<< problem 86 - Cuboid route | Product-sum numbers - problem 88 >> |

# Problem 87: Prime power triples

(see projecteuler.net/problem=87)

The smallest number expressible as the sum of a prime square, prime cube, and prime fourth power is 28.

In fact, there are exactly four numbers below fifty that can be expressed in such a way:

28 = 2^2 + 2^3 + 2^4

33 = 3^2 + 2^3 + 2^4

49 = 5^2 + 2^3 + 2^4

47 = 2^2 + 3^3 + 2^4

How many numbers below fifty million can be expressed as the sum of a prime square, prime cube, and prime fourth power?

# Algorithm

A simple prime sieve is responsible to find all primes i < sqrt{50000000}.

Then three nested loops compute all sums of all combinations of such primes where a^2 + b^3 + c^4 < 50000000.

Be careful: `b*b*b`

and `c*c*c*c`

can easily exceed an `unsigned int`

.

Those sums are sorted and duplicate sums are removed. Now we have a nice `std::vector`

with all sums.

The original problem is solved now (just display `sums.size()`

) but the Hackerrank problem is slightly tougher.

## Modifications by HackerRank

In order to find the the number of sums which are below a certain input value, I search through the sorted container with `std::upper_bound`

and then compute the distance to the beginning of the container.

That's extremely fast (`std::upper_bound`

most likely uses binary search) and easily processed thousands of test cases per second.

# My code

… was written in C++11 and can be compiled with G++, Clang++, Visual C++. You can download it, too.

#include <vector>
#include <algorithm>
#include <iostream>
int main()
{
const unsigned int MaxLimit = 100 * 1000 * 1000; // Hackerrank: 10^7 instead of 5*10^6
// prime sieve
std::vector<unsigned int> primes;
primes.push_back(2);
for (unsigned int i = 3; i*i < MaxLimit; i += 2)
{
bool isPrime = true;
// test against all prime numbers we have so far (in ascending order)
for (auto p : primes)
{
// next prime is too large to be a divisor ?
if (p*p > i)
break;
// divisible ? => not prime
if (i % p == 0)
{
isPrime = false;
break;
}
}
// yes, we have a prime number
if (isPrime)
primes.push_back(i);
}
// just three nested loops where I generate all sums
std::vector<unsigned int> sums;
for (auto a : primes)
for (auto b : primes)
for (auto c : primes)
{
auto a2 = a*a;
auto b3 = (unsigned long long)b*b*b;
auto c4 = (unsigned long long)c*c*c*c;
auto sum = a2 + b3 + c4;
// abort if too big
if (sum > MaxLimit)
break;
sums.push_back(sum);
}
// sort ascendingly
std::sort(sums.begin(), sums.end());
// a few sums occur twice, let's remove them !
auto last = std::unique(sums.begin(), sums.end());
// process test cases
unsigned int tests = 1;
std::cin >> tests;
while (tests--)
{
unsigned int limit = MaxLimit;
std::cin >> limit;
// find next sum which is bigger than the limit
auto pos = std::upper_bound(sums.begin(), last, limit);
// how many sums are inbetween 28 and limit ?
auto num = std::distance(sums.begin(), pos);
std::cout << num << std::endl;
}
return 0;
}

This solution contains 11 empty lines, 12 comments and 3 preprocessor commands.

# Interactive test

You can submit your own input to my program and it will be instantly processed at my server:

This is equivalent to`echo "1 50" | ./87`

Output:

*Note:* the original problem's input `50000000`

__cannot__ be entered

because just copying results is a soft skill reserved for idiots.

*(this interactive test is still under development, computations will be aborted after one second)*

# Benchmark

The correct solution to the original Project Euler problem was found in 0.16 seconds on a Intel® Core™ i7-2600K CPU @ 3.40GHz.

Peak memory usage was about 20 MByte.

(compiled for x86_64 / Linux, GCC flags: `-O3 -march=native -fno-exceptions -fno-rtti -std=c++11 -DORIGINAL`

)

See here for a comparison of all solutions.

**Note:** interactive tests run on a weaker (=slower) computer. Some interactive tests are compiled without `-DORIGINAL`

.

# Changelog

March 12, 2017 submitted solution

May 5, 2017 added comments

# Hackerrank

see https://www.hackerrank.com/contests/projecteuler/challenges/euler087

My code solves **9** out of **9** test cases (score: **100%**)

# Difficulty

Project Euler ranks this problem at **20%** (out of 100%).

Hackerrank describes this problem as **easy**.

*Note:*

Hackerrank has strict execution time limits (typically 2 seconds for C++ code) and often a much wider input range than the original problem.

