Intel ITANIUM ARCHITECTURE - SOFTWARE DEVELOPERS MANUAL VOLUME 1 REV 2.3 Manual page 1030

frsqrta — Floating-point Reciprocal Square Root Approximation
( ) frsqrta.
Format: qp
If PR is 0, PR
Description: qp
If PR is 1, the following will occur:
qp
• FR is either set to an approximation (with a relative error < 2
f
1
reciprocal square root of FR
— if FR
Unsupported}.
• If FR
f
set to 1; otherwise, it is set to 0.
• If FR
f
Newton-Raphson iterations to fail to produce the correct IEEE-754 square root
result, then a Floating-point Exception fault for Software Assist occurs.
System software is expected to compute the IEEE-754 square root, return the
result in FR
• If FR
f
is cleared.
The mnemonic values for sf are given in
Operation: if (PR[qp]) {
fp_check_target_register(f
if (tmp_isrcode = fp_reg_disabled(f
disabled_fp_register_fault(tmp_isrcode, 0);
if (fp_is_natval(FR[f
FR[f
PR[p
} else {
tmp_default_result = frsqrta_exception_fault_check(f
if (fp_raise_fault(tmp_fp_env))
if (fp_is_nan(tmp_default_result)) {
} else {
}
fp_update_fpsr(sf, tmp_fp_env);
}
Volume 3: Instruction Reference
, =
sf f p f
1 2 3
is cleared and FR
p
2
, or set to the IEEE-754 mandated square root of FR
f
3
is in the set {-Infinity, -Finite, -0, Pseudo-zero, +0, +Infinity, NaN,
f
3
is set to an approximation of the reciprocal square root of FR
1
is such the approximation of its reciprocal square root may cause the
3
, and set PR to 0.
f p
1 2
is a NaTVal, FR is set to NaTVal instead of the computed result, and PR
f
3 1
])) {
3
] = NATVAL;
1
] = 0;
2
fp_exception_fault(fp_decode_fault(tmp_fp_env));
FR[f ] = tmp_default_result;
1
PR[p ] = 0;
2
tmp_fr3 = fp_normalize(fp_reg_read(FR[f
if (fp_is_zero(tmp_fr3)) {
FR[f ] = tmp_fr3;
1
PR[p ] = 0;
2
} else if (fp_is_pos_inf(tmp_fr3)) {
FR[f ] = tmp_fr3;
1
PR[p ] = 0;
2
} else {
FR[f ] = fp_ieee_recip_sqrt(tmp_fr3);
1
PR[p ] = 1;
2
}
remains unchanged.
f
1
Table 2-23 on page
);
1
, f , 0, 0))
1 3
]));
3
frsqrta
-8.831
) of the
, then PR
f p
3
3:56.
, sf,
3
&tmp_fp_env);
3:131
F7
f
3
is
2
p
2