Line Drawing Algorithms in Computer Graphics MCQ Questions and Answers
1. The principal goal of line drawing algorithms in raster graphics is to
A) Compute polygon normals
B) Fill polygons with color
C) Determine which pixels best approximate the continuous line
D) Compute lighting per vertex
Answer: C
2. Which algorithm is known for using floating-point arithmetic and incremental updates?
A) DDA (Digital Differential Analyzer)
B) Bresenham’s algorithm
C) Wu’s algorithm
D) Midpoint circle algorithm
Answer: A
3. Which algorithm is preferred when integer-only arithmetic is desired?
A) DDA
B) Bresenham’s line algorithm
C) Supersampling
D) Liang–Barsky
Answer: B
4. The DDA algorithm steps the major axis by 1 and increments the minor axis by
A) 0 or 1 randomly
B) The slope m (Δy/Δx) or its reciprocal
C) 2Δy − Δx
D) The decision parameter p
Answer: B
5. Bresenham’s algorithm decides the next pixel using a
A) Floating-point slope calculation each step
B) Lookup table of pixels
C) Decision (error) term updated by integer additions/subtractions
D) Random selection
Answer: C
6. In Bresenham’s algorithm for lines with 0 ≤ m ≤ 1, the initial decision parameter p₀ is often
A) 2Δx − Δy
B) 2Δy − Δx
C) Δy/Δx
D) m
Answer: B
7. For a line with slope greater than 1, standard practice in integer algorithms is to
A) Use DDA only
B) Swap x and y roles (steep-line handling)
C) Interchange the roles of x and y when stepping
D) Use Wu’s algorithm exclusively
Answer: C
8. Wu’s algorithm is primarily used for
A) Integer-only rasterization
B) Clipping lines
C) Anti-aliased line drawing using intensity interpolation
D) Filling triangles
Answer: C
9. The midpoint line algorithm chooses a pixel based on the sign of a function evaluated at
A) The endpoint only
B) The midpoint between two candidate pixels
C) The origin
D) The previous pixel only
Answer: B
10. Which algorithm is best suited when minimizing arithmetic per step is critical on integer hardware?
A) DDA
B) Bresenham’s algorithm
C) Wu’s algorithm
D) Liang–Barsky
Answer: B
11. The DDA algorithm requires computing the value of y incrementally using the slope; this can introduce rounding errors because it uses
A) Integer division only
B) Floating-point accumulation
C) XOR operations
D) Bit-shift only
Answer: B
12. Bresenham’s algorithm is particularly efficient because it replaces multiplications and divisions with
A) Trigonometric evaluations
B) Simple additions and subtractions (and bit shifts if optimized)
C) Floating-point operations
D) Square roots
Answer: B
13. When implementing Bresenham for negative slopes, one must
A) Reverse the display order only
B) Use Wu’s intensity scheme
C) Adjust increment/decrement directions for x and/or y
D) Use DDA instead
Answer: C
14. A common trick to handle all octants in integer line algorithms is to
A) Use floating arithmetic for all cases
B) Normalize line to the first octant by symmetry (swap/reflect coordinates)
C) Use Wu’s algorithm only for octant 1
D) Clip the line into octant-sized pieces
Answer: B
15. The decision parameter p in Bresenham’s algorithm is initially computed using Δx and Δy where Δx = x₂ − x₁ and Δy = y₂ − y₁. For a gentle slope (Δx > Δy), the main update uses additions of
A) Δx only
B) 2Δy and 2(Δy − Δx) depending on choice
C) Δy²
D) Δx + Δy
Answer: B
16. Which algorithm will generally produce smoother-looking lines without explicit anti-aliasing operations?
A) Bresenham integer-only output
B) Wu’s anti-aliased algorithm
C) Raw DDA without intensity control
D) Cohen–Sutherland clipping
Answer: B
17. When using DDA, the number of steps equals
A) Δy only
B) Δx + Δy
C) max(|Δx|, |Δy|)
D) min(|Δx|, |Δy|)
Answer: C
18. Which line-drawing algorithm can be extended most naturally to draw circles by symmetry and integer arithmetic?
A) DDA direct extension
B) Wu’s algorithm
C) Midpoint (Bresenham-like) circle algorithm
D) Liang–Barsky
Answer: C
19. Which algorithm uses intensity (alpha) values proportional to fractional coverage of pixel area?
A) Bresenham integer only
B) Wu’s line algorithm
C) Cohen–Sutherland
D) DDA with rounding only
Answer: B
20. In the context of line algorithms, “aliasing” refers to
A) Incorrect clipping results
B) Staircase effect due to discrete sampling of a continuous line
C) Incorrect color mapping
D) Lossy compression artifacts
Answer: B
21. The main computational cost per step of DDA compared to Bresenham is typically higher because DDA uses
A) No arithmetic at all
B) Floating point addition (and rounding) per step
C) Only bitwise operations
D) Table lookups only
Answer: B
22. Which algorithm is deterministic and guarantees symmetric results for octants when implemented with integer arithmetic and coordinate normalization?
