[docs] Add LICENSE.txt to the root of the mono-repo
[llvm-project.git] / llvm / unittests / Support / BranchProbabilityTest.cpp
blob94626e796e2ec0f362d58dedbd9436e4695fd707
1 //===- unittest/Support/BranchProbabilityTest.cpp - BranchProbability tests -=//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
9 #include "llvm/Support/BranchProbability.h"
10 #include "llvm/Support/raw_ostream.h"
11 #include "gtest/gtest.h"
13 using namespace llvm;
15 namespace llvm {
16 void PrintTo(BranchProbability P, ::std::ostream *os) {
17 *os << P.getNumerator() << "/" << P.getDenominator();
20 namespace {
22 typedef BranchProbability BP;
23 TEST(BranchProbabilityTest, Accessors) {
24 EXPECT_EQ(306783378u, BP(1, 7).getNumerator());
25 EXPECT_EQ(1u << 31, BP(1, 7).getDenominator());
26 EXPECT_EQ(0u, BP::getZero().getNumerator());
27 EXPECT_EQ(1u << 31, BP::getZero().getDenominator());
28 EXPECT_EQ(1u << 31, BP::getOne().getNumerator());
29 EXPECT_EQ(1u << 31, BP::getOne().getDenominator());
32 TEST(BranchProbabilityTest, Operators) {
33 EXPECT_TRUE(BP(1, 7) < BP(2, 7));
34 EXPECT_TRUE(BP(1, 7) < BP(1, 4));
35 EXPECT_TRUE(BP(5, 7) < BP(3, 4));
36 EXPECT_FALSE(BP(1, 7) < BP(1, 7));
37 EXPECT_FALSE(BP(1, 7) < BP(2, 14));
38 EXPECT_FALSE(BP(4, 7) < BP(1, 2));
39 EXPECT_FALSE(BP(4, 7) < BP(3, 7));
41 EXPECT_FALSE(BP(1, 7) > BP(2, 7));
42 EXPECT_FALSE(BP(1, 7) > BP(1, 4));
43 EXPECT_FALSE(BP(5, 7) > BP(3, 4));
44 EXPECT_FALSE(BP(1, 7) > BP(1, 7));
45 EXPECT_FALSE(BP(1, 7) > BP(2, 14));
46 EXPECT_TRUE(BP(4, 7) > BP(1, 2));
47 EXPECT_TRUE(BP(4, 7) > BP(3, 7));
49 EXPECT_TRUE(BP(1, 7) <= BP(2, 7));
50 EXPECT_TRUE(BP(1, 7) <= BP(1, 4));
51 EXPECT_TRUE(BP(5, 7) <= BP(3, 4));
52 EXPECT_TRUE(BP(1, 7) <= BP(1, 7));
53 EXPECT_TRUE(BP(1, 7) <= BP(2, 14));
54 EXPECT_FALSE(BP(4, 7) <= BP(1, 2));
55 EXPECT_FALSE(BP(4, 7) <= BP(3, 7));
57 EXPECT_FALSE(BP(1, 7) >= BP(2, 7));
58 EXPECT_FALSE(BP(1, 7) >= BP(1, 4));
59 EXPECT_FALSE(BP(5, 7) >= BP(3, 4));
60 EXPECT_TRUE(BP(1, 7) >= BP(1, 7));
61 EXPECT_TRUE(BP(1, 7) >= BP(2, 14));
62 EXPECT_TRUE(BP(4, 7) >= BP(1, 2));
63 EXPECT_TRUE(BP(4, 7) >= BP(3, 7));
65 EXPECT_FALSE(BP(1, 7) == BP(2, 7));
66 EXPECT_FALSE(BP(1, 7) == BP(1, 4));
67 EXPECT_FALSE(BP(5, 7) == BP(3, 4));
68 EXPECT_TRUE(BP(1, 7) == BP(1, 7));
69 EXPECT_TRUE(BP(1, 7) == BP(2, 14));
70 EXPECT_FALSE(BP(4, 7) == BP(1, 2));
71 EXPECT_FALSE(BP(4, 7) == BP(3, 7));
73 EXPECT_TRUE(BP(1, 7) != BP(2, 7));
74 EXPECT_TRUE(BP(1, 7) != BP(1, 4));
