IBM 7090 - Page 80

Floating Add Magnitude F AM +0304 (Min I, E, 4L)
(Max
I, E, 13L)
This instruction algebraically adds the magnitude of the floating-point number stored
at the location designated by the address to the floating-point number in the accumulator.
The sequence of operations is identical to that of FAD except that the SR sign is set
positive during the E cycle. This control line is on Systems 2.09.95.1.
Unnormalized Add Magnitude
U AM -0304 (Min I, E, 3L)
(Max I, E, 8L)
This instruction operates the same as F AD, with the exceptions pointed out in UF A
and FAM.
Floating Subtract FSB +0302 (Min I, E, 4L)
(Max
I, E, 13L)
This instruction algebraically subtracts the floating point number stored at the loca-
tion indicated by the address from the floating-point number in the accumulator. FSB
operates the same as FAD, except that the sign of the word in the SR is inverted during
the E cycle. This is shown on Systems 2.09.95. 1.
Unnormalized Floating Subtract
UFS -0302 (Min I, E, 3L)
(Max
I, E, 7L)
This instruction algebraically subtracts two floating-pOint numbers without normal-
izing the result. Execution is the same as F AD with the exceptions noted in UF A and
FSB.
Floating Subtract Magnitude FSM +0306 (Min I, E, 4L)
(Max
I,~, 13L)
This instruction algebraically subtracts the magnitude of the floating-point number
stored at the location indicated by the address from the floating-point number in the
accumulator. Execution of this instruction is the same as FAD, except that the sign of
the word in the SR is forced minus during the E cycle. This is shown on Systems
2.09.95. 1.
Unnormalized Subtract Magnitude USM -0306 (Min I, E, 3L)
(Max I, E, 9L)
This instruction operates the same as FSM, with the exceptions explained under
UFA.
Floating Multiply
FMP +0260 (Min I, E) Figure 5.3-24
(Max I, E, IlL)
This instruction multiplies the floating-point number stored at the location deSignated
by the address by the floating-point number stored in the MQ. The product appears as
two floating-point numbers: the most significant part in the accumulator, and the least
significant part in the MQ. The signs of both registers are set to the algebraic sign of
the product. If the multiplicand is zero, the product will be two normal zeros with
proper algebraic signs. If the AC fraction or the multiplier fraction is zero, the AC is
reset to a normal zero and no characteristic is assigned to the MQ. If the AC fraction
is not zero, the MQ is assigned a characteristic 2710 less than the AC characteristic.
Remember that when multiplying numbers using exponents, the fractions are multi-
plied and the exponents are added. Multiplying the fractions is done much the same as
in a MPY operation, by adding and shifting. Adding the characteristic is not enough
79