232
|
1 |
// Part 2 about a "Compiler" for the Brainf*** language
|
|
2 |
//======================================================
|
|
3 |
|
404
|
4 |
object M5b {
|
232
|
5 |
|
|
6 |
// !!! Copy any function you need from file bf.scala !!!
|
|
7 |
//
|
|
8 |
// If you need any auxiliary function, feel free to
|
|
9 |
// implement it, but do not make any changes to the
|
|
10 |
// templates below.
|
|
11 |
|
|
12 |
|
|
13 |
def time_needed[T](n: Int, code: => T) = {
|
|
14 |
val start = System.nanoTime()
|
|
15 |
for (i <- 0 until n) code
|
|
16 |
val end = System.nanoTime()
|
|
17 |
(end - start)/(n * 1.0e9)
|
|
18 |
}
|
|
19 |
|
|
20 |
type Mem = Map[Int, Int]
|
|
21 |
|
|
22 |
|
|
23 |
import io.Source
|
|
24 |
import scala.util._
|
|
25 |
|
|
26 |
def load_bff(name: String) : String =
|
|
27 |
Try(Source.fromFile(name)("ISO-8859-1").mkString).getOrElse("")
|
|
28 |
|
|
29 |
def sread(mem: Mem, mp: Int) : Int =
|
|
30 |
mem.getOrElse(mp, 0)
|
|
31 |
|
|
32 |
def write(mem: Mem, mp: Int, v: Int) : Mem =
|
|
33 |
mem.updated(mp, v)
|
|
34 |
|
|
35 |
def jumpRight(prog: String, pc: Int, level: Int) : Int = {
|
|
36 |
if (prog.length <= pc) pc
|
|
37 |
else (prog(pc), level) match {
|
|
38 |
case (']', 0) => pc + 1
|
|
39 |
case (']', l) => jumpRight(prog, pc + 1, l - 1)
|
|
40 |
case ('[', l) => jumpRight(prog, pc + 1, l + 1)
|
|
41 |
case (_, l) => jumpRight(prog, pc + 1, l)
|
|
42 |
}
|
|
43 |
}
|
|
44 |
|
|
45 |
def jumpLeft(prog: String, pc: Int, level: Int) : Int = {
|
|
46 |
if (pc < 0) pc
|
|
47 |
else (prog(pc), level) match {
|
|
48 |
case ('[', 0) => pc + 1
|
|
49 |
case ('[', l) => jumpLeft(prog, pc - 1, l - 1)
|
|
50 |
case (']', l) => jumpLeft(prog, pc - 1, l + 1)
|
|
51 |
case (_, l) => jumpLeft(prog, pc - 1, l)
|
|
52 |
}
|
|
53 |
}
|
|
54 |
|
|
55 |
def compute(prog: String, pc: Int, mp: Int, mem: Mem) : Mem = {
|
|
56 |
if (0 <= pc && pc < prog.length) {
|
|
57 |
val (new_pc, new_mp, new_mem) = prog(pc) match {
|
|
58 |
case '>' => (pc + 1, mp + 1, mem)
|
|
59 |
case '<' => (pc + 1, mp - 1, mem)
|
|
60 |
case '+' => (pc + 1, mp, write(mem, mp, sread(mem, mp) + 1))
|
|
61 |
case '-' => (pc + 1, mp, write(mem, mp, sread(mem, mp) - 1))
|
|
62 |
case '.' => { print(sread(mem, mp).toChar); (pc + 1, mp, mem) }
|
|
63 |
case '[' =>
|
|
64 |
if (sread(mem, mp) == 0) (jumpRight(prog, pc + 1, 0), mp, mem) else (pc + 1, mp, mem)
|
|
65 |
case ']' =>
|
|
66 |
if (sread(mem, mp) != 0) (jumpLeft(prog, pc - 1, 0), mp, mem) else (pc + 1, mp, mem)
|
|
67 |
case _ => (pc + 1, mp, mem)
|
|
68 |
}
|
|
69 |
compute(prog, new_pc, new_mp, new_mem)
|
|
70 |
}
|
|
71 |
else mem
|
|
72 |
}
|
|
73 |
|
|
74 |
def run(prog: String, m: Mem = Map()) = compute(prog, 0, 0, m)
|
|
75 |
|
|
76 |
|
233
|
77 |
// The baseline to what we can compare our "compiler"
|
|
78 |
// implemented below. It should require something like
|
|
79 |
// 60 seconds for the calculation on my laptop
|
232
|
80 |
//
|
|
81 |
//time_needed(1, run(load_bff("benchmark.bf")))
|
|
82 |
|
|
83 |
|
|
84 |
|
233
|
85 |
// DEBUGGING INFORMATION!!!
