Page vi: Replace 5.4.3 LP-Norms with 5.4.3 LP-Norms.
Page vi: Missing 5.3 Stability, reorder all subsequent entries to 5.4 Quantization
down to 5.12 DCT Processor, Cont.
Line 9: Remove unneeded hyphen in "circuit"
Page 10, Figure 1.8: NAND-gates should be replaced by AND-gates.
Page 10, Figure 1.9: The plus-sign in the description of F should have a ring around it
(it should be an xor sign).
Page 11, 2nd paragraph, line 3: The first letter in "integrated" should also be in italic.
Page 12, Figure 1.11: Hyphen should be replaced by a slash in "Verification/Validation"
Page 12, Figure 1.12: Figure text should not end with a dot.
Page 21, Equation 1.5: Remove the extra equal sign. The last line in the equation should read
O(nlogc(b))
Page 32, 2nd paragraph, 3rd line: Last letter in wells should be italic.
Page 35, text under equation 2.1c: The equation should have both Tox and L in the denominator.
Page 35, text under equation 2.2c: The equation should have both Tox and L in the denominator.
Page 38, the equation on the middle of the page should read: Vout,low
= R2/(R1+R2)Vdd (.i.e., multiply
Vdd with R2)
Page 39,equations at the middle of the page: Remove spacing after n-
and p- in n-transistor and p-transistor .
Page 42: Figure 2.14 the arrows on the right right side should switch
places, i.e., CL=105fF should point to the rightmost line and
CL=55fF should point to the second rightmost line.
Page 44, line 8 from bottom: the number 103 in the right side of equation
should be 103.
Page 47, line 2 should read: .... air velocity, v, and the ..
Page 47, equation 2.11: right side of the equation should be -0.3.
Page 53, Section 2.5, first paragraph: Remove the 3rd sentence ("As
of ....by the end of the century").
Page 55, problem 2.2: Determine the ID-VDS ... should be
Determine the ID-VDS ...
Page 63: equation 3.3 reads ...e-jwT should read ...e-jwnT
Page 63: Example 3.1, bottom equation: Replace n ≤ 0 ≤
n0 with 0 ≤ n ≤ n0
Page 65: Example 3.2, first equation: Replace an with an
Page 72, under Figure 3.11: Equal signs should be aligned horisontally.
More space between if and x1.
Page 73: The first equation under section 3.10 replace y(ejwnT)
= ... with y(n) = ...
Page 73: Equation 3.12 replace y(ejwT) = ... with Y(ejwT)
Page 77: Last text line: See problem 3.16 should read See problem 3.15
Page 78, Figure 3.15: Replace with new figure with correct axes: tg
should not be using superscript.
Page 81, Second set of equations in Example 3.5: Exponent of z is wrong for the last terms of the equations. They should
be: +b2z-2U(z) and +a2z-2U(z)
Page 85, Line 5 at the first paragraph after 3.14.2 RLS: ... requires
many more operations then the stochastic-gradient algorithm ... should
read ... requires many more operations than the stochastic-gradient algorithm
...
Page 86, Equation at end of page: Replace n = 0, 1, ..., N-1 with k
= 0, 1, ..., N-1.
Page 87, Box 3.2 text: Remove dot at end of "Direct computation of the
DFT"
Page 87, Equation under Box 3.2: Tmult = 10242
4 100 ns = 0.419s
Page 89: Box 3.3 p:=Digit_Reverse(k*2^M1,NumofStage); should read p=Digit_Reverse(k/2^M1,NumofStage);
Page 90: Figure 3.23 The text below X3 denoted X(0,0,0) should
read X3(0,0,1). To the left of this text on the arc x(0,0,1)
between X2 and X3 is a multiplier missing. The text
X3(1,1,0) should read X3(0,1,1). The text X3(0,1,1)
should read X3(1,1,0).
Page 93, fourth line, equation X2(k0,k1)=Sum
X1(k1,n0)W(2k1+k0)n0=... should
be X2(k0,k1)=Sum X1(k0,n0)W(2k1+k0)n0=...
Page 94, Box 3.6 text: Remove ending dot.
Page 95, Figure 3.26 text: "Sande-Tukey's FFT for N=8." Remove ending
dot.
Page 96, Equation (3.22): X(N-0)/N should be X(N-n)/N.
Page 97, last equation: ...= 2*1024/(16*166) =... should
be ...= 2*1024/(16*106) =...
Page 98: third line text reads ....(See problem 3.29). should read ....(See
problem 3.28).
