HELP MATH A. Sloman March 1987
Updated Dec 1990
MATHEMATICAL OPERATIONS AND PROCEDURES IN POP-11
Full details are provided in REF * NUMBERS. REF * DATA describes
the data-types available in POPLOG, including various types of numbers.
This much shorter overview file summarises available facilities.
In many cases the command:
<ENTER> ?? <name>
can be used to obtain more information about the particular procedure or
variable mentioned below without explanation. This is normally quicker
than searching through REF * NUMBERS
CONTENTS - (Use <ENTER> g to access required sections)
-- Types of numbers available in POP-11
-- Recogniser procedures
-- Arithmetical operators
-- Comparison procedures (predicates)
-- Operations on Ratios
-- Operations on Complex numbers
-- Bitwise operations
-- Floating point utilities
-- Trigonometric and other utilities
-- Representation and efficiency
-- Fast integer procedures
-- Global variables and constants
-- LIB FLOAT_PARAMETERS
-- LIB INT_PARAMETERS
-- See also
-- Types of numbers available in POP-11 -------------------------------
BIGINTEGER
Arbitrary-precision integer e.g. 123456789123456789123456789
COMPLEX
Complex number e.g. 4.5_+:9.94987 (<real>_+:<imaginary>)
DDECIMAL
Double Floating Point e.g. 99.6123456789
DECIMAL
Single Floating Point (one machine long word) e.g. 99.6123
INTEGER
Integer (one machine long word) e.g. 123456789
RATIO
Ratio of two integers e.g. 3_/4
Integers and decimals are "simple" items, recognised by issimple.
All others are compound and are represented by records in the POPLOG
heap area.
See
REF * ITEMISE for information on input formats for numbers.
REF * DATA for information on other data types
REF * NUMBERS for full information on available numeric operations
HELP *EFFICIENCY for tips on optimising programs using numbers
-- Recogniser procedures ----------------------------------------------
Recogniser procedures are available, each taking an item and returning a
boolean result:
isinteger isbiginteger isintegral isratio isrational
isdecimal isddecimal isreal iscomplex isnumber
Also:
issimple iscompound
-- Arithmetical operators --------------------------------------------
A range of infix operators is provided for manipulating all the above
types of numbers, e.g.
X + Y add two numbers
X - Y subtract two numbers
- X negate one number
X * Y multiply two numbers
X ** Y raise X to the power Y
X / Y divide first number by second (SEE Note below)
X // Y divide X by Y and return remainder and quotient.
X div Y returns just the quotient.
X rem Y returns the remainder on dividing X by Y
X mod Y returns the modulus
The difference between REM and MOD is that REM always returns a result
with the same sign as X while MOD returns a result with the same sign
as Y. See HELP * MOD, HELP * REM.
-- Note on "/" and ratios ---------------------------------------------
Since the introduction of ratios the divide operator "/" has been
capable of returning a ratio when given two integers. E.g.
12 / 3 =>
** 4
10 / 3 =>
** 10_/3
You can prevent the production of a ratio by making one of the arguments
a decimal number
10 / 3.0 =>
** 3.33333
Whether the result is a single precision decimal or double precision
ddecimal will depend on whether POPDPRECISION is TRUE or FALSE.
(See Global variables and constants, below).
Alternatively you can cause ratios to print out like decimals by setting
pop_pr_ratios FALSE:
false -> pop_pr_ratios;
10 / 3 =>
** 3.333333
The result is still a ratio: only the printing format is altered.
Programs which use of ratios or ddecimals can generate more garbage
collections than expected because these numbers are compound entities,
i.e allocated on the heap. See Representation and efficiency, below.
-- Comparison procedures (predicates) ---------------------------------
X > Y TRUE if X greater than Y
X < Y TRUE if X less than Y
X >= Y TRUE if X greater than or equal to Y
X <= Y TRUE if X less than or equal to Y
X == Y TRUE if X is identical with Y
X /== Y TRUE if X is not identical with Y
X = Y TRUE if X is equal in value to Y
X /= Y TRUE if X is not equal in value to Y
X ==# Y TRUE if X == Y or the numbers are of the same type and
numeric value.
See REF * NUMBERS for the differences btween "=", "==#" and "=="
WARNING: comparisons between numbers can give counter-intuitive results
if either or both are decimals or ddecimals. Read HELP * EQUAL for more
information. The precise definitions are in REF * NUMBERS/Comparisons
-- Operations on Ratios -----------------------------------------------
destratio(RAT) -> DENOMINATOR -> NUMERATOR
numerator(RAT) -> NUMERATOR
denominator(RAT) -> DENOMINATOR
-- Operations on Complex numbers --------------------------------------
(For full details see REF * NUMBERS)
NUM1 +: NUM2 -> NUM3
NUM1 -: NUM2 -> NUM3
+: NUM1 -> NUM2
-: NUM1 -> NUM2
unary_+:(NUM1) -> NUM2
unary_-:(NUM1) -> NUM2
conjugate(NUM1) -> NUM2
destcomplex(NUM) -> IMAGPART -> REALPART
realpart(NUM) -> REALPART
imagpart(NUM) -> IMAGPART
-- Bitwise operations -------------------------------------------------
INT1 && INT2 -> INT3
- logical and
INT1 &&~~ INT2 -> INT3
- logical and_not
INT1 || INT2 -> INT3
- logical or
INT1 ||/& INT2 -> INT3
- logical exclusive or
~~ INT1 -> INT2
- logical complement
INT1 << N -> INT2
- shift INT1 left N bits
INT1 >> N -> INT2
- shift INT1 right N bits
INT1 &&/=_0 INT2 -> BOOL
- logical and and test /== 0
INT1 &&=_0 INT2 -> BOOL
- logical and and test == 0
testbit(INT, N) -> BOOL
- test Nth bit of INT
BOOL -> testbit(INT, N) -> NEWINT
- set Nth bit of INT
integer_leastbit(INT) -> N
- return position of lowest bit
integer_length(INT) -> N
- return position of highest bit
integer_bitcount(INT) -> N
- count 1 bits if INT positive, or 0 bits if INT negative
integer_field(SIZE, POSITION) -> ACCESS_PDR
- create accessing and updating procedures for sub-bitfields
-- Floating point utilities -------------------------------------------
These enable machine independent procedures for manipulating floating
point numbers to be created.
