dsp56800e Freescale Semiconductor, Inc, dsp56800e Datasheet - Page 682

• Manufacturer Part Number: dsp56800e
• Description: 16-bit Digital Signal Controller Core
• Manufacturer: Freescale Semiconductor, Inc

The L bit is latched once it is set, and it is cleared only by processor reset or by an instruction that explicitly
clears it.
B.2.3
The extension in use bit (E) indicates whether the extension portion of a data ALU result contains
significant bits or whether they are just sign extension. If E is cleared, then bits 35–31 of the result are
00000 or 11111. E is set otherwise.
The E bit is calculated based on the 20- or 36-bit internal result that is computed by the data ALU. E is thus
calculated regardless of whether the specified destination contains an extension register. See Section B.1.1,
“Operand Size and Type,” for more information on how the internal result is calculated.
The calculation of E is not affected by the condition code mode. If the MAC output limiter is enabled, the
E bit is set based on the result before it passes through the limiter. Saturation that occurs in the MAC
output limiter can result in the E bit being set even though the final result does not use the extension
portion of the accumulator.
B.2.4
The unnormalized bit (U) is set to indicate that the result of a data ALU operation is not normalized. It is
set if the 2 most significant bits of the MSP of the result are the same, and it is cleared otherwise. The exact
computation is as follows:
U is not affected by the condition code mode. When the MAC output limiter is enabled, and when limiting
is performed, this bit is always cleared. If no limiting is done by the MAC output limiter, U is computed
using the preceding equations.
In general, the calculation of the U bit when a CMP.W instruction is executed is not affected by the state of
the MAC output limiter. However, if the first operand to CMP.W is not a register (that is, it is a memory
location or an immediate value), and if the second operand is X0, Y0, or Y1, then the calculation of U can
be affected if the MAC output limiter is enabled.
B.2.5
The negative bit (N) indicates whether or not the result of a data ALU or an AGU calculation is negative. A
value is considered negative if the MSB of the result is set. Note that N is set based on the size of the
destination for the result, not on the internal (20- or 36-bit) size.
The condition code mode has no effect on the calculation of N for 8-, 16-, 24-, and 32-bit results. For 20-
or 36-bit results, N is based on bits 15 or 31 (respectively) if 32-bit condition code mode is selected. When
saturation is enabled (SA = 1), the N bit is set based on the result before it passes through the MAC output
limiter.
Calculation of the N bit is somewhat different for the ASLL.W, ASLL.L, ASRAC, LSRAC, and IMPY.W
instructions. See Section B.3.3, “Special Calculation Rules for Certain Instructions,” for more information.
B-6
U = (Bit 31 ⊕ Bit 30) for 20-, 32-, and 36-bit calculations
U = (Bit 15 ⊕ Bit 14) for 8- and 16-bit calculations
Extension in Use Bit (E)
Unnormalized Bit (U)
Negative Bit (N)
DSP56800E Core Reference Manual
Freescale Semiconductor