Page 100: Equation 3.25 should not have an equation number (it is part
of equation 3.24)
Page 101: definition of ck called 3.25 should not have an
equation numbering.
Page 102, Equation 3.29: The constant before the summation should be
the square root of 2/(N-1)
Page 102, Equation 3.29: Replace sqrt(2/(2N-1)) with sqrt(2/(N-1)) and
replace ck with cn.
Page 102, Equation 3.30: Definition of ck should be 1/2 for
k = 0 or k = N-1. This definition should not be called equation 3.29.
Page 103, Matrix at top of page: Value in upper right corner should
be 1/2.
Page 104, line 6: ck should be 1/2 when k=0 or k=N-1.
Page 110, Reference 45: Replace "Weidong K." with "Kou W.".
Page 110: problem 3.3 b) text reads an x(n) <-> X(z/-a) should read
anx(n) <-> X(z/a)
Page 124, First paragraph: Delete last line in paragraph ("This structure
... discussed in Chapter 11")
Page 128, Figure 4.9: The gray box representing the stopband requirement should touch the wT- axes.
Page 130, Last paragraph in section 4.6.1, first sentence: Replace B05
with P05 to make the sentence ... uniquely described by its order, P05,
except for the passband ....
Page 134, line 8: G should be equal to 1.16187662*1012.
Page 136, Figure 4.15: Text on axes should have italic T, and tg should
use non-superscript g.
Page 144, line 8: the parathesis should also include e=...
Page 147, line 17: the last equation inside the parathesis should be
V1=V2.
Page 147, Last three equations: B1 = A2 + alpha(A2
- A1) is equation 4.42, B2 = A1 + alpha(A2
- A1) is equation 4.43, and alpha = (R1-R2)/(R1+R2) is equation
4.44.
Page 163, Figure 4.52: The subtraction should have negative input swapped,
i.e., the lower branch should be subtracted from the upper branch.
Page 165, last paragraph, second to last sentence: Infinity sign should
be in subscript.
Page 169, Box 4.1: the begin on the fourth line should be in bold. Indentation
of the fifth line should be the same as on the sixth line. The fifth line
should read x1(Ln+i) := 0; and the sixth line should read
Compute:y(Ln+i);
Page 173, Figure 4.64: should read: Spectrum of the input signal and the interpolated signal.
Page 179, last paragraph, line 2 and line 3: ...35.2 * 106 adaptors/s
corresponding to 35.2 * 106 multiplications/s and 105 * 106 additions ...
should be ...35.2 * 106 adaptors/s
corresponding to 35.2 * 106 multiplications/s and 105 * 106 additions ...
Page 184, line 4 and 5: the minus signs should be longer for sp2,3
and sp4,5.
Page 189, Figure 5.1: Change text to read: "... with two quantizers"
Page 190, line 11 from bottom: ... value Q/2 which ... should be ...
value -Q/2 which ...
Page 190, line 7 from bottom: ... therefore 18.285Q should be ... therefore
-18.285Q.
Page 196, Second paragraph, fourth line: ak(n)2
bk(n)2 > 0 sholud be ak(n)2
- bk(n)2 > 0
Page 196, Equation 5.3: Should be called equation 5.2
Page 196, Second paragraph after equation 5.3, fourth line: Q = 210
should be Q = 2-10.
Page 197, Third line: ... to [2,2], since ... should be ... to [-2,2[,
since ...
Page 197, Seventh line: normal range [1,1[ at ... should be normal range
[-1,1[ at ...
Page 199, line 1 at the second paragraph: ... multiplie;s .. should
be ... multipliers ...
Page 199, expression after the fourth paragraph: |v(n)| = ... , an
extra right bracket should be deleted.
Page 201, Line 9: |Y| = |(UV) Wp| = |UV| ... should be |Y|
= |(U - V) Wp| = |U - V| ...
Page 202, Equation 5.6: Remove the exponent p from the equation. Set
starting value for integration to be minus infinity.
Page 203, Equation 5.10: Use smaller fonts for wT under max.
Page 203, Equation 5.11: Use smaller fonts for ||2 and ||2p. Move ||q
down to become a subscript (||q).
Page 204, Coefficient after second paragraph: ... a11 = 1.0565532
... should be ... a11 = -1.0565532 ...
Page 205, Box 5.1: constant declaration for the following coefficients
should be negative, a11, b21, a12, b22, a13, b23 and b24.
Page 205, Last line: ... a12 = 0.34260417 ... should be ...
a12 = -0.34260417 ...