A) DDA without adjustments
B) Bresenham’s algorithm with octant normalization
C) Wu’s algorithm without symmetry handling
D) Randomized rasterization
Answer: B
23. For high-quality antialiasing with GPU pipelines, modern approaches often prefer multisampling or shader-based techniques over per-line Wu’s algorithm because they integrate better with
A) Bresenham integer math
B) CPU-bound rasterization only
C) Fragment processing and multisample buffers on GPUs
D) DDA loops
Answer: C
24. Which algorithm is most sensitive to cumulative floating-point rounding errors on long lines?
A) Bresenham (integer)
B) DDA
C) Wu’s algorithm (uses fractional parts but less cumulative error due to per-step recomputation)
D) Midpoint integer algorithm
Answer: B
25. To draw a line from (2,3) to (10,7) using DDA, Δx = 8, Δy = 4, number of steps = max(8,4) = 8. The x increment is 1 and y increment per step is
A) 2
B) 4
C) 0.5
D) 1.5
Answer: C
26. In integer Bresenham’s algorithm, when the error term becomes ≥ 0 you typically
A) Do nothing to y
B) Increment y and adjust the error by adding 2(Δy − Δx)
C) Decrement x
D) Swap axes
Answer: B
27. The midpoint test function f(x,y) for a line implicitly evaluates which expression?
A) Distance squared only
B) A linear expression derived from the line equation (e.g., f(x,y) = Δy·x − Δx·y + constant)
C) A trigonometric function
D) Pixel intensity
Answer: B
28. Which line algorithm is simplest to implement but least efficient due to floating arithmetic?
A) Simple incremental DDA
B) Bresenham optimized version
C) Wu’s algorithm
D) Cohen–Sutherland
Answer: A
29. Bresenham’s algorithm can be modified to draw thick lines by
A) Using floating point only
B) Drawing multiple offset parallel lines or using a filled rectangle approach
C) Changing Δx and Δy only
D) Using Wu’s intensity interpolation only
Answer: B
30. For a vertical line (Δx = 0), most line algorithms should handle it by
A) Calculating slope m = Δy/0 and proceeding normally
B) Special-case: increment y only (step through y)
C) Using Wu’s algorithm only
D) Rotating the line 45° first
Answer: B
31. Which statement about Bresenham’s algorithm is FALSE?
A) It uses integer operations for the decision variable
B) It can be implemented using only additions and subtractions
C) It requires dividing Δy by Δx at every step
D) It can handle various octants by symmetry transforms
Answer: C
32. Which algorithm requires per-step computation of fractional parts for anti-aliasing intensity?
A) DDA integer-only variant
B) Wu’s algorithm
C) Cohen–Sutherland
D) Midpoint integer-only algorithm
Answer: B
33. A common optimization for Bresenham’s algorithm on hardware is to replace multiplications by
A) Floating point divisions
B) Left shifts and additions (where applicable)
C) Trigonometric LUTs
D) Square root approximations
Answer: B
34. Which algorithm is often taught as “Bresenham for circles” analog because both use the same principle?
A) DDA circle algorithm
B) Wu’s circle algorithm
C) Midpoint circle algorithm (Bresenham-like)
D) Liang–Barsky circle clipping
Answer: C
35. For steep lines (|m| > 1), a robust implementation strategy is to swap x and y so that the primary loop increments along
A) The minor axis always
B) The axis with the larger absolute extent (the major axis)
C) Only negative coordinates
D) The origin first
Answer: B
36. When implementing antialiasing for lines, the aim is to reduce visual artifacts by approximating the continuous line’s coverage of each pixel and adjusting
A) The pixel coordinates only
B) The pixel intensity (brightness/alpha)
C) The line endpoints only
D) The clipping region only
Answer: B
37. Which algorithm can be described as evaluating a polynomial (linear) function sign at midpoints to choose pixels?
A) DDA floating accumulation
B) Midpoint algorithm
C) Wu’s anti-aliasing
D) Liang–Barsky parametric method
Answer: B
38. The error accumulation in DDA can be limited by periodically reinitializing coordinates using exact arithmetic to avoid drift — this technique is called
A) Anti-aliasing
B) Re-centering or periodic correction
C) Floating reset
D) Bit shifting
Answer: B
39. The reason Bresenham’s algorithm is often faster than naive DDA on integer-only processors is that it avoids
A) All additions
B) Floating-point operations and expensive rounding
C) Drawing pixels
D) Using decision parameters
Answer: B
40. Which technique provides higher quality but is more computationally expensive: supersampling or simple Wu anti-aliasing?