|
|
86 |
//
|
|
87 |
// Compiler, even real ones, are fiedishly difficult to get
|
|
88 |
// to prduce correct code. The point is that for example for
|
|
89 |
// the sierpinski program, they need to still generate code
|
|
90 |
// that displays such a triangle. If yes, then one usually
|
|
91 |
// can take comfort that all is well. If not, then something
|
|
92 |
// went wrong during the optimisations.
|
|
93 |
|
|
94 |
|
232
|
95 |
|
|
96 |
// (5) Write a function jtable that precomputes the "jump
|
|
97 |
// table" for a bf-program. This function takes a bf-program
|
|
98 |
// as an argument and Returns a Map[Int, Int]. The
|
|
99 |
// purpose of this map is to record the information
|
233
|
100 |
// that given on the position pc is a '[' or a ']',
|
|
101 |
// then to which pc-position do we need to jump next?
|
232
|
102 |
//
|
|
103 |
// For example for the program
|
|
104 |
//
|
|
105 |
// "+++++[->++++++++++<]>--<+++[->>++++++++++<<]>>++<<----------[+>.>.<+<]"
|
|
106 |
//
|
|
107 |
// we obtain the map
|
|
108 |
//
|
|
109 |
// Map(69 -> 61, 5 -> 20, 60 -> 70, 27 -> 44, 43 -> 28, 19 -> 6)
|
|
110 |
//
|
|
111 |
// This states that for the '[' on position 5, we need to
|
|
112 |
// jump to position 20, which is just after the corresponding ']'.
|
|
113 |
// Similarly, for the ']' on position 19, we need to jump to
|
|
114 |
// position 6, which is just after the '[' on position 5, and so
|
233
|
115 |
// on. The idea is to not calculate this information each time
|
|
116 |
// we hit a bracket, but just look up this information in the
|
|
117 |
// jtable. You can use the jumpLeft and jumpRight functions
|
|
118 |
// from Part 1 for calculating the jtable.
|
232
|
119 |
//
|
233
|
120 |
// Then adapt the compute and run functions from Part 1 in order
|
|
121 |
// to take advantage of the information stored in the jtable.
|
|
122 |
// This means whenever jumpLeft and jumpRight was called previously,
|
|
123 |
// you should look up the jump address in the jtable.
|
232
|
124 |
|
|
125 |
|
|
126 |
def jtable(pg: String) : Map[Int, Int] =
|
|
127 |
(0 until pg.length).collect { pc => pg(pc) match {
|
|
128 |
case '[' => (pc -> jumpRight(pg, pc + 1, 0))
|
|
129 |
case ']' => (pc -> jumpLeft(pg, pc - 1, 0))
|
|
130 |
}}.toMap
|
|
131 |
|
|
132 |
|
|
133 |
// testcase
|
|
134 |
// jtable("""+++++[->++++++++++<]>--<+++[->>++++++++++<<]>>++<<----------[+>.>.<+<]""")
|
|
135 |
// => Map(69 -> 61, 5 -> 20, 60 -> 70, 27 -> 44, 43 -> 28, 19 -> 6)
|
|
136 |
|
|
137 |
|
|
138 |
def compute2(pg: String, tb: Map[Int, Int], pc: Int, mp: Int, mem: Mem) : Mem = {
|
|
139 |
if (0 <= pc && pc < pg.length) {
|
|
140 |
val (new_pc, new_mp, new_mem) = pg(pc) match {
|
|
141 |
case '>' => (pc + 1, mp + 1, mem)
|
|
142 |
case '<' => (pc + 1, mp - 1, mem)
|
|
143 |
case '+' => (pc + 1, mp, write(mem, mp, sread(mem, mp) + 1))
|
|
144 |
case '-' => (pc + 1, mp, write(mem, mp, sread(mem, mp) - 1))
|
|
145 |
case '.' => { print(sread(mem, mp).toChar); (pc + 1, mp, mem) }
|
336
|
146 |
case '[' =>
|
232
|
147 |
if (sread(mem, mp) == 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
148 |
case ']' =>
|
|
149 |
if (sread(mem, mp) != 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
150 |
case _ => (pc + 1, mp, mem)
|
|
151 |
}
|
|
152 |
compute2(pg, tb, new_pc, new_mp, new_mem)
|
|
153 |
}
|
|
154 |
else mem
|
|
155 |
}
|
|
156 |
|
|
157 |
|
|
158 |
def run2(pg: String, m: Mem = Map()) =
|
|
159 |
compute2(pg, jtable(pg), 0, 0, m)
|
|
160 |
|
|
161 |
//time_needed(1, run2(load_bff("benchmark.bf")))
|
|
162 |
|
|
163 |
|
233
|
164 |
|
232
|
165 |
// (6) Write a function optimise which deletes "dead code" (everything
|
|
166 |
// that is not a bf-command) and also replaces substrings of the form
|
233
|
167 |
// [-] by a new command 0. The idea is that the loop [-] just resets the
|
232
|
168 |
// memory at the current location to 0. In the compute3 and run3 functions
|
233
|
169 |
// below you implement this command by writing the number 0 to mem(mp),
|
|
170 |
// that is write(mem, mp, 0).