Page 206, Coefficient after second paragraph: ... a13 = 0.608176
... should be ... a13 = -0.608176 ...
Page 206, Last line: ... a11 = 0.287702 ... should be ...
a11 = -0.287702 ...
Page 207, Coefficient after second paragraph: ... a12 = 0.2273435
... should be ... a12 = -0.2273435 ...
Page 207, Coefficient after third paragraph: ... a13 = 0.6403554
... should be ... a13 = -0.6403554 ...
Page 208, Equation 5.13: m should be equal to (Qc-1)Q/2 in
the case of truncation.
Page 208, Last paragraph, second to last sentence: ... and Q/2 for ...
should be ... and -Q/2 for ...
Page 211, line 7: ... more than N1 such ... should be ...
more than N-1 such ...
Page 212, First line: Replace (N 1)sigmaB2 with
(N-1)sigmaB2.
Page 212, Equation at bottom of page: Set starting index on lower summation
to be k=1.
Page 213, Figure 5.20: Change figure text to "Recursive filter with
error feedback"
Page 213, Second equation: E(z) = V(z) X(z) ... should be E(z) = V(z)-
X(z) - ...
Page 213, Paragraph under last equation, fifth line: replace denuminator
with denominator
Page 217, all equations: Use smaller fonts for all occurrencens of ||2
and ||1.
Page 218, line 15: right side of the equation is 8(1-2-M)sigma2B.
Page 222, problem 5.8: Use smaller fonts for ||2 and ||2p.
Page 222, problem 5.12: ... and a0=a1 should be
... and a0=-a1.
Page 223, Fig. P5.18a should read fsample1 = 44.1 kHz.
Page 224, problem 5.19: H(ejwT)=ejwT... should
be H(ejwT)=e-jwT...
Page 224, problem 5.19 (line 10): ... with an accuracy of Q10
should be ... with an accuracy of Q-10.
Page 230, Figure 6.6:
Bottom adder should have an arrow on the left input.
Page 230, second paragraph, last line: ...algorithm, however, is useable
for low ...
Page 231, Figure 6.9 b:
1/Throughput is defined as distance between MSB of first and second
output.
Page 232, Figure 6.10 a:
The LSB bits should be named as xwd-1, ywd-1,
and sigmawd-1.
Figure 6.10 b:
The LSB bits should be named as xwd-2, xwd-1,
ywd-2, ywd-1, sigmawd-2, and sigmawd-1.
The output LSB should be shifted one bit to the right, putting the
sigmawd-1 in the same column as xwd-2 and ywd-2.
Page 233, Figure 6.11:
The output values values should all be denoted y, i.e., change the
bottom xwd-1 to ywd-1.
Page 235, Box. 6.1, entry #1:
...with transmittance = 1. Transmittances of -1 can often be propagated
into adjacent branches.
Page 235, Box. 6.1 text: Remove ending dot.
Page 242, equation system after the first paragraph: Add equation
v2(n+1) = v1(n)
Page 246, Figure 6.30 text: Remove ending dot.
Page 247, line 4: ... one clock cycle longer that ... should be ... one clock cycle longer than ...
Page 251, first paragraph, second line: the line is ended prematurely.
Should be continued with the text from lines 3-5.
Page 253, line 15: x(n)-x(n-2)u(n-2) should be x(n)-x(n-2)-u(n-2).
Page 257, equation after the first paragraph: remove the last two derivations.
Page 258, Figure 6.51: Remove extra line between adaptor a1 and a5.
Change ending subtraction to subtract the lower branch from the upper branch.
Page 258, Equation at middle of page: Remove factor 2 from Tmult,
Tmin = (Tadd + Tmult + Tadd)/2
Page 259, Figure 6.53: Change names of the two last PEs in the sequence
to PE2 and PE3.
Page 261, First two equations: The equation set x(n+1) = A... and y(n+1)
= C... is equation 6.2. Replace y(n+1) in this equation with y(n)
Page 262, First equation: The equation set x(n+L+1) = A'... and y(n+L)
= C' ... is equation 6.3.
Page 263, equation 6.6: the sum should start with k=0 and end with
k=M for the term ckz-k.
Page 264, line 14: the denominator at the right side of equation should
be
1-(b1-c1)z-1-b1c1z-2.
Page 264, the fourth equation: denominator for the second term
should be 1-b12z-2 in stead of 1-b12z-1.
Page 265, line 1 in the fourth paragraph: An increase in the ... should be where c0 = 1. An increase in the....