75 EXPECT_TRUE(BP(5, 7) != BP(3, 4));
76 EXPECT_FALSE(BP(1, 7) != BP(1, 7));
77 EXPECT_FALSE(BP(1, 7) != BP(2, 14));
78 EXPECT_TRUE(BP(4, 7) != BP(1, 2));
79 EXPECT_TRUE(BP(4, 7) != BP(3, 7));
81 EXPECT_TRUE(BP(1, 7) == BP(2, 14));
82 EXPECT_TRUE(BP(1, 7) == BP(3, 21));
83 EXPECT_TRUE(BP(5, 7) == BP(25, 35));
84 EXPECT_TRUE(BP(99999998, 100000000) < BP(99999999, 100000000));
85 EXPECT_TRUE(BP(4, 8) == BP(400000000, 800000000));
88 TEST(BranchProbabilityTest, MoreOperators) {
89 BP A(4, 5);
90 BP B(4U << 29, 5U << 29);
91 BP C(3, 4);
93 EXPECT_TRUE(A == B);
94 EXPECT_FALSE(A != B);
95 EXPECT_FALSE(A < B);
96 EXPECT_FALSE(A > B);
97 EXPECT_TRUE(A <= B);
98 EXPECT_TRUE(A >= B);
100 EXPECT_FALSE(B == C);
101 EXPECT_TRUE(B != C);
102 EXPECT_FALSE(B < C);
103 EXPECT_TRUE(B > C);
104 EXPECT_FALSE(B <= C);
105 EXPECT_TRUE(B >= C);
107 BP BigZero(0, UINT32_MAX);
108 BP BigOne(UINT32_MAX, UINT32_MAX);
109 EXPECT_FALSE(BigZero == BigOne);
110 EXPECT_TRUE(BigZero != BigOne);
111 EXPECT_TRUE(BigZero < BigOne);
112 EXPECT_FALSE(BigZero > BigOne);
113 EXPECT_TRUE(BigZero <= BigOne);
114 EXPECT_FALSE(BigZero >= BigOne);
117 TEST(BranchProbabilityTest, ArithmeticOperators) {
118 BP Z(0, 1);
119 BP O(1, 1);
120 BP H(1, 2);
121 BP Q(1, 4);
122 BP Q3(3, 4);
124 EXPECT_EQ(Z + O, O);
125 EXPECT_EQ(H + Z, H);
126 EXPECT_EQ(H + H, O);
127 EXPECT_EQ(Q + H, Q3);
128 EXPECT_EQ(Q + Q3, O);
129 EXPECT_EQ(H + Q3, O);
130 EXPECT_EQ(Q3 + Q3, O);
132 EXPECT_EQ(Z - O, Z);
133 EXPECT_EQ(O - Z, O);
134 EXPECT_EQ(O - H, H);
135 EXPECT_EQ(O - Q, Q3);
136 EXPECT_EQ(Q3 - H, Q);
137 EXPECT_EQ(Q - H, Z);
138 EXPECT_EQ(Q - Q3, Z);
140 EXPECT_EQ(Z * O, Z);
141 EXPECT_EQ(H * H, Q);
142 EXPECT_EQ(Q * O, Q);
143 EXPECT_EQ(O * O, O);
144 EXPECT_EQ(Z * Z, Z);
146 EXPECT_EQ(Z * 3, Z);
147 EXPECT_EQ(Q * 3, Q3);
148 EXPECT_EQ(H * 3, O);
149 EXPECT_EQ(Q3 * 2, O);
150 EXPECT_EQ(O * UINT32_MAX, O);
152 EXPECT_EQ(Z / 4, Z);
153 EXPECT_EQ(O / 4, Q);
154 EXPECT_EQ(Q3 / 3, Q);
155 EXPECT_EQ(H / 2, Q);
156 EXPECT_EQ(O / 2, H);
157 EXPECT_EQ(H / UINT32_MAX, Z);
159 BP Min(1, 1u << 31);
161 EXPECT_EQ(O / UINT32_MAX, Z);
162 EXPECT_EQ(Min * UINT32_MAX, O);
165 TEST(BranchProbabilityTest, getCompl) {
166 EXPECT_EQ(BP(5, 7), BP(2, 7).getCompl());
167 EXPECT_EQ(BP(2, 7), BP(5, 7).getCompl());
168 EXPECT_EQ(BP::getZero(), BP(7, 7).getCompl());
169 EXPECT_EQ(BP::getOne(), BP(0, 7).getCompl());
172 TEST(BranchProbabilityTest, scale) {
173 // Multiply by 1.0.
