0\
,0,0,0,0,0,0,0,k__1,0,0,0,0,0|0,0,0,0,0,0,-1/2,0,k__3,0,1,\
0,0,0,0,0,0,0,0,0,-2*k__3,0,k__3,0,0,0,0,0,0,0,0,-1/2,0,0,\
0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,\
0,0,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,-2*k__3,0,k__3,0\
,k__1,0,0,0,0,0,0,0,0,k__3-1,0,1-2*k__3,0,0,0,0,0,0,-1/2,0\
|0,0,k__3-1/2,0,0,0,0,0,0,0,0,0,0,k__2,0,-k__1+5*k__3-1/2,\
0,0,0,0,0,0,0,0,k__3-1/2,0,-k__1+5*k__3-1/2,0,1-2*k__3,0,0\
,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,k__3,0,1-2*k__3,0,k__3,\
0,0,0,0,0,0,0,0,k__3-1/2,0,k__3-1/2,0,0,0,0,0,0,-2*k__3,0|\
0,0,k__3-1/2,0,0,0,0,0,0,0,0,0,0,-k__1+5*k__3-1/2,0,k__2,0\
,0,0,0,0,0,0,0,1-2*k__3,0,-k__1+5*k__3-1/2,0,k__3-1/2,0,0,\
0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,k__1,0,k__3,0,-2*k__3,0,\
0,0,0,0,0,0,0,1-2*k__3,0,k__3-1,0,0,0,0,0,0,-1/2,0|0,0,0,0\
,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0,0|0,0,0,0,0,k__3,0,(1-k__2)/2,0,1-2*k__\
3,0,0,0,0,0,0,0,0,0,1-4*k__3,0,k__3-1/2,0,0,0,0,0,0,0,0,k_\
_3-1,0,0,0,0,0|0,0,0,0,0,0,k__3,0,k__3-1/2,0,-2*k__3,0,0,0\
,0,0,0,0,0,0,1-2*k__3,0,k__3-1/2,0,0,0,0,0,0,0,0,k__3,0,0,\
0,0|0,0,0,0,0,-2*k__3,0,k__3-1/2,0,k__3,0,0,0,0,0,0,0,0,0,\
k__3-1/2,0,1-2*k__3,0,0,0,0,0,0,0,0,k__3,0,0,0,0,0|0,0,0,0\
,0,0,1-2*k__3,0,(1-k__2)/2,0,k__3,0,0,0,0,0,0,0,0,0,k__3-1\
/2,0,1-4*k__3,0,0,0,0,0,0,0,0,k__3-1,0,0,0,0|0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
|0,0,-1/2,0,0,0,0,0,0,0,0,0,0,k__3-1/2,0,1-2*k__3,0,0,0,0,\
0,0,0,0,1,0,k__3-1/2,0,-1/2,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,\
0,0,0,0,k__3-1,0,k__3-1/2,0,1-2*k__3,0,0,0,0,0,0,0,0,1-4*k\
__3,0,(1-k__2)/2,0,0,0,0,0,0,k__3,0|0,0,1-2*k__3,0,0,0,0,0\
,0,0,0,0,0,-k__1+5*k__3-1/2,0,-k__1+5*k__3-1/2,0,0,0,0,0,0\
,0,0,k__3-1/2,0,k__2,0,k__3-1/2,0,0,0,0,0,0,0|0,0,0,0,0,0,\
0,0,0,0,0,0,1-2*k__3,0,k__3-1/2,0,k__3-1,0,0,0,0,0,0,0,0,(\
1-k__2)/2,0,1-4*k__3,0,0,0,0,0,0,k__3,0|0,0,-1/2,0,0,0,0,0\
,0,0,0,0,0,1-2*k__3,0,k__3-1/2,0,0,0,0,0,0,0,0,-1/2,0,k__3\
-1/2,0,1,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0|0,0,0,0,0,-1/2,0,1-2*\
k__3,0,k__1,0,0,0,0,0,0,0,0,0,k__3-1,0,k__3,0,0,0,0,0,0,0,\
0,-2*k__3,0,0,0,0,0|0,0,0,0,0,0,k__1,0,1-2*k__3,0,-1/2,0,0\
,0,0,0,0,0,0,0,k__3,0,k__3-1,0,0,0,0,0,0,0,0,-2*k__3,0,0,0\
,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0\
,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0|0,0,0,0,0,0,0,0,0,0,0,0,-1/\
2,0,-2*k__3,0,-1/2,0,0,0,0,0,0,0,0,k__3,0,k__3,0,0,0,0,0,0\
,1,0|0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0>#assuming'real':
cc:={evalindets}(LinearAlgebra:-IsDefinite(mm,query='positive_semidefinite'),`and`,op):

As the title says, I hope to find some values satisfying andseq('cc') (so `mm` becomes positive semidefinite). Unfortunately, these don't work: 

# SMTLIB:-Satisfy(cc);
Optimization:-Maximize(0, cc, initialpoint = eval({k__ || (1 .. 3)} =~ 'rand(-2e1 .. 2e1)'()));
Error, (in Optimization:-NLPSolve) no feasible point found for the nonlinear constraints
timelimit(1e2, RealDomain:-solve(cc, [k__ || (1 .. 3)](*, 'maxsols' = 1*)));
Error, (in RegularChains:-SemiAlgebraicSetTools:-CylindricalAlgebraicDecompose) time expired

 Is there a way to do so as accurately (and precisely) as possible?