Page 265, line 15 from bottom: e1=b1 c1
c2 should be e1=b1-c1-c2.
Page 266, line 5: zM should be z-M.
Page 266, bottom paragraph, first sentence: Should read: This transfer
function can be realized by a number of cascaded FIR structures followed
by a recursive structure as shown in Figure 6.59.
Page 268, figure 6.60: Replace y(m) with y(4m), y(m+1) with y(4m+1),
y(m+2) with y (4m+2), and y(m+3) with y(4m+3).
Page 272, problem 6.5b: Replace 4.64 with 4.65.
Page 272, problem 6.9: Replace 6.53 with 6.60 and 4.64 with 4.65.
Page 272, problem 6.10: Replace 3.21 with 3.18.
Page 275, problem 6.31: Replace 6.28 with 6.30.
Page 283, top equation: use lower-case t: tFFT = ...
Page 283, second equation: Replace 0.372 in the denominator with 0.32
Page 286, Box 7.1, 15th line from bottom: Replace loop variable n with
variable m.
Page 286, Box 7.1, 3rd line from bottom: hyphen after the semicolon
should be moved right to the start of the comment.
Page 288, definition of k2: Second line should read k1
+ N/2 Stage ≥ 2
Page 288, Box 7.3: The calls to Butterfly should be replaced with
Butterfly(x_tmp, k1, k1Ns, Wcos, Wsin);
if (stage = 1) then
Butterfly(x_tmp, k2, k2Ns, Wsin, -Wcos);
else
Butterfly(x_tmp, k2, k2Ns, Wcos, Wsin);
end if;
Also move the comments "-- for loop" at "end loop;" to the right
Page 289, definitions of k2: First line should read k1
+ Ns/2 Stage = 1
Page 289, Box 7.4, 2nd line from top: Align comment to be as the first
line.
Page 289, Box 7.4: The calls to Butterfly should be replaced with
Butterfly(x_tmp, k1, k1Ns, Wcos, Wsin);
if (stage = 1) then
Butterfly(x_tmp, k2, k2Ns, Wsin, -Wcos);
else
Butterfly(x_tmp, k2, k2Ns, Wcos, Wsin);
end if;
Page 291, Box 7.5, 9 lines from bottom: hypen after the semicolon should
be moved right to the start of the comment. Remove definition of variables
TempRe and TempIm. New variables:
variable x_tmp1, x_tmp2, x_tmp3, x_tmp4 : complex;
Remove the indent for the calculations of Wcos and Wsin.
Change the code after calculation of Wsin up to the end loop to be:
Memory(Rflag,x_tmp1,x_tmp2,x_tmp3,x_tmp4,k1,k1Ns,k2,k2Ns);
Butterfly(x_tmp1,x_tmp2,Wcos,Wsin);
if(Stage = 1) then
Butterfly(x_tmp3,x_tmp4,Wsin,-Wcos); -- Butterflies
else
Butterfly(x_tmp3,x_tmp4,Wcos,Wsin);
end if;
Memory(Wflag,x_tmp1,x_tmp2,x_tmp3,x_tmp4,k1,k1Ns,k2,k2Ns);
Should also add a new definition of the Butterfly procedure:
Procedure Butterfly(variable x_tmp1, x_tmp2 : inout complex;
variable Wcos, Wsin : real) is
variable TempRe, TempIm : real;
begin
TempRe := xtmp_1.re - xtmp_2.re;
TempIm := xtmp_1.im - xtmp_2.im;
xtmp_1.re := xtmp_1.re + xtmp_2.re;
xtmp_1.im := xtmp_1.im + xtmp_2.im;
xtmp_2.re := TempRe * Wcos - TempIm * Wsin;
xtmp_2.im := TempIm * Wcos + TempRe * Wsin;
end Butterfly;
Replace the Memory_Read(i, data) with Memory_Read(ind, data);
Page 296, Figure 7.12: Replace y(n) with y(n-1)
Page 299, Figure 7.18: Replace y(n) with y(n-1)
Page 300, figure 7.19: Replace x(n-1) with x(2n), x(n) with x(2n+1),
y(n-2) with y(2n-1), and y(n-1) with y(2n)
Page 300, last paragraph, line six to seven: reads ... The number of
unit delays is 92/2 = 46 per sample... should read ... The number of unit
delays is 76/2 = 38 per sample...
Page 301, Figure 7.20: Replace x(n-1) with x(2n), x(n) with x(2n+1),
y(n-2) with y(2n-1), and y(n-1) with y(2n). Move the output dot of the
adder generating y(2n-1) down on the symbol.