174 EXPECT_EQ(UINT64_MAX, BP(1, 1).scale(UINT64_MAX));
175 EXPECT_EQ(UINT64_MAX, BP(7, 7).scale(UINT64_MAX));
176 EXPECT_EQ(UINT32_MAX, BP(1, 1).scale(UINT32_MAX));
177 EXPECT_EQ(UINT32_MAX, BP(7, 7).scale(UINT32_MAX));
178 EXPECT_EQ(0u, BP(1, 1).scale(0));
179 EXPECT_EQ(0u, BP(7, 7).scale(0));
181 // Multiply by 0.0.
182 EXPECT_EQ(0u, BP(0, 1).scale(UINT64_MAX));
183 EXPECT_EQ(0u, BP(0, 1).scale(UINT64_MAX));
184 EXPECT_EQ(0u, BP(0, 1).scale(0));
186 auto Two63 = UINT64_C(1) << 63;
187 auto Two31 = UINT64_C(1) << 31;
189 // Multiply by 0.5.
190 EXPECT_EQ(Two63 - 1, BP(1, 2).scale(UINT64_MAX));
192 // Big fractions.
193 EXPECT_EQ(1u, BP(Two31, UINT32_MAX).scale(2));
194 EXPECT_EQ(Two31, BP(Two31, UINT32_MAX).scale(Two31 * 2));
195 EXPECT_EQ(9223372036854775807ULL, BP(Two31, UINT32_MAX).scale(UINT64_MAX));
197 // High precision.
198 EXPECT_EQ(UINT64_C(9223372045444710399),
199 BP(Two31 + 1, UINT32_MAX - 2).scale(UINT64_MAX));
202 TEST(BranchProbabilityTest, scaleByInverse) {
203 // Divide by 1.0.
204 EXPECT_EQ(UINT64_MAX, BP(1, 1).scaleByInverse(UINT64_MAX));
205 EXPECT_EQ(UINT64_MAX, BP(7, 7).scaleByInverse(UINT64_MAX));
206 EXPECT_EQ(UINT32_MAX, BP(1, 1).scaleByInverse(UINT32_MAX));
207 EXPECT_EQ(UINT32_MAX, BP(7, 7).scaleByInverse(UINT32_MAX));
208 EXPECT_EQ(0u, BP(1, 1).scaleByInverse(0));
209 EXPECT_EQ(0u, BP(7, 7).scaleByInverse(0));
211 auto MAX_DENOMINATOR = BP::getDenominator();
213 // Divide by something very small.
