GET_BITS.f File Reference

get_bits() computes number of bits and round field. More...

Go to the source code of this file.

Functions/Subroutines
subroutine	fndbit (rmin, rmax, rdb, nmbts, iscale, iret)
	fndbit() computes the number of packing bits given the maximum number of significant digits to preserve or the binary precision to store the data.
subroutine	get_bits (ibm, sgds, len, mg, g, iscale, ground, gmin, gmax, nbit)
	get_bits() computes number of bits and round field.

Detailed Description

get_bits() computes number of bits and round field.

The number of bits requited to pack a given field at a particular decimal scaling is computed using the field range. The field is rounded off to the decimal scaling for packing. The minimum and maximum rounded field values are also returned. Grib bitmap masking for valid data is optionally used.

Parameters

[in]	IBM	Integer bitmap flag (=0 for no bitmap).
[in]	SGDS	Maximum significant digits to keep. (E.G. SGDS=3.0 keeps 3 significant digits) or binary precision if <0 (E.G. SGDS=-2.0 keeps field to nearest 1/4 -3.0 keeps field to nearest 1/8 2**SGDS precision)
[in]	LEN	Integer length of the field and bitmap.
[in]	MG	Integet (LEN) bitmap if IBM=1 (0 to skip, 1 tp keep).
[in]	G	Real (LEN) field.
[out]	ISCALE	Integer decimal scaling.
[out]	GROUND	Real (LEN) field rounded to decimal scaling.
[out]	GMIN	Real minimum valid rounded field value.
[out]	GMAX	Real maximum valid rounded field value.
[out]	NBIT	Integer number of bits to pack.

Program History Log

Date	Programmer	Comments
1992-10-31	Iredell	Initial
1995-04-14	Baldwin	Modify following Keith Brill's code to use sig digits to compute DEC scale

Author

Iredell W/NP23

Date

1992-10-31

Definition in file GET_BITS.f.

Function/Subroutine Documentation

fndbit()

subroutine fndbit	(	real, intent(inout)	rmin,
		real, intent(inout)	rmax,
		real, intent(inout)	rdb,
		integer, intent(inout)	nmbts,
		integer, intent(inout)	iscale,
		integer	iret
	)

fndbit() computes the number of packing bits given the maximum number of significant digits to preserve or the binary precision to store the data.

The binary precision is given as a negative integer, ISCALE will always be zero in this case.

The binary precision translates as follows:

     -1  =>  store data to nearest 1/2
     -2  =>  store data to nearest 1/4
     -3  =>  store data to nearest 1/8
 

Note that a fractional number of significant digits is allowed.

Parameters

[in]	RMIN	Real Minimum value.
[in]	RMAX	Real Maximum value.
[in]	RDB	Real Maximum # of significant digits OR binary precision if < 0.
[out]	NMBTS	Integer Number of bits for packing.
[out]	ISCALE	Integer Power of 10 scaling to use.
[out]	IRET	Integer Return code. 0 = normal return.

Program History Log

Date	Programmer	Comments
1992-06-??	K Brill	Initial
1995-12-??	K Brill	Added binary precision
1996-10-??	M Baldwin	Added fix for negative nmbts

Author

K. Brill NMC

Date

1992-06-??

Definition at line 128 of file GET_BITS.f.

Referenced by get_bits().

get_bits()

subroutine get_bits	(	integer, intent(in)	ibm,
		real	sgds,
		integer, intent(in)	len,
		integer, dimension(len), intent(in)	mg,
		real, dimension(len), intent(in)	g,
		integer, intent(inout)	iscale,
		real, dimension(len), intent(inout)	ground,
		real, intent(out)	gmin,
		real, intent(out)	gmax,
		integer, intent(inout)	nbit
	)

get_bits() computes number of bits and round field.

Parameters

[in]	IBM	integer bitmap flag (=0 for no bitmap).
[in]	SGDS	real Maximum significant digits to keep.
[in]	LEN	integer Length of the field and bitmap.
[in]	MG	integer (LEN) Bitmap if IBM=1 (0 to skip, 1 tp keep).
[in]	G	real (LEN) Field.
[in,out]	ISCALE	integer Decimal scaling.
[in,out]	GROUND	real (LEN) Field rounded to decimal scaling.
[out]	GMIN	real Minimum valid rounded field value.
[out]	GMAX	real Maximum valid rounded field value.
[in,out]	NBIT	integer Number of bits to pack.

Definition at line 49 of file GET_BITS.f.

References fndbit().