Page 302, figure 7.21: Replace all inputs to be x( instead of y(.
Page 303, figure 7.22: Replace x(5n-4) with x(5n), x(5n-3) with x(5n+1),
x(5n-2) with x(5n+2), x(5n-1) with x(5n+3), x(5n) with x(5n+4), N3
with N1, N1 with N3, N0 with
N4, y(5n-4) with y(5n), y(5n-3) with y(5n+1), y(5n-2) with y(5n+2),
y(5n-1) with y(5n+3), and y(5n) with y(5n+4)
Page 305, figure 7.24: Replace x(6n-5) with x(6n), x(6n-4) with x(6n+1),
x(6n-3) with x(6n+2), x(6n-2) with x(6n+3), x(6n-1) with x(6n+4), x(6n)
with x(6n+5), N3 with N2, N2 with N3,
N1 with N4, N0 with N5, y(6n-5)
with y(6n), y(6n-4) with y(6n+1), y(6n-3) with y(6n+2), y(6n-2) with y(6n+3),
y(6n-1) with y(6n+4), and y(6n) with y(6n+5)
Page 309, Figure 7.34: Replace y(4n1) with y(4n+1) in the second figure.
Page 310, paragraph before equation 7.4: Replace with "The computation
time for one bit-serial multiplier is Wf+Wd clock
cycles, which in most cases is longer than the minimal sample period Tmin.
Therefore, the schedule must cover at least m sample periods, where m must
be selected as"
Page 310, equation 7.4: Replace with: m ≥ nearest larger integer
of[(Wd+max{alphaif})/Tmin]
Page 315, Section 7.6.4, last sentence: Should read ... scheduling the
operations that do not belong ...
Page 316, equation 7.5: the lower bound of the sums at the right should
be i=ta and i=t0.
Page 316, equation 7.6: the lower bound of the sum at the right should
be i=t0.
Page 317, Figure 7.44, at step 3 and 4, the assignment for operation
a1, combination Self: step 3 reads 4/3 should be read 1/3, step
4 reads -4/3 should be read -2/3.
Page 319, lines 13 and 14 from the bottom: (a) and (b) should not be
in bold.
Page 319, last sentence: Remove ("However, the approach...")
Page 322, box 7.6: On the seventh line should the second repeat statement
be bold. On the 9th line should the lower-case o be replaced by 0 as in
if dCost(i,k) ≤ 0 then accept.
Page 323, Heading 7.7: The word CONT. should be in upper case.
Page 324, figure 7.53: it should be 62.5 ns in stead of 62.5 us in
the figure.
Page 326, figure 7.55: the address should be delivered to W4
in stead of R4.
Page 331, figure 7.68: Replace PE1 with PE1, PE2 with PE2,
and PE3 with PE3.
Page 332, box 7.7, step 1: replace "processes having the longest lifetimes"
with "process having the longest lifetime". step 4: After Stop on the 5th
line in this step should one line with an "end if;" in bold be added. The
lines following these should be indented to match the inital if in step
4 (i.e., moved to the left one level).
Page 334, paragraph below Figure 7.72, third sentence: Replace 2Wd
with 2(Wd+3)
Page 337, Figure 7.77: Variable 19a should end at time 40, i.e., put
a black square at time 40. Rename outputs y(m+2) with y(4m-2), y(m+3) with
y(4m-1), y(m) with y(4m), and y(m+1) with y(4m+1).
Page 338, Figure 7.78: Swap 2b and 2a, swap 7a and 7b.
Page 339, Figure 7.79: Add Negation of variable 7b, replace 16b with
16a, add ending of variable 11a (ends at time 8), adjust ending of 6a to
be on the middle of time 28.
Page 339, Figure 7.80: Replace 7a with 7b, 4a with 4b, add negation
of 16b, 10b, 8b, 17b, 11b, and 13b.
Page 340, Figure 7.81: Add negation of variable 7b, replace 11b with
11a.
Page 340, Figure 7.82: Add negation of variable 16b, move the end of
variable 16b and the start of variable 1b two time units to the right.
Add the name 13b to the variable in cell2 starting at 47 and ending at
4. Make the figure caption read "Cell assignment for memory 2"
Page 343, Figure 7.86: Remove the fourth column of Butterfly operations.
The lower output from the top-left butterfly should start at the bottom
of the operation.