214 EXPECT_EQ(UINT64_MAX, BP(1, UINT32_MAX).scaleByInverse(UINT64_MAX));
215 EXPECT_EQ(uint64_t(UINT32_MAX) * MAX_DENOMINATOR,
216 BP(1, MAX_DENOMINATOR).scaleByInverse(UINT32_MAX));
217 EXPECT_EQ(MAX_DENOMINATOR, BP(1, MAX_DENOMINATOR).scaleByInverse(1));
219 auto Two63 = UINT64_C(1) << 63;
220 auto Two31 = UINT64_C(1) << 31;
222 // Divide by 0.5.
223 EXPECT_EQ(UINT64_MAX - 1, BP(1, 2).scaleByInverse(Two63 - 1));
224 EXPECT_EQ(UINT64_MAX, BP(1, 2).scaleByInverse(Two63));
226 // Big fractions.
227 EXPECT_EQ(2u, BP(Two31, UINT32_MAX).scaleByInverse(1));
228 EXPECT_EQ(2u, BP(Two31 - 1, UINT32_MAX).scaleByInverse(1));
229 EXPECT_EQ(Two31 * 2, BP(Two31, UINT32_MAX).scaleByInverse(Two31));
230 EXPECT_EQ(Two31 * 2, BP(Two31 - 1, UINT32_MAX).scaleByInverse(Two31));
231 EXPECT_EQ(UINT64_MAX, BP(Two31, UINT32_MAX).scaleByInverse(Two63 + Two31));
233 // High precision. The exact answers to these are close to the successors of
234 // the floor. If we were rounding, these would round up.
235 EXPECT_EQ(UINT64_C(18446744060824649767),
236 BP(Two31 + 2, UINT32_MAX - 2)
237 .scaleByInverse(UINT64_C(9223372047592194056)));
238 EXPECT_EQ(UINT64_C(18446744060824649739),
239 BP(Two31 + 1, UINT32_MAX).scaleByInverse(Two63 + Two31));
242 TEST(BranchProbabilityTest, scaleBruteForce) {
243 struct {
244 uint64_t Num;
245 uint32_t Prob[2];
246 uint64_t Result;
247 } Tests[] = {
248 // Data for scaling that results in <= 64 bit division.
249 { 0x1423e2a50ULL, { 0x64819521, 0x7765dd13 }, 0x10f418888ULL },
250 { 0x35ef14ceULL, { 0x28ade3c7, 0x304532ae }, 0x2d73c33bULL },
251 { 0xd03dbfbe24ULL, { 0x790079, 0xe419f3 }, 0x6e776fc2c4ULL },
252 { 0x21d67410bULL, { 0x302a9dc2, 0x3ddb4442 }, 0x1a5948fd4ULL },
253 { 0x8664aeadULL, { 0x3d523513, 0x403523b1 }, 0x805a04cfULL },
254 { 0x201db0cf4ULL, { 0x35112a7b, 0x79fc0c74 }, 0xdf8b07f8ULL },
255 { 0x13f1e4430aULL, { 0x21c92bf, 0x21e63aae }, 0x13e0cba26ULL },
256 { 0x16c83229ULL, { 0x3793f66f, 0x53180dea }, 0xf3ce7b6ULL },
257 { 0xc62415be8ULL, { 0x9cc4a63, 0x4327ae9b }, 0x1ce8b71c1ULL },
258 { 0x6fac5e434ULL, { 0xe5f9170, 0x1115e10b }, 0x5df23dd4cULL },