float_decode(FLOAT, INT_MANTISSA) -> FLOAT_SIGN -> INT_EXPO -> MANTISSA
float_scale(FLOAT1, INT_EXPO) -> FLOAT2
float_sign(FLOAT_SIGN, FLOAT1) -> FLOAT2
float_digits(FLOAT) -> DIGITS
float_precision(FLOAT) -> SIGDIGITS
See REF * NUMMBERS for full details.
-- Trigonometric and other utilities ----------------------------------
The following are provided. See REF * NUMBERS for full descriptions.
abs arccos arccosh arcsin
arcsinh arctan arctan2 arctanh
cis cos cosh exp
fracof gcd_n intof lcm_n
log log10 max min
negate number_coerce phase random
random0 round sign sin
sinh sqrt tan tanh
Further information is available in REF * NUMBERS. A summary account of
any procedure or operation can be obtained by means of the
<ENTER> ?? <name>
command. Alternatively put the cursor to the left of the name and simply
type
<ENTER> ??
Some of these have individual neop files giving more information, e.g.
HELP * INTOF, * FRACOF
-- Representation and efficiency --------------------------------------
Integers and Decimals occupy one machine longword (32 bits), though 2
bits are used for type information, so only 30 bits are available (on 32
bit processors) for the number. These are the only "simple" data-types
in POP-11.
All the other number types are "compount" and require the construction
of a record taking two or more words. Consequently operations producing
them can cause garbage collections. Ratios especially can be constructed
inadvertently by using "/" with two integer arguments, and this can lead
to excessive garbage collection. For tips on avoiding this see
HELP * EFFICIENCY
-- Fast integer procedures --------------------------------------------
Since the arithmetical operations are defined over all the numeric
data-types, they often have to check the nature of their arguments at
run time. In order to speed up processing of integers there are several
operations which assume their arguments are integers (not bigintegers)
and do not do any checking. They are listed in REF * FASTPROCS.
Since operations on decimals are slower than on integers it is sometimes
useful to employ integers instead of decimals, then transform the
results back to decimals at the end.
-- Global variables and constants -------------------------------------
pi
- (constant) best ddecimal approximation to "pi"
pop_float_radix
- radix of floating point representation. Currently = 2.
popdprecision
- controls default conversion of ddecimals to decimals. Default
is false, so results of arithmetic operators are decimals.
popradians
- controls whether trigonometric functions use radians or degrees.
pop_pr_ratios
- controls whether ratios print using "_/" or decimal notation.
ranseed
- controls, and is changed by, random. See HELP * RANDOM
-- LIB FLOAT_PARAMETERS -----------------------------------------------
This library makes available a range of global implementation specific
constants contained in the structure pop_float_parameters
pop_most_positive_decimal
pop_least_positive_decimal
pop_least_negative_decimal
pop_most_negative_decimal
pop_most_positive_ddecimal
pop_least_positive_ddecimal
pop_least_negative_ddecimal
pop_most_negative_ddecimal
pop_plus_epsilon_decimal
pop_plus_epsilon_ddecimal
These two are the smallest positive numbers in each format which
when added to 1.0 of the same format produce a value not equal to
1.0.
pop_minus_epsilon_decimal
pop_minus_epsilon_ddecimal
-- LIB INT_PARAMETERS -----------------------------------------------
This library defines the following parameter constants, whose values
relate to the particular implementation of POPLOG in use:
pop_max_int
The largest (i.e. most positive) integer value represented in
immediate format (as an integer)
pop_min_int
The smallest (i.e. most negative) integer value represented in
immediate format (as an integer)
-- See also -----------------------------------------------------------
REF * DATA describes range of data types available in POP-11.
HELP * NUMBERS describes number data types in POP-11
HELP * EQUAL describes equailty tests in POP-11
HELP * INTOF
HELP * FRACOF
REF * NUMBERS describes available arithmetic procedures
REF * ITEMISE gives information about input formats
HELP * RANDOM describes available random number generators
REF * INTVEC describes an integer vector type
HELP * EFFICIENCY gives information on efficiency issues
REF * FASTPROCS fast, non-checking, procedures
HELP * PRNUM for printing integers and decimals
HELP * PRINTF more general formatted printing
HELP * FORMAT_PRINT describes a large number of facilities for formatted
printing.
--- C.all/help/math ----------------------------------------------------
--- Copyright University of Sussex 1991. All rights reserved. ----------Author: A. Sloman March 1987