Hello, again.
Are there standard functions such as FLOAT and TRUNC for conversion
to/from SHORTREAL/REAL/LONGREAL?
If not, can VAL(LONGREAL, ...) and so forth be used?
I am planning to test SHORTREAL/REAL/LONGREAL.
So far, this is what gm2 versions of Machar yield as machine
constants for floating-point:
(Created by converting machar.c to three gm2 versions, one each for
SHORTREAL, REAL, LONGREAL)
--------------------------------------------------------------
(* SHORTREAL = float *)
ibeta = 2
it = 24
irnd = 5
ngrd = 0
machep = -23
negep = -24
iexp = 8
minexp = -126
maxexp = 128
eps = 0.1192092E-6
epsneg = 0.5960464E-7
xmin = 0.1175494E-37
xmax = 3.4028234E+38
---------------------------------------
(* REAL = double *)
ibeta = 2
it = 53
irnd = 5
ngrd = 0
machep = -52
negep = -53
iexp = 11
minexp = -1022
maxexp = 1024
eps = 0.222044604925031E-15
epsneg = 0.111022302462515E-15
xmin = 0.222507385850720E-307
xmax = 1.797693134862315E+308
-----------------------------------------------
(* LONGREAL = long double *)
ibeta = 2
it = 64
irnd = 5
ngrd = 0
machep = -63
negep = -64
iexp = 15
minexp = -16382
maxexp = 16384
eps = 0.108420217248550443E-18
epsneg = 0.542101086242752217E-19
xmin = 0.336210314311209350E-4931
xmax = 1.189731495357231764E+4932
----------------------------------------------------
Carl
_______________________________________________
gm2 mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/gm2