### Table 2.6:Computational complexity of Rx Pulse Shaping Filter in FDD-mode specific design for 1 frame The complexities of Rx Pulse Shaping Filter in FDD and HSDPA -specific modes are exactly the same. Computational complexity of Rx Pulse Shaping Filter in TDD-mode specific design for 1 slot period is in Table 2.7.

### Table 2: Complexity of spatial-domain versus DCT do-

"... In PAGE 4: ... #2812#29 is reduced to N F #02 N G exactly. The computation complexity for one output 8 #02 8 block when compositing is performed in the spatial- domain versus DCT-domain is shown in Table2 . Note... ..."

### Table 3: Compared complexity of the algorithms counting all oating point operations but the ones necessary to compute the componentwise exact partial products Routine Priest apos;s Shewchuk apos;s Ours

### Table 3.17: Computational complexity of Frequency Estimator in FDD-mode of multi-mode design for 1 frame Computational complexity of Frequency Estimator in HSDPA-mode specific design is exactly the same as in FDD. Note that the complexity figures for FDD and HSDPA of multi-mode have been slightly increased compared to those for single mode because some overhead signal control regarding reconfigurability control is required to share the module in an appropriate way between modes and no further optimization have been required in algorithm point of view. Computational complexity of Descrambler in TDD-mode of multi-mode design is in Table 3.18.

### Table 4.3: Asymptotic adder complexities (unit-gate model).

1997

### Table 4.3: Asymptotic adder complexities (unit-gate model).

1997

### Table 4.3: Asymptotic adder complexities (unit-gate model).

1997

### Table 1 Post-construction average time complexities for 5 di erent problems using standard and distributed su x trees with k computing nodes. r is the number of strings for the All Pairs Su x-Pre x problem and the number of patterns for Exact Set Matching.

"... In PAGE 12: ... We suppose that there are k computing nodes and assume for simplicity that k = jzj, where is the alphabet size and z is the xed pre x. Table1 compares the expected running times for the solution of these 5 problems using the fastest serial method (based on a standard su x tree) and a parallel method (based on distributed su x trees). The running time of (distributed) su x tree construction is not included.... ..."