|
232
|
171 |
//
|
|
172 |
// The easiest way to modify a string in this way is to use the regular
|
399
|
173 |
// expression """[^<>+-.\[\]""", which recognises everything that is
|
232
|
174 |
// not a bf-command and replace it by the empty string. Similarly the
|
|
175 |
// regular expression """\[-\]""" finds all occurences of [-] and
|
|
176 |
// by using the Scala method .replaceAll you can repplace it with the
|
233
|
177 |
// string "0" standing for the new bf-command.
|
232
|
178 |
|
337
|
179 |
def optimise(s: String) : String = {
|
399
|
180 |
s.replaceAll("""[^<>+-.\[\]]""","")
|
337
|
181 |
.replaceAll("""\[-\]""", "0")
|
|
182 |
}
|
232
|
183 |
|
233
|
184 |
|
232
|
185 |
def compute3(pg: String, tb: Map[Int, Int], pc: Int, mp: Int, mem: Mem) : Mem = {
|
|
186 |
if (0 <= pc && pc < pg.length) {
|
|
187 |
val (new_pc, new_mp, new_mem) = pg(pc) match {
|
|
188 |
case '0' => (pc + 1, mp, write(mem, mp, 0))
|
|
189 |
case '>' => (pc + 1, mp + 1, mem)
|
|
190 |
case '<' => (pc + 1, mp - 1, mem)
|
|
191 |
case '+' => (pc + 1, mp, write(mem, mp, sread(mem, mp) + 1))
|
|
192 |
case '-' => (pc + 1, mp, write(mem, mp, sread(mem, mp) - 1))
|
|
193 |
case '.' => { print(sread(mem, mp).toChar); (pc + 1, mp, mem) }
|
|
194 |
case '[' =>
|
|
195 |
if (sread(mem, mp) == 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
196 |
case ']' =>
|
|
197 |
if (sread(mem, mp) != 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
198 |
case _ => (pc + 1, mp, mem)
|
|
199 |
}
|
|
200 |
compute3(pg, tb, new_pc, new_mp, new_mem)
|
|
201 |
}
|
|
202 |
else mem
|
|
203 |
}
|
|
204 |
|
|
205 |
def run3(pg: String, m: Mem = Map()) = {
|
|
206 |
val pg_opt = optimise(pg)
|
|
207 |
compute3(pg_opt, jtable(pg_opt), 0, 0, m)
|
|
208 |
}
|
|
209 |
|
|
210 |
|
233
|
211 |
// testcases
|
|
212 |
|
337
|
213 |
//println(optimise(load_bff("collatz.bf")))
|
233
|
214 |
//optimise(load_bff("benchmark.bf")) // should have inserted 0's
|
463
|
215 |
//optimise(load_bff("mandelbrot.bf")).length // => 11203
|
233
|
216 |
|
|
217 |
//time_needed(1, run3(load_bff("benchmark.bf")))
|
420
|
218 |
//time_needed(1, run3(load_bff("mandelbrot.bf")))
|
233
|
219 |
|
232
|
220 |
|
|
221 |
|
233
|
222 |
// (7) Write a function combine which replaces sequences
|
|
223 |
// of repated increment and decrement commands by appropriate
|
|
224 |
// two-character commands. For example for sequences of +
|
|
225 |
//
|
|
226 |
// orig bf-cmds | replacement
|
|
227 |
// ------------------------------
|
|
228 |
// + | +A
|
|
229 |
// ++ | +B
|
|
230 |
// +++ | +C
|
|
231 |
// |
|
|
232 |
// ... |
|
|
233 |
// |
|
|
234 |
// +++....+++ | +Z
|
|
235 |
// (where length = 26)
|
|
236 |
//
|
|
237 |
// Similar for the bf-command -, > and <. All other commands should
|
|
238 |
// be unaffected by this change.
|
|
239 |
//
|
|
240 |
// Adapt the compute4 and run4 functions such that they can deal
|
|
241 |
// appropriately with such two-character commands.