A) Wu’s always more expensive
B) Supersampling (sampling at higher resolution then downsampling)
C) Bresenham’s integer-only pure output
D) DDA with rounding only
Answer: B
41. The walking algorithm that updates an error term by adding 2Δy each step is part of which family?
A) Wu family
B) Bresenham/midpoint family
C) Liang–Barsky family
D) DDA floating family
Answer: B
42. Line clipping should usually be performed
A) After rasterization for simplicity
B) Before scan conversion to minimize work
C) During anti-aliasing only
D) Never — draw whole infinite line
Answer: B
43. For hardware rasterizers, which representation of lines is typically used to feed the scan conversion stage?
A) Pixel list only
B) Vertex endpoints with edge equations for interpolation
C) Color palette indices
D) Compressed run-length encoding
Answer: B
44. The Bresenham algorithm generalization to higher dimensions (e.g., 3D line voxel traversal) is commonly known as
A) DDA 3D only
B) 3D Bresenham or voxel traversal (e.g., Amanatides–Woo for rays)
C) Wu 3D algorithm
D) Liang–Barsky 3D only
Answer: B
45. When drawing a line with slope exactly 1 from (0,0) to (n,n) using Bresenham, which sequence of pixels is generated?
A) Alternating x and y only
B) Only y increments, x constant
C) Both x and y increment at each step producing pixels (0,0),(1,1),(2,2),…
D) None of the above
Answer: C
46. In Wu’s algorithm the intensity assigned to adjacent pixels is proportional to
A) Square of fractional distance
B) The fractional part of the exact y (or x) coordinate (distance from pixel center)
C) Product of Δx and Δy
D) Sign of decision variable only
Answer: B
47. A major drawback of Wu’s algorithm compared to Bresenham is that Wu’s requires
A) No arithmetic
B) Per-pixel intensity calculation and gamma-aware blending
C) Integer-only math
D) Region codes
Answer: B
48. Which algorithm inherently supports symmetrical line drawing about octant boundaries with minimal extra logic?
A) Straight DDA without adjustments
B) Bresenham with coordinate normalization
C) Wu’s only
D) Liang–Barsky only
Answer: B
49. Which method is best described as parametric and useful in clipping and intersection tests as well as line drawing?
A) Bresenham integer loop
B) Parametric representation P(t) = P₀ + t(P₁ − P₀)
C) Wu intensity interpolation
D) Octant mapping only
Answer: B
50. If Δx = 0 and Δy ≠ 0, Bresenham’s algorithm degenerates to a simple loop incrementing y — this case is handled by treating line as
A) Diagonal line only
B) Vertical special-case
C) Slope 1 case
D) Clipped away
Answer: B
51. Which algorithm is most appropriate when implementing thick, stylized strokes (variable width) and brush effects?
A) Bresenham bare pixels
B) DDA integer-only
C) Polygonal or area-based stroke rendering and textured brush masks
D) Cohen–Sutherland clipping
Answer: C
52. The midpoint line algorithm and Bresenham’s algorithm are related because both use
A) Floating slopes per step
B) A decision parameter evaluating closeness between candidate pixels
C) Supersampling inherently
D) Parametric t updates only
Answer: B
53. To reduce banding or quantization artifacts in Wu’s algorithm, one should adjust intensities using
A) Integer flooring only
B) Proper gamma correction before blending
C) Random dithering only
D) Clipping to white only
Answer: B
54. Which algorithm naturally provides sub-pixel information useful for antialiasing?
A) Integer Bresenham only
B) Wu’s algorithm (uses fractional parts of coordinates)
C) Cohen–Sutherland
D) Octant mapping only
Answer: B
55. The incremental update in Bresenham’s loop uses the previous error to compute the next without division. This is a demonstration of the technique called
A) Floating-point rolling
B) Incremental (recursive) computation
C) Direct evaluation each step
D) Recursive anti-aliasing
Answer: B
56. Which of the following is an advantage of using midpoint/Bresenham-style integer algorithms for embedded graphics hardware?