Page 344, figure 7.87: Replace RAM0 with RAM0 and RAM1 with
RAM1
Page 345, figure 7.88: The top left butterfly operation should have
its lower output start at the bottom of the operation. Replace PE0 with
PE0 and PE1 with PE1.Replace exponent on the lower
left butterfly to be 7 instead of 3. Replace exponent on the lower right
butterfly to be 0 instead of 1 or 3.
Page 345, Figure 7.89: The text should read "Exclusion graph for Pes
for a four-point FFT"
Page 346, Figure 7.90: There should be a dashed line from the lower
output of the Butterfly in row 1 in stage 1 to the Butterfly in row 3 stage
2.
Page 346, second line from the bottom: should read "those in row..."
Page 347, figure 7.92: The top left buttefly named W0 should
have its lower output start at the bottom of the operation. Replace PE0
with PE0 and PE1 with PE1. The exponent
of the lower right butterfly should be 0 instead of 3.
Page 349, figure 7.95: Replace PE1 with PE1 and
PE2 with PE2.
Page 353, problem 7.12, first line: ...using only one type of PE that
can....
Page 354, problem 7.19: Replace Figure 7.19 with Figure P7.19.
Page 362, Figure 8.7: Replace the text DAM Bus on the right in the figure
with DMA Bus.
Page 368, Figure 8.11: Make the lines inside the dotted line box in
figure 8.11 thicker
Page 369, Figure 8.12: Make the box of the NAND gate thicker.
Page 374, figure 8.21: it misses four dots in the cross points in
the binary tree.
Page 390, Example 9.1, first equation: The index to ai should
be in italic.
Page 390, last paragraph, third sentence and fourth sentence:
...to the cleared accumulator, and x(n) is loaded into the first shift
register position. In the next time step the shift register is shifted,
and the product aN-2x(n-N+2) is computed and added to the accumulator.
Page 391, Table 9.1: x(n) and x(n+1) should be moved one column to the
left. y(n) moved one column to the left, y(n-1) removed, and y(n+1) added
at the bottom of the rightmost column.
Page 391, figure 9.5: replace ... for the direct form FIR ... with
... for the transposed, direct form FIR ...
Page 393, Example 9.4, first equation. Remove the +1 in the summation
and add a closing bracket, i.e., the summation is done over ai[x(n-i)
+ x(n-2Q-1+i)].
Page 395, First paragraph in 9.4.1, last line: methodology
Page 400, equation 9.4: The second equal sign should be removed and
the vector and array moved closer together.
Page 402, equation 9.5: The second equal sign should be removed and
the vector and array moved closer together.
Page 402, Example 9.7, bottom line: replace cy with c4
Page 403, equation 9.6: The second equal sign should be removed and
the vector and array moved closer together.
Page 406, figure 9.20: Replace the CacheK with Cache1
for the 3rd Cache (connected to PEN), and replace Cache1
with CacheK for the fourth Cache (also connected to PEN).
Page 406, Equation 9.8: Replace expression with K ≤ max(i,t){ni(t)}
Page 408, second line: ... is their clock frequency... should read ...
is their clock period...
Page 412, figure 9.29: Replace x(n) with x(2n), x(n+1) with x(2n+1),
y(n) with y(2n-1), and y(n+1) with y(2n)
Page 417,
Page 419, Figure 9.35: Replace top W1 with R1,
top W2 with R2, top W3 with R3,
and top W4 with R4.
Page 420, Figure 9.36: Box labeled Stage PE should be labeled Stage.
Page 420, Box 9.1: On the third line should begin be bold. On the sixth
line should else be bold.
Page 421, Box 9.2, third line: begin should be bold.
Page 422, figure 9.37: Both P and i out from Base index Generator should
be in italic.
Page 422, Figure 9.37: Box labeled Stage PE should be labeled Stage.
Page 423, Box 9.3: On the third line should begin be bold. On the fourth
line should the if be bold.
Page 423, Box 9.4: On the third line should begin be bold. The empty
6th line from the bottom should be removed. Replace the end; at the second
line from the bottom with end if;
Page 423, Box 9.5, third line: begin should be bold.
Page 424, Box 9.5: The third and fourth line should be moved to left
to align the if with the else on the fifth line. The sixth line should
be moved to right to align with BF := on the fourth line. The eighth line
should be moved to the right to align with the seventh line. The else on
line 11 should be bold and moved to the right to align with seventh line.
Add a line with end if; between lines 14 and 15, aligned with the if on
line 15. Replace Do with Do_in on line 15.
