ifpsp060/src/fplsp.S

5572 #--we want P+p = W+w  but  |p| <= half ulp of P
5573 #--Then, we need to compute  A := R-P   and  a := r-p
5583 #--Now we need to normalize (A,a) to  "new (R,r)" where R+r = A+a but
6003 #--we want P+p = W+w  but  |p| <= half ulp of P
6004 #--Then, we need to compute  A := R-P   and  a := r-p
6014 #--Now we need to normalize (A,a) to  "new (R,r)" where R+r = A+a but
6817 #	Notes:	If AdjFlag = 0, we have X = Mlog2 + Jlog2/64 + R,	#
6837 #		Note also that we use the FMOVEM instruction to move X	#
6965 #		inefficiency, we stay away from denormalized numbers	#
6966 #		the best we can. For |X| >= 2^(-16312), the		#
7543 # here, we build the result in a tmp location so as not to disturb the input
7596 #		overflow. Thus, we calculate sinh(X) as follows:	#
7710 #          Thus, we calculate sinh(X) as follows:			#
7833 #	6. (|X| >= 50 log2) Tanh(X) = +-1 (round to nearest). Thus, we	#
9242 # we may need to read the fpsr for the last fp op in t_catch2().
9267 # the multiply factor that we're trying to create should be a denorm
9268 # for the multiply to work. therefore, we're going to actually do a
9271 # later. we don't do this with the DENORMs above because this method
9275 	subi.l		&-0x3fff,%d0		# how many should we shift?
9323 # The source input is below 1, so we check for denormalized numbers
10025 #	For t_ovfl_sc() we take special care not to lose the INEX2 bit.	#
10037 # dst op is a DENORM. we have to normalize the mantissa to see if the
10039 # dst so we don't screw up the version passed to us.
10123 # occurred and has been logged. Now we need to see if an inexact	#
10156 # 060FPLSP should be inexact. So, if inexact is enabled, then we create	#
10159 #	Must check, too, if the result was zero, in which case we just	#
10239 # values are already in fp0 and fp1 so we do nothing here.