A) Use of multiply/divide pipelines
B) Low computational overhead and predictable timing
C) High-quality anti-aliasing without extra steps
D) No need to handle octants
Answer: B
57. When converting a generalized line-drawing algorithm to draw in all directions, a standard technique is to perform coordinate transforms: reflect, swap, and translate — these are examples of using
A) Randomization
B) Symmetry and affine transformations
C) Clipping only
D) Texture mapping
Answer: B
58. The accuracy of Bresenham’s algorithm is due to exact arithmetic on integer grid; however, when plotting on non-square pixels (pixel aspect ≠ 1), you must
A) Use Wu only
B) Adjust Δx and Δy by pixel aspect ratio or rescale coordinates
C) Use DDA only
D) Ignore aspect and proceed
Answer: B
59. To draw an anti-aliased line using a shader on modern GPUs, you would typically compute the distance from the pixel center to the ideal line and use that to determine
A) The next vertex only
B) The alpha/coverage value for blending the line color
C) The slope to pass to DDA
D) The clipping code only
Answer: B
60. The classic Bresenham derivation is based on comparing distances to the ideal line of two candidate pixels; those distances are proportional to evaluating the line function at a sample point — this sample point is usually chosen as
A) The left endpoint always
B) The midpoint between the two candidate pixel centers
C) The right endpoint always
D) The origin
Answer: B
61. For high-performance software rasterizers, a common micro-optimization for integer Bresenham is to replace multiplication by repeated addition or precomputed difference terms — this is known as
A) Supersampling
B) Incremental differencing
C) Floating normalization
D) Pixel dithering
Answer: B
62. Which algorithm will produce exact symmetry for lines drawn from (x1,y1)→(x2,y2) and (x2,y2)→(x1,y1) if implemented correctly with coordinate normalization?
A) Naive DDA with accumulation only
B) Bresenham with proper octant handling
C) Wu without symmetry handling
D) Random dithering
Answer: B
63. When the major axis is x, the stepping rate is 1 pixel per step in x; when the major axis is y (steep lines), after swapping, the stepping rate corresponds to
A) 1 pixel per step in both axes simultaneously
B) 1 pixel per step in the swapped major axis (now x after swap)
C) Floating increments in both axes
D) No stepping needed
Answer: B
64. Which algorithm is useful when you want to compute all pixels for a line without worrying about slope sign or octant logic (but possibly slower)?
A) Optimized Bresenham only
B) Symmetric DDA that uses max(|Δx|,|Δy|) steps and computes x and y increments
C) Wu only
D) Midpoint only
Answer: B
65. The main difference between midpoint and Bresenham’s formulations is primarily the perspective in deriving the decision variable; both yield similar integer-step algorithms — midpoint evaluates the line function at the midpoint, whereas Bresenham’s original derivation focuses on
A) Floating slope integration
B) Difference of distances or error accumulation
C) Parametric t extremes
D) Anti-aliasing weight
Answer: B
66. For machine architectures where floating-point is cheap and vectorized, DDA may be competitive because it offers
A) Integer-only speedup always
B) Straightforward SIMD vectorization of coordinate increments
C) No need to normalize coordinates
D) Built-in anti-aliasing
Answer: B
67. In Bresenham’s algorithm optimized for drawing many adjacent lines, precomputing which quantities can reduce per-line overhead?
A) Colors only
B) 2Δy and 2(Δy − Δx) or their equivalents for given Δx, Δy
C) Vertex normals
D) Texture coordinates
Answer: B
68. The line drawing problem for GPU scan conversion is usually transformed into edge functions and interpolants so that per-pixel tests reduce to evaluating a signed linear function; this approach generalizes Bresenham style to
A) DDA only
B) Arbitrary polygon edge rasterization with edge equations
C) Wu’s anti-aliasing only
D) Clipping algorithms only
Answer: B
69. A variant of Bresenham’s algorithm useful for rasterizing thick stroked lines by filling spans is to treat the stroke as a polygon and rasterize its edges using
A) DDA only
B) Standard polygon fill and edge-walking rasterization
C) Wu’s scalar blending only
D) Cohen–Sutherland clipping only
Answer: B
70. The major benefit of the parametric form P(t)=P₀ + t(P₁ − P₀) for line drawing is that it allows straightforward extension to clipping, interpolation, and
A) Integer-only math only
B) Subdivision, interpolation of attributes (color, texture coord) along the line
C) Anti-aliasing automatically
D) Eliminating error terms
Answer: B
71. When Δx and Δy are both large, integer algorithms still run in O(n) where n is the number of steps — the constant factors matter for performance. The complexity O(n) refers to n ≈
A) Δx × Δy
B) max(|Δx|,|Δy|)
C) min(|Δx|,|Δy|)