259 { 0x1929375f2ULL, { 0x3a851375, 0x76c08456 }, 0xc662b083ULL },
260 { 0x243c89db6ULL, { 0x354ebfc0, 0x450ef197 }, 0x1bf8c1663ULL },
261 { 0x310e9b31aULL, { 0x1b1b8acf, 0x2d3629f0 }, 0x1d69c93f9ULL },
262 { 0xa1fae921dULL, { 0xa7a098c, 0x10469f44 }, 0x684413d6eULL },
263 { 0xc1582d957ULL, { 0x498e061, 0x59856bc }, 0x9edc5f4ecULL },
264 { 0x57cfee75ULL, { 0x1d061dc3, 0x7c8bfc17 }, 0x1476a220ULL },
265 { 0x139220080ULL, { 0x294a6c71, 0x2a2b07c9 }, 0x1329e1c75ULL },
266 { 0x1665d353cULL, { 0x7080db5, 0xde0d75c }, 0xb590d9faULL },
267 { 0xe8f14541ULL, { 0x5188e8b2, 0x736527ef }, 0xa4971be5ULL },
268 { 0x2f4775f29ULL, { 0x254ef0fe, 0x435fcf50 }, 0x1a2e449c1ULL },
269 { 0x27b85d8d7ULL, { 0x304c8220, 0x5de678f2 }, 0x146e3befbULL },
270 { 0x1d362e36bULL, { 0x36c85b12, 0x37a66f55 }, 0x1cc19b8e7ULL },
271 { 0x155fd48c7ULL, { 0xf5894d, 0x1256108 }, 0x11e383604ULL },
272 { 0xb5db2d15ULL, { 0x39bb26c5, 0x5bdcda3e }, 0x72499259ULL },
273 { 0x153990298ULL, { 0x48921c09, 0x706eb817 }, 0xdb3268e7ULL },
274 { 0x28a7c3ed7ULL, { 0x1f776fd7, 0x349f7a70 }, 0x184f73ae2ULL },
275 { 0x724dbeabULL, { 0x1bd149f5, 0x253a085e }, 0x5569c0b3ULL },
276 { 0xd8f0c513ULL, { 0x18c8cc4c, 0x1b72bad0 }, 0xc3e30642ULL },
277 { 0x17ce3dcbULL, { 0x1e4c6260, 0x233b359e }, 0x1478f4afULL },
278 { 0x1ce036ce0ULL, { 0x29e3c8af, 0x5318dd4a }, 0xe8e76195ULL },
279 { 0x1473ae2aULL, { 0x29b897ba, 0x2be29378 }, 0x13718185ULL },
280 { 0x1dd41aa68ULL, { 0x3d0a4441, 0x5a0e8f12 }, 0x1437b6bbfULL },
281 { 0x1b49e4a53ULL, { 0x3430c1fe, 0x5a204aed }, 0xfcd6852fULL },
282 { 0x217941b19ULL, { 0x12ced2bd, 0x21b68310 }, 0x12aca65b1ULL },
283 { 0xac6a4dc8ULL, { 0x3ed68da8, 0x6fdca34c }, 0x60da926dULL },
284 { 0x1c503a4e7ULL, { 0xfcbbd32, 0x11e48d17 }, 0x18fec7d37ULL },
285 { 0x1c885855ULL, { 0x213e919d, 0x25941897 }, 0x193de742ULL },
286 { 0x29b9c168eULL, { 0x2b644aea, 0x45725ee7 }, 0x1a122e5d4ULL },
287 { 0x806a33f2ULL, { 0x30a80a23, 0x5063733a }, 0x4db9a264ULL },
288 { 0x282afc96bULL, { 0x143ae554, 0x1a9863ff }, 0x1e8de5204ULL },
289 // Data for scaling that results in > 64 bit division.