In my opinion, Hackerrank's modified problems are usually a lot harder to solve. As a rule thumb: brute-force is rarely an option.

# Links

projecteuler.net/thread=87 - **the** best forum on the subject (*note:* you have to submit the correct solution first)

Code in various languages:

Python: www.mathblog.dk/project-euler-87-sum-power/ (written by Kristian Edlund)

Java: github.com/nayuki/Project-Euler-solutions/blob/master/java/p087.java (written by Nayuki)

Scala: github.com/samskivert/euler-scala/blob/master/Euler087.scala (written by Michael Bayne)

# Heatmap

green problems solve the original Project Euler problem and have a perfect score of 100% at Hackerrank, too.

yellow problems score less than 100% at Hackerrank (but still solve the original problem).

gray problems are already solved but I haven't published my solution yet.

blue problems are solved and there wasn't a Hackerrank version of it at the time I solved it or I didn't care about it because it differed too much.

red problems are solved but exceed the time limit of one minute.

*Please click on a problem's number to open my solution to that problem:*

1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 |

26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 |

51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | 61 | 62 | 63 | 64 | 65 | 66 | 67 | 68 | 69 | 70 | 71 | 72 | 73 | 74 | 75 |

76 | 77 | 78 | 79 | 80 | 81 | 82 | 83 | 84 | 85 | 86 | 87 | 88 | 89 | 90 | 91 | 92 | 93 | 94 | 95 | 96 | 97 | 98 | 99 | 100 |

101 | 102 | 103 | 104 | 105 | 106 | 107 | 108 | 109 | 110 | 111 | 112 | 113 | 114 | 115 | 116 | 117 | 118 | 119 | 120 | 121 | 122 | 123 | 124 | 125 |

126 | 127 | 128 | 129 | 130 | 131 | 132 | 133 | 134 | 135 | 136 | 137 | 138 | 139 | 140 | 141 | 142 | 143 | 144 | 145 | 146 | 147 | 148 | 149 | 150 |

151 | 152 | 153 | 154 | 155 | 156 | 157 | 158 | 159 | 160 | 161 | 162 | 163 | 164 | 165 | 166 | 167 | 168 | 169 | 170 | 171 | 172 | 173 | 174 | 175 |

176 | 177 | 178 | 179 | 180 | 181 | 182 | 183 | 184 | 185 | 186 | 187 | 188 | 189 | 190 | 191 | 192 | 193 | 194 | 195 | 196 | 197 | 198 | 199 | 200 |

201 | 202 | 203 | 204 | 205 | 206 | 207 | 208 | 209 | 210 | 211 | 212 | 213 | 214 | 215 | 216 | 217 | 218 | 219 | 220 | 221 | 222 | 223 | 224 | 225 |

226 | 227 | 228 | 229 | 230 | 231 | 232 | 233 | 234 | 235 | 236 | 237 | 238 | 239 | 240 | 241 | 242 | 243 | 244 | 245 | 246 | 247 | 248 | 249 | 250 |

251 | 252 | 253 | 254 | 255 | 256 | 257 | 258 | 259 | 260 | 261 | 262 | 263 | 264 | 265 | 266 | 267 | 268 | 269 | 270 | 271 | 272 | 273 | 274 | 275 |

276 | 277 | 278 | 279 | 280 | 281 | 282 | 283 | 284 | 285 | 286 | 287 | 288 | 289 | 290 | 291 | 292 | 293 | 294 | 295 | 296 | 297 | 298 | 299 | 300 |

301 | 302 | 303 | 304 | 305 | 306 | 307 | 308 | 309 | 310 | 311 | 312 | 313 | 314 | 315 | 316 | 317 | 318 | 319 | 320 | 321 | 322 | 323 | 324 | 325 |

326 | 327 | 328 | 329 | 330 | 331 | 332 | 333 | 334 | 335 | 336 | 337 | 338 | 339 | 340 | 341 | 342 | 343 | 344 | 345 | 346 | 347 | 348 | 349 | 350 |

351 | 352 | 353 | 354 | 355 | 356 | 357 | 358 | 359 | 360 | 361 | 362 | 363 | 364 | 365 | 366 | 367 | 368 | 369 | 370 | 371 | 372 | 373 | 374 | 375 |

I scored 12,626 points (out of 14300 possible points, top rank was 20 out ouf ≈60000 in July 2017) at Hackerrank's Project Euler+.

Look at my progress and performance pages to get more details.

My username at Project Euler is

**stephanbrumme**while it's stbrumme at Hackerrank.

<< problem 86 - Cuboid route | Product-sum numbers - problem 88 >> |