D) Δx + Δy squared
Answer: B
72. Which of the following is true about Bresenham vs. Wu for visible-quality tradeoffs?
A) Bresenham always looks better than Wu
B) Wu gives smoother lines but costs more per pixel due to intensity calculations
C) Wu is integer-only and cheapest
D) Bresenham does anti-aliasing automatically
Answer: B
73. The “error” variable in Bresenham’s algorithm represents
A) The pixel brightness value
B) A scaled measure of distance from the true line to the plotted pixel center
C) The clipping code
D) The texture index
Answer: B
74. Which approach can reduce aliasing by changing sampling patterns rather than per-pixel intensity blending?
A) Wu’s per-pixel intensity only
B) Supersampling / multisampling (render at higher resolution and downsample)
C) Bresenham integer stepping
D) Decision variable doubling only
Answer: B
75. When plotting anti-aliased lines on an LCD with non-square pixels, what must be considered?
A) Nothing; same as CRT
B) Pixel aspect ratio — adjust distance calculations or rescale coordinates
C) Only color depth matters
D) Use Bresenham only
Answer: B
76. The Bresenham algorithm can be extended to draw parametric curves by plotting successive approximated line segments; this is computationally feasible but may require
A) No changes at all
B) Subdivision and many small integer line draws
C) Wu’s algorithm only
D) Liang–Barsky only
Answer: B
77. If you need perfectly symmetric lines about both endpoints, a reliable approach is to compute pixel positions by rounding the parametric sample at t = k / n for k = 0..n instead of accumulating floating error — this technique is an instance of
A) Incremental DDA in float form only
B) Uniform parametric sampling with rounding (avoids cumulative drift)
C) Wu’s intensity only
D) Random jittering
Answer: B
78. Which algorithm naturally lends itself to hardware acceleration using simple integer adders and shift registers?
A) DDA float accumulator
B) Bresenham’s integer decision algorithm
C) Wu anti-aliased intensity computation
D) Liang–Barsky parametric clipping
Answer: B
79. The principal weakness of naive Bresenham when implemented carelessly on modern CPUs is that it does not exploit
A) Integer arithmetic
B) SIMD/vector instructions for parallel pixel generation
C) Predictable branches
D) Clipping functionality
Answer: B
80. In parametric sampling P(t) = P₀ + t(P₁ − P₀), stepping t uniformly ensures even sampling along the line in Euclidean distance only if t is stepped by
A) 1/n always
B) Step size proportional to 1 / max(|Δx|, |Δy|) (or by actual arc-length increments for perfect even spacing)
C) Δx only
D) Δy only
Answer: B
81. Which of the following is TRUE about midpoint and Bresenham derivations?
A) Midpoint uses non-incremental floating recalculations each step
B) Both can be formulated to produce identical pixel choices by appropriate scaling
C) Bresenham always plots anti-aliased intensities
D) Midpoint cannot be implemented using integer math
Answer: B
82. When implementing a high-quality line renderer for vector graphics editors (zoomable), it’s advantageous to use vector representation and only rasterize at target resolution because this preserves
A) Integer arithmetic only
B) Crispness at any zoom level and avoids re-computing from low-res bitmaps
C) Anti-aliasing automatically
D) Clipping enforcement only
Answer: B
83. For lines drawn with variable color or texture along their length, the parametric form is useful because it allows interpolating attributes per step by linear interpolation of