290 { 0x23ca5f2f672ca41cULL, { 0xecbc641, 0x111373f7 }, 0x1f0301e5c76869c6ULL },
291 { 0x5e4f2468142265e3ULL, { 0x1ddf5837, 0x32189233 }, 0x383ca7bad6053ac9ULL },
292 { 0x277a1a6f6b266bf6ULL, { 0x415d81a8, 0x61eb5e1e }, 0x1a5a3e1d1c9e8540ULL },
293 { 0x1bdbb49a237035cbULL, { 0xea5bf17, 0x1d25ffb3 }, 0xdffc51c5cb51cf1ULL },
294 { 0x2bce6d29b64fb8ULL, { 0x3bfd5631, 0x7525c9bb }, 0x166ebedd9581fdULL },
295 { 0x3a02116103df5013ULL, { 0x2ee18a83, 0x3299aea8 }, 0x35be89227276f105ULL },
296 { 0x7b5762390799b18cULL, { 0x12f8e5b9, 0x2563bcd4 }, 0x3e960077695655a3ULL },
297 { 0x69cfd72537021579ULL, { 0x4c35f468, 0x6a40feee }, 0x4be4cb38695a4f30ULL },
298 { 0x49dfdf835120f1c1ULL, { 0x8cb3759, 0x559eb891 }, 0x79663f6e3c8d8f6ULL },
299 { 0x74b5be5c27676381ULL, { 0x47e4c5e0, 0x7c7b19ff }, 0x4367d2dfb22b3265ULL },
300 { 0x4f50f97075e7f431ULL, { 0x9a50a17, 0x11cd1185 }, 0x2af952b30374f382ULL },
301 { 0x2f8b0d712e393be4ULL, { 0x1487e386, 0x15aa356e }, 0x2d0df3649b2b19fcULL },
302 { 0x224c1c75999d3deULL, { 0x3b2df0ea, 0x4523b100 }, 0x1d5b481d160dd8bULL },
303 { 0x2bcbcea22a399a76ULL, { 0x28b58212, 0x48dd013e }, 0x187814d0610c8a56ULL },
304 { 0x1dbfca91257cb2d1ULL, { 0x1a8c04d9, 0x5e92502c }, 0x859cf7d19e83ad0ULL },
305 { 0x7f20039b57cda935ULL, { 0xeccf651, 0x323f476e }, 0x25720cd9054634bdULL },
306 { 0x40512c6a586aa087ULL, { 0x113b0423, 0x398c9eab }, 0x1341c03dbb662054ULL },
307 { 0x63d802693f050a11ULL, { 0xf50cdd6, 0xfce2a44 }, 0x60c0177b667a4feaULL },
308 { 0x2d956b422838de77ULL, { 0xb2d345b, 0x1321e557 }, 0x1aa0ed16b094575cULL },
309 { 0x5a1cdf0c1657bc91ULL, { 0x1d77bb0c, 0x1f991ff1 }, 0x54097ee9907290eaULL },
310 { 0x3801b26d7e00176bULL, { 0xeed25da, 0x1a819d8b }, 0x1f89e96a616b9abeULL },
311 { 0x37655e74338e1e45ULL, { 0x300e170a, 0x5a1595fe }, 0x1d8cfb55ff6a6dbcULL },
312 { 0x7b38703f2a84e6ULL, { 0x66d9053, 0xc79b6b9 }, 0x3f7d4c91b9afb9ULL },
313 { 0x2245063c0acb3215ULL, { 0x30ce2f5b, 0x610e7271 }, 0x113b916455fe2560ULL },
314 { 0x6bc195877b7b8a7eULL, { 0x392004aa, 0x4a24e60c }, 0x530594fabfc81cc3ULL },
315 { 0x40a3fde23c7b43dbULL, { 0x4e712195, 0x6553e56e }, 0x320a799bc205c78dULL },
316 { 0x1d3dfc2866fbccbaULL, { 0x5075b517, 0x5fc42245 }, 0x18917f00745cb781ULL },
317 { 0x19aeb14045a61121ULL, { 0x1bf6edec, 0x707e2f4b }, 0x6626672aa2ba10aULL },
318 { 0x44ff90486c531e9fULL, { 0x66598a, 0x8a90dc }, 0x32f6f2b097001598ULL },
319 { 0x3f3e7121092c5bcbULL, { 0x1c754df7, 0x5951a1b9 }, 0x14267f50d4971583ULL },
320 { 0x60e2dafb7e50a67eULL, { 0x4d96c66e, 0x65bd878d }, 0x49e317155d75e883ULL },
321 { 0x656286667e0e6e29ULL, { 0x9d971a2, 0xacda23b }, 0x5c6ee3159e1deac3ULL },
322 { 0x1114e0974255d507ULL, { 0x1c693, 0x2d6ff }, 0xaae42e4be5f9f8dULL },
323 { 0x508c8baf3a70ff5aULL, { 0x3b26b779, 0x6ad78745 }, 0x2c983876178ed5b1ULL },
324 { 0x5b47bc666bf1f9cfULL, { 0x10a87ed6, 0x187d358a }, 0x3e1767153bea720aULL },
325 { 0x50954e3744460395ULL, { 0x7a42263, 0xcdaa048 }, 0x2fe739f0944a023cULL },
326 { 0x20020b406550dd8fULL, { 0x3318539, 0x42eead0 }, 0x186f326307c0d985ULL },
327 { 0x5bcb0b872439ffd5ULL, { 0x6f61fb2, 0x9af7344 }, 0x41fa1e3c47f0f80dULL },