Page 424, Box 9.6: The begin on line three should be bold. Replace Do
with Do_in on lines four, 10, and 24. Remove the empty 7th line. The else
on line 13 should be bold.
Page 432, problem 9.8: alpha1=3/16, alpha3=3/16,
alpha5=3/16 should be alpha1=-99/1024, alpha3=381/1024,
alpha5=-1509/2048.
Page 435, problem 9.16: Replace 9.39 with 9.38
Page 435, problem 9.17: Replace 9.40 with 9.39
Page 438, second paragraph, last line: ... handling of don't care values...
Page 446, picture 10.16: The middle transistor in the last stage should
be called p5 instead of n5.
Page 450, line 6 from bottom: ... step n1 ... should be ... step n-1...
Page 454, equation 10.8: D(n+N) = T(n+N-1)T(n+N-2)...T(n)
Page 457, Figure 10.34: In step 1 should the first subtrellis have its
top leftmost line thicker (7--). The second subtrellis should have the
bottom leftmost line and the diagonal line (left low, right high) in the
middle to be thicker, forming (=7\). The third subtrellis should have the
leftmost diagonal and the two lower lines thicker, forming (X==). The fourth
subtrellis should have its lower left line, upper right, and middle diagonal
thicker, forming (=X=). The second step should have the first diagonal
(lower left, upper right), left middle lower line, middle right diagonal
(lower right, upper left), and right diagonal (lower left, upper right)
marked, forming (7>XX).
Page 461, equation 11.1: The last part consisting of = 1 should be removed.
Page 462, fourth paragraph, first sentence: ... in the range -1 <
x < 1. Hence...
Page 463
(-0.828125)10 = (1.110101)SM
First paragraph: ... in the range -1 ≤ x ≤ 1. Hence, ...
Second paragraph:
...lie in the range -1+Q ≤ x ≤ 1-Q where... ...as either
+0 or -0. This may... ...values +1 and -1 can not....
Page 464, Section 11.2.3, second paragraph:
...in the range -1+Q ≤ x ≤ 1-Q where... ...values +1
and -1 can not...
(-0.828125)10 = (1.001010)1C
Page 464, equation 11.8: rk should be rk.
Page 464, last paragraph, 3rd sentence: Replace ...are simpler to implement...
with ...are more complicated to implement...
Page 465, Section 11.2.4, second paragraph: Add minussigns and change
exponent to become:
...the range -1 ≤ x ≤ 1-Q where Q = 2-Wd+1...
Page 466, Example 11.3:
Assume that the numbers are 6/8 = (0.110)2C, 4/8 = (0.100)2C,
and -7/8 = (1.001)2C.
We first add
S1 = 6/8 + 4/8
= (0.110)2C + (0.100)2C
= {overflow} = (1.010)2C = -6/8
Page 466, Figure 11.1: Change binary representation for 2/8 to 0.010
Page 467, top of page:
change all number indexes from 2 to 2C.
Section 11.2.5, second paragraph: Add minus signs and change exponent:
...in the range: -1 ≤ x ≤ 1-Q, where Q = 2-Wd+1.
(+0.828125)10 = (1.110101)BO
(-0.828125)10 = (0.001011)BO
Page 468, end of second paragraph:
Typically digits can take one of three values: -1, 0, or +1.
Section 11.3.1
...the range -2+Q ≤ x ≤ 2-Q, where Q = 2-Wd+1,
is... Change summation start index to i=0
Section 11.3.1, second paragraph
...represented by (0.1000-1)SDC or by (0.01111)SDC.
The decimal number (-15/32)10 = (1.10001)2C can be
represented by (0.-10001)SDC or (0.0-1-1-1-1)SDC.
Example 11.4
...numbers are (1-11-1)SDC = (5)10 and (0-111)SDC
= (-1)10.
Page 469, Table 11.1:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Section 11.3.2, second paragraph:
...(15/32)10 is represented by (0.1000-1)CSDC
and (-15/32)10 by (0.-10001)CSDC.
Page 469, bottom equation: Move subtraction sign before second exponential
term on the left half of the equation. 2k+n+1-2k
= 2k+n...
Page 470, first paragraph:
...0s, followed by a -1. For example, (0.011111)2C = (31/64)10
corresponds to (0.10000-1)CSDC = ((32-1)/64)10.
Second paragraph:
...example (...00111100...)2C is converted into (...01000-100...)CSDC...
...reexamined and pairs of type (-1,1) are changed to (0,-1) and triplets
of type (0,1,1) are changed to (1, 0, -1)... ...(0.110101101101)2C
= (1.00-10-100-10-101)CSDC. ...