|
232
|
242 |
|
|
243 |
def splice(cs: List[Char], acc: List[(Char, Int)]) : List[(Char, Int)] = (cs, acc) match {
|
|
244 |
case (Nil, acc) => acc
|
|
245 |
case ('[' :: cs, acc) => splice(cs, ('[', 1) :: acc)
|
|
246 |
case (']' :: cs, acc) => splice(cs, (']', 1) :: acc)
|
|
247 |
case ('.' :: cs, acc) => splice(cs, ('.', 1) :: acc)
|
|
248 |
case ('0' :: cs, acc) => splice(cs, ('0', 1) :: acc)
|
|
249 |
case (c :: cs, Nil) => splice(cs, List((c, 1)))
|
|
250 |
case (c :: cs, (d, n) :: acc) =>
|
|
251 |
if (c == d && n < 26) splice(cs, (c, n + 1) :: acc)
|
|
252 |
else splice(cs, (c, 1) :: (d, n) :: acc)
|
|
253 |
}
|
|
254 |
|
|
255 |
def spl(s: String) = splice(s.toList, Nil).reverse
|
|
256 |
|
234
|
257 |
//spl(load_bff("benchmark.bf"))
|
232
|
258 |
|
233
|
259 |
def combine(s: String) : String = {
|
|
260 |
(for ((c, n) <- spl(s)) yield c match {
|
232
|
261 |
case '>' => List('>', (n + '@').toChar)
|
|
262 |
case '<' => List('<', (n + '@').toChar)
|
|
263 |
case '+' => List('+', (n + '@').toChar)
|
|
264 |
case '-' => List('-', (n + '@').toChar)
|
|
265 |
case _ => List(c)
|
|
266 |
}).flatten.mkString
|
|
267 |
}
|
|
268 |
|
|
269 |
|
234
|
270 |
//combine(load_bff("benchmark.bf"))
|
232
|
271 |
|
|
272 |
def compute4(pg: String, tb: Map[Int, Int], pc: Int, mp: Int, mem: Mem) : Mem = {
|
|
273 |
if (0 <= pc && pc < pg.length) {
|
|
274 |
val (new_pc, new_mp, new_mem) = pg(pc) match {
|
|
275 |
case '0' => (pc + 1, mp, write(mem, mp, 0))
|
|
276 |
case '>' => (pc + 2, mp + (pg(pc + 1) - '@'), mem)
|
|
277 |
case '<' => (pc + 2, mp - (pg(pc + 1) - '@'), mem)
|
|
278 |
case '+' => (pc + 2, mp, write(mem, mp, sread(mem, mp) + (pg(pc + 1) - '@')))
|
|
279 |
case '-' => (pc + 2, mp, write(mem, mp, sread(mem, mp) - (pg(pc + 1) - '@')))
|
|
280 |
case '.' => { print(sread(mem, mp).toChar); (pc + 1, mp, mem) }
|
336
|
281 |
case '[' =>
|
232
|
282 |
if (sread(mem, mp) == 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
283 |
case ']' =>
|
|
284 |
if (sread(mem, mp) != 0) (tb(pc), mp, mem) else (pc + 1, mp, mem)
|
|
285 |
case _ => (pc + 1, mp, mem)
|
|
286 |
}
|
|
287 |
compute4(pg, tb, new_pc, new_mp, new_mem)
|
|
288 |
}
|
|
289 |
else mem
|
|
290 |
}
|
|
291 |
|
|
292 |
def run4(pg: String, m: Mem = Map()) = {
|
233
|
293 |
val pg_opt = combine(optimise(pg))
|
232
|
294 |
compute4(pg_opt, jtable(pg_opt), 0, 0, m)
|
|
295 |
}
|
|
296 |
|
234
|
297 |
// testcases
|
384
|
298 |
//println(combine(optimise(load_bff("mandelbrot.bf").drop(123))))
|
337
|
299 |
|
234
|
300 |
//combine(optimise(load_bff("benchmark.bf"))) // => """>A+B[<A+M>A-A]<A[[....."""
|
233
|
301 |
|
232
|
302 |
//time_needed(1, run4(load_bff("benchmark.bf")))
|
233
|
303 |
|
|
304 |
//time_needed(1, run(load_bff("sierpinski.bf")))
|
|
305 |
//time_needed(1, run4(load_bff("sierpinski.bf")))
|
|
306 |
|
290
|
307 |
//println(time_needed(1, run4(load_bff("mandelbrot.bf"))))
|
232
|
308 |
|
|
309 |
|
384
|
310 |
|
|
311 |
|
|
312 |
|
290
|
313 |
}
|
336
|
314 |
|
340
|
315 |
/*
|
336
|
316 |
import CW10b._
|
|
317 |
println(time_needed(1, run(load_bff("collatz.bf"))))
|
|
318 |
println(time_needed(1, run2(load_bff("collatz.bf"))))
|
|
319 |
println(time_needed(1, run3(load_bff("collatz.bf"))))
|
|
320 |
println(time_needed(1, run4(load_bff("collatz.bf"))))
|
340
|
321 |
*/
|