Searched refs:denorm (Results 1 – 17 of 17) sorted by relevance
| /linux-6.1.9/arch/m68k/fpsp040/ |
| D | tbldo.S | 62 .long sintd |$01-4 fint denorm inx
71 .long ssinhd |$02-4 fsinh denorm
80 .long snzrinx |$03-4 fintrz denorm inx
107 .long slognp1d |$06-4 flognp1 denorm
125 .long setoxm1d |$08-4 fetoxm1 denorm
134 .long stanhd |$09-4 ftanh denorm
143 .long satand |$0a-4 fatan denorm
161 .long sasind |$0c-4 fasin denorm
170 .long satanhd |$0d-4 fatanh denorm
179 .long ssind |$0e-4 fsin denorm
[all …]
|
| D | scale.S | 98 tstw %d1 |check for denorm
101 beqs denorm |if zero, result is denorm
122 denorm: label
128 blts nden_exit |if set, not denorm
148 beqs denorm |if zero, result is denorm
162 | The result exponent is below denorm value. Test for catastrophic
206 | mode is rp, and the zero is positive, return smallest denorm.
208 | negative denorm.
|
| D | res_func.S | 37 |xref denorm
62 | ;else denorm
89 bne mon_dnrm |else denorm
91 bne normal |require normalization of denorm
604 | condition exists with a very large norm and a denorm. One
650 | if(both operands are denorm)
654 | else(dest is denorm)
660 | else(dest is denorm)
664 | if(both operands are denorm)
694 cmpb #0xff,DNRM_FLG(%a6) |if both ops denorm,
[all …]
|
| D | round.S | 280 | would go less than 0. In that case the number becomes a denorm - the
329 subl %d0,%d5 |make op a denorm by shifting bits
383 | set (otherwise this would have been tagged a zero not a denorm).
398 | denorm --- denormalize an intermediate result
418 .global denorm
419 denorm: label
445 | ; shifted off in denorm routine)
458 | ; shifted off in denorm routine)
514 movel %d1,%d0 |copy the denorm threshold
534 movew %d0,LOCAL_EX(%a0) |exponent = denorm threshold
[all …]
|
| D | x_unfl.S | 33 |xref denorm
160 | force the precision to extended for the denorm routine. Use
196 | WARNING: a0 and d0 are assumed to be intact between the denorm and
209 bsrl round |returns rounded denorm at (a0)
|
| D | get_op.S | 48 | detects this and tags the number as a denorm. The routine
49 | res_func sees the denorm tag and converts the denorm to a
|
| D | util.S | 475 | ;smallest +ext denorm
477 | ;smallest +sgl denorm
479 | ;smallest +dbl denorm
533 leal EXT_PSML,%a1 |answer is negative smallest denorm
548 leal EXT_PSML,%a1 |answer is positive smallest denorm
|
| D | slog2.S | 122 bsr slognd | ...log(X), X denorm.
149 bsr slognd | ...log(X), X denorm.
|
| D | bindec.S | 172 | to signal denorm. If the input is unnormalized, normalize
178 clrb BINDEC_FLG(%a6) |clr norm/denorm flag
183 | Normalize the denorm
201 st BINDEC_FLG(%a6) |set flag for denorm
242 tstb BINDEC_FLG(%a6) |check for denorm
|
| D | do_func.S | 287 | ;00 = norm or denorm
436 bclrl #2,%d0 |alias denorm into norm
437 bclrl #2,%d1 |alias denorm into norm
|
| D | slogn.S | 478 | Simply return the denorm
|
| D | fpsp.h | 163 .set denorm_bit,7 | bit determines if denorm or unnorm
|
| D | srem_mod.S | 383 bra t_avoid_unsupp |check for denorm as a
|
| D | gen_except.S | 250 | x is an extended denorm can report an underflow exception.
|
| /linux-6.1.9/arch/m68k/ifpsp060/src/ |
| D | fpsp.S | 1427 # or double precision denorm, inf, or nan, the operand needs to be
1480 bne.b fso_dbl_dnrm # it's a skewed denorm
2339 mov.w &0x3c01,%d1 # pass denorm threshold
2340 bsr.l dnrm_lp # denorm it
3329 # denorm operand in the sgl or dbl format. NANs also become skewed and must be
9626 # the multiply factor that we're trying to create should be a denorm
10064 # preceding fmul was a denorm. but, it better not have been since the
10067 # bra t_avoid_unsupp # check for denorm as a
10166 # Entry point for scale w/ extended denorm. The function does
10184 tst.b LOCAL_EX(%a0) # is denorm negative?
[all …]
|
| D | pfpsp.S | 1426 # or double precision denorm, inf, or nan, the operand needs to be
1479 bne.b fso_dbl_dnrm # it's a skewed denorm
2338 mov.w &0x3c01,%d1 # pass denorm threshold
2339 bsr.l dnrm_lp # denorm it
3328 # denorm operand in the sgl or dbl format. NANs also become skewed and must be
|
| D | fplsp.S | 8553 # Simply return the denorm
8559 # satanhd(): computes the inverse hyperbolic tangent of a denorm input #
8759 bsr slognd # log(X), X denorm.
8804 bsr slognd # log(X), X denorm.
9269 # multiply with a denorm which will cause an unimplemented data type
9798 bsr.l unnorm_fix # convert to norm,denorm,or zero
10893 # a0 = input operand has been converted to a norm, denorm, or #
10920 cmp.w %d0, %d1 # will denorm push exp < 0?
10921 bgt.b unnorm_nrm_zero # yes; denorm only until exp = 0
|