Third paragraph:
...0 or 1 and the other part the -1 digits....
Page 473, figure 11.4: The rightmost carry input should be c4
Page 475, figure 11.5: all multiplications with a0 should
have a minus sign in front, except for a0x0. The
term a1x1 should not have a minus sign. All terms
with x0 except a0x0 should have minus
signs.
Page 476, line 3 and 1 from bottom: For example, (0.5)10=
... should be (-0.5)10= ..., and (0.25)10 should
be (-0.25)10.
Page 476, Figure 11.6: All a0 bits should be replaced with
-a0, and -aWc-1 should be replaced with aWc-1
Page 477, The proof for modified Booth encoding. Remove the -1 from
the exponent of 2 in the second to last term on line 1. Change -1 to +1
on the exponent of 2 in the middle summation on line 2. Change summation
index end to 8 and change -1 to +1 on the exponent of 2 on the middle summation
on line three. First line of the equations should read
8 7 x = sum ... = sum (x2i-12-2i+1) + sum (x2i2-2i) - x020 1 1Second line of the equations should read
8 7 7 x = sum (x2i-12-2i+1) + sum (x2i2-2i+1) - 2 sum (x2i2-2i-1) - x020 1 1 1Third line of the equations should read
8 8 8 x = sum (x2i-12-2i+1) + sum (x2i2-2i+1) - 2 sum (x2(i-1)2-2i+1) - x020 1 1 2Page 479, Figure 11.10: The three leftmost columns at the top of the figure should each have circles (corresponding to (2,2) counters).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Octant | a | a0a1a2 | b | (C+S)/2 | (C-S)/2 |
| 0 | 0 ≤ a < π/4 | 000 | a | 1/sqrt(2)sin(π/4-b) | 1/sqrt(2)cos(π/4-b) |
| 1 | π/4 ≤ a < 2π/4 | 001 | a - π/4 | -1/sqrt(2)sin(b) | 1/sqrt(2)cos(b) |
| 2 | 2π/4 ≤ a < 3π/4 | 010 | a-2π/4 | -1/sqrt(2)cos(π/4-b) | 1/sqrt(2)sin(π/4-b) |
| 3 | 3π/4 ≤ a < 4π/4 | 011 | a-3π/4 | -1/sqrt(2)cos(b) | -1/sqrt(2)sin(b) |
Problem 11.5: Change indexes on numbers from 2 to 2C, i.e., (0.0101)2C and (1.010)2C.
Problem 11.6:
The coefficient is a=(0.100-1)CSDC.
Problem 11.7:
... a=(1.11011)2C...
Problem 11.8:
All numbers are two's complement (index 2C).
Page 527, Problem 11.10 a):
...multiplications by -1 that may appear...
Problem 11.14: Change indexes from 2 to 2C and change 0.375 to -0.375.
Page 530, problem 11.25 (a): a=21(1+22(1+22))
should be a=2-1(1+2-2(1+2-2)).
Page 530, problem 11.28: R = Y + 350 Cr/256-175/256 G = Y - 86 Cb/256
- 178Cr/256 + 132/256 B = Y + 444Cb/256 - 222/256 Cb = (-44R -87G + 131B)/256
+ 128/256 Cr = (131R - 110G - 21B)/256 + 128/256
Page 545, heading 12.4: Change CONT to be in upper case only.
Page 545, last sentence in the first paragraph of section 12.4: The
required read and write frequency is 31 MHz should be The required read
and write frequency is 32 MHz.
Page 548, second line from bottom: ... with five 21-bit ... should be
... with six 21-bit ...
Page 548, last line: ... the other with ten 21-bit ... should be ...
the other with nine 21-bit ...
Page 549, first line: ... only five are used, ... should be ... only
six are used, ...
Page 549, fourth line: ... only 10 are used, ... should be ... only
9 are used, ...
Page 550, Figure 12.19: Add a thick horizontal line between the multiplier
block (overflow, sub, multiplier, add) and the two RAMs.
Page 551, line 3 at the sixth paragraph: ... dice per wafer ...
should be ... dies per wafer ...
Page 551, Equation 12.1: Replace Dice with Dies
Page 551, Equation 12.1: Replace -2 with -1
Page 552, line 2: ... the size of the dice. ... should be ... the size of the dies. ...
Page 552, line 2 at the fourth paragraph: ... when the dice. ...
should be ... when the dies. ...
Page 557: Change I2L to I2L
Page 561: Change Yields to Yield