328 { 0x7a670f365db87a53ULL, { 0x417e102, 0x3bb54c67 }, 0x8642a551d0f41b0ULL },
329 { 0x1ef0db1e7bab1cd0ULL, { 0x2b60cf38, 0x4188f78f }, 0x147ae0d63fc0575aULL }
332 for (const auto &T : Tests) {
333 EXPECT_EQ(T.Result, BP(T.Prob[0], T.Prob[1]).scale(T.Num));
337 TEST(BranchProbabilityTest, NormalizeProbabilities) {
338 const auto UnknownProb = BranchProbability::getUnknown();
340 SmallVector<BranchProbability, 2> Probs{{0, 1}, {0, 1}};
341 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
342 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[0].getNumerator());
343 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[1].getNumerator());
346 SmallVector<BranchProbability, 2> Probs{{0, 1}, {1, 1}};
347 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
348 EXPECT_EQ(0u, Probs[0].getNumerator());
349 EXPECT_EQ(BranchProbability::getDenominator(), Probs[1].getNumerator());
352 SmallVector<BranchProbability, 2> Probs{{1, 100}, {1, 100}};
353 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
354 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[0].getNumerator());
355 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[1].getNumerator());
358 SmallVector<BranchProbability, 2> Probs{{1, 1}, {1, 1}};
359 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
360 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[0].getNumerator());
361 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[1].getNumerator());
364 SmallVector<BranchProbability, 3> Probs{{1, 1}, {1, 1}, {1, 1}};
365 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
366 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
367 Probs[0].getNumerator());
368 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
369 Probs[1].getNumerator());
370 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
371 Probs[2].getNumerator());
374 SmallVector<BranchProbability, 2> Probs{{0, 1}, UnknownProb};
375 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
376 EXPECT_EQ(0U, Probs[0].getNumerator());
377 EXPECT_EQ(BranchProbability::getDenominator(), Probs[1].getNumerator());
380 SmallVector<BranchProbability, 2> Probs{{1, 1}, UnknownProb};
381 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
382 EXPECT_EQ(BranchProbability::getDenominator(), Probs[0].getNumerator());
383 EXPECT_EQ(0U, Probs[1].getNumerator());
386 SmallVector<BranchProbability, 2> Probs{{1, 2}, UnknownProb};
387 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
388 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[0].getNumerator());
389 EXPECT_EQ(BranchProbability::getDenominator() / 2, Probs[1].getNumerator());
392 SmallVector<BranchProbability, 4> Probs{
393 {1, 2}, {1, 2}, {1, 2}, UnknownProb};
394 BranchProbability::normalizeProbabilities(Probs.begin(), Probs.end());
395 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
396 Probs[0].getNumerator());
397 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
398 Probs[1].getNumerator());
399 EXPECT_EQ(BranchProbability::getDenominator() / 3 + 1,
400 Probs[2].getNumerator());
401 EXPECT_EQ(0U, Probs[3].getNumerator());