A) Texture only
B) Any vertex attributes (color, texture coords, normals) as functions of t
C) Only color with integer math
D) Clipping codes only
Answer: B
84. A typical integer Bresenham implementation uses which initializations before the main loop?
A) Only pixel plotting function pointer
B) Compute Δx, Δy, initial decision p, and the increments for p on pixel choice
C) Compute antialiasing weights only
D) Build the entire pixel array first
Answer: B
85. Which choice best describes the role of symmetry transforms (swap, reflect) in line drawing libraries?
A) To change color palette only
B) To map all lines into a single canonical octant so a single algorithm can be used
C) To accelerate clipping only
D) To avoid integer arithmetic entirely
Answer: B
86. In a multi-threaded software renderer, plotting long lines could be parallelized by dividing the line into segments and processing them concurrently; this requires careful handling to ensure
A) Overlapping of buffers always
B) Deterministic ordering and no shared-state write conflicts
C) Using Wu exclusively
D) Dynamic deletion of pixels
Answer: B
87. For accurate line thickness rendering in vector-to-raster conversion, a robust strategy is to treat the stroke as a polygon swept by the line and then rasterize it using edge fill rules; this is preferred to naive multiple-offset Bresenham because it correctly handles
A) Clipping only
B) Joins, caps, and miter limits for stroke geometry
C) Anti-aliasing only
D) Octant transforms only
Answer: B
88. In anti-aliased line rendering, pre-multiplied alpha blending is often recommended because it
A) Avoids integer arithmetic
B) Simplifies compositing math and avoids certain artifacts when blending partially-covered pixels
C) Increases error accumulation
D) Requires more gamma correction only
Answer: B
89. When drawing lines on displays with known non-linear gamma, alpha-weighted intensities should be corrected by converting linear intensities through gamma encoding before writing to the framebuffer to avoid
A) Integer overflow
B) Visible brightness mismatches and incorrect blending
C) Faster performance
D) Clipping failures
Answer: B
90. For lines that must be drawn repeatedly with different transforms, it is efficient to store the original vector primitives and re-run the chosen line algorithm for each transform instead of storing raster bitmaps because this approach preserves
A) Lower memory only
B) Resolution independence and eliminates blurring on scaling/rotation
C) Integer-only math only
D) Anti-aliasing automatically
Answer: B
91. Which of the following statements about Bresenham and midpoint algorithms is correct?
A) They are unrelated and solve different problems
B) They are mathematically equivalent when decision variables are scaled appropriately
C) Only Bresenham can be used for circles
D) Midpoint always uses floating arithmetic and is slower
Answer: B
92. In practice, to speed up line drawing for many short segments, a renderer may batch operations into scanline spans; this converts per-pixel plotting into per-span filling and benefits from
A) Increased branching per pixel
B) Cache locality and span-fill optimizations
C) More floating multiplications
D) Breaking clipping rules
Answer: B
93. Which of the following is a potential problem when naive integer Bresenham is ported to platforms with different integer sizes?
A) Floating-point overflow only
B) Overflow of intermediate decision variables if not sized appropriately
C) Lack of gamma correction
D) DDA incompatibility only
Answer: B
94. The Bresenham algorithm for circles and ellipses is widely used because it produces symmetric curves using only integer arithmetic and uses which geometric property?
A) Parametric arc-sampling only
B) Symmetry of octants and incremental decision variables
C) Floating interpolation only
D) Edge equations only
Answer: B
95. If you observe slight bias (systematic shift) when plotting a line with DDA, the most likely cause is
A) Wrong color palette
B) Cumulative rounding error in incremental floating accumulation
C) Incorrect clipping algorithm only
D) Bresenham implementation bug only
Answer: B
96. For very long lines across large coordinate spaces, an algorithm that recomputes absolute coordinates from parametric t occasionally is preferred to pure accumulation to avoid
A) Antialiasing automatically
B) Accumulated error due to repeated additions (drift)
C) Deciding octant only
D) Using Wu’s method
Answer: B
97. A hardware rasterizer implementing Bresenham-style edges in FPGA would most benefit from which feature?
A) Floating-point unit
B) Wide integer adders and pipeline-friendly incremental updates
C) Large caches only
D) Shader cores only
Answer: B
98. Which approach is best when you need both speed and high visual quality on modern GPUs?
A) Pure integer Bresenham on CPU and blit result
B) Use GPU triangle/edge rasterization with MSAA or shader-based anti-aliasing
C) Wu’s algorithm on CPU only
D) DDA on single core only
Answer: B
99. In some implementations, Bresenham’s decision variable is scaled by 2 to avoid fractional halves; scaling decision terms preserves correctness because the sign tests are invariant under positive scaling. This is an example of
A) Floating-point conversion
B) Scale invariance of comparison-based algorithms
C) Bit-rot prevention
D) Gamma correction
Answer: B
100. Which property is most desirable for a general-purpose line drawing routine in a graphics library?
A) Only integer arithmetic regardless of context
B) Correctness for all octants, reasonable performance, and predictable visual quality (and optional anti-aliasing modes)
C) Only anti-aliased output irrespective of speed
D) Dependence on specific display hardware only
Answer: B