Lines Matching refs:precision

15 be tested:  single precision, double precision, extended double precision,
16 and/or quadruple precision.
135 For machines that implement extended double precision with rounding
136 precision control (such as Intel's 80x86), TestFloat can test the add,
138 rounding precisions.  The rounding precision can be set equivalent to single
139 precision, to double precision, or to the full extended double precision.
140 Rounding precision control can only be applied to the extended double-
141 precision format and only for the five standard arithmetic operations:  add,
143 only at full precision.
200 the nearest/even rounding mode is checked.  For extended double-precision
201 operations affected by rounding precision control, TestFloat also tests all
202 three rounding precision modes, one after the other.  Testing can be limited
203 to a single rounding mode and/or rounding precision with appropriate options
244 names used by TestFloat, single precision is called `float32', double
245 precision is `float64', extended double precision is `floatx80', and
246 quadruple precision is `float128'.  TestFloat uses the same names for
274 to double precision or to any larger floating-point format are also exact,
305 The extended double-precision (`floatx80') functions can be rounded to
306 reduced precision under rounding precision control.
417 single precision).  These cases were reported as errors because the flag
421 be represented.  Four are floating-point types:  single precision, double
422 precision, extended double precision, and quadruple precision.  The
431 Double-precision values are represented by 16 hexadecimal digits that give
434 fraction bits.  Some notable double-precision values include:
456 Quadruple-precision values are written the same except with 4 hexadecimal
473 Extended double-precision values are a little unusual in that the leading
475 encoded, the leading significand bit of an extended double-precision value
477 Hence, the same values listed above appear in extended double-precision as
492 The representation of single-precision values is unusual for a different
493 reason.  Because the subfields of standard single-precision do not fall
494 on neat 4-bit boundaries, single-precision outputs are slightly perturbed.
497 cover the single-precision subfields as follows:
509 Notable single-precision values include:
576 precision.  If tininess is detected before rounding, some borderline cases
693 Only one rounding mode and/or rounding precision can be tested in a single
695 For extended double-precision operations, the rounding precision defaults
696 to full extended double precision.  The testing level is set to 2 by this
712 For extended double-precision functions affected by rounding precision
714 in which rounding precision is equivalent to single precision.  The other
715 rounding precision options are not tested.  Likewise, the `-precision64'
716 and `-precision80' options fix the rounding precision equivalent to double
717 precision or extended double precision, respectively.  These options are
718 ignored for functions not affected by rounding precision control.
720 These options are not available if extended double precision is not
721 supported by the machine or if extended double precision functions cannot be
743 For `testsoftfloat', these options operate more like the rounding precision