1.#inf "immediately"

[badguy]

hey guys

when i want to print out amount of "seed"("seed*=pow(x,y)")-seed is a long double variable,it print"1.#INF" and after some changes it print "1.#IO",in big numbers.

how to solve this(any suggestion)?

thanks in advance.

ask for immediately answer.

[jonnin]

your result is too large. to fix it you may need to use a large number class, such as can be found on BOOST.ORG site, or other sites. What value did you have for x, y and initial value of seed?

Basically, the biggest floating point value you can support on most machines is 10-bytes, and that can be a pain, ususally its just 8 bytes, which has limitations. For integers usually 64 bit is all you have, which also is not that big really.

If you can, try to factor or reduce the equation somehow to avoid the problem first. For example, after 10 every factorial is 10 X value, so you can store them with one less exponent after 10 (there are several factors that can be removed this way to save space on factorials but eventually you have to use bigger number objects).

[badguy]

the value that i have is sth like this (i know this is ridiculous)

long double seed=24;
for(int n=0;n<=380;n++,year++){
cout<<setw(4)<<year;
for(float i=0.05;i<=0.10;i+=0.01){
seed*=pow((1+i),n);
cout<<setw(12)<<seed;
}
so what is the point?

[jonnin]

That code does not go INF immediately. It prints this, as expected:

24
24
24
24
24
24
25.2
26.712
28.5818
30.8684
33.6465
37.0112
40.8048
45.8483
52.4917
61.2264
72.7431
88.0191
101.893
121.356
148.667
187.277
242.529
322.807
392.374
495.362
649.319
883.392
1246.98
1825.7
2330.11
3118.21
4373.46
6426.04
9887.27
15923.5
21339.1
30269.9
45426.9
72086.8
120897
214176
301367
453145
727651
1.24707e+006
2.27969e+006
4.44247e+006
6.56354e+006
1.04613e+007
1.79744e+007
3.32695e+007
6.62915e+007
1.42102e+008
2.20446e+008
3.72439e+008
6.84715e+008
1.36875e+009
2.97278e+009
7.00965e+009
1.1418e+010
2.04478e+010
4.0224e+010
8.68407e+010
2.05584e+011
5.33231e+011
9.12006e+011
1.73126e+012
3.64404e+012
8.4966e+012
2.19249e+013
6.25542e+013
1.12338e+014
2.26046e+014
5.091e+014
1.282e+015
3.60584e+015
1.13167e+016
2.13393e+016
4.55151e+016
1.09684e+017
2.98301e+017
9.14532e+017
3.15721e+018
6.25106e+018
1.4133e+019
3.64425e+019
1.07039e+020
3.57695e+020
1.35835e+021
2.8239e+021
6.76764e+021
1.86721e+022
5.92312e+022
2.15749e+023
9.01236e+023
1.96728e+024
4.99759e+024
1.47537e+025
5.05454e+025
2.00681e+026
9.22123e+026
2.11352e+027
5.69123e+027
1.79776e+028
6.65173e+028
2.87863e+029
1.45499e+030
3.50161e+030
9.99478e+030
3.37817e+031
1.34992e+032
6.36776e+032
3.54043e+033
8.94647e+033
2.70684e+034
9.78937e+034
4.2248e+035
2.17225e+036
1.32853e+037
3.52498e+037
1.13051e+038
4.37472e+038
2.03904e+039
1.14276e+040
7.68793e+040
2.14183e+041
7.28127e+041
3.01486e+042
1.51763e+043
9.27092e+043
6.86071e+044
2.00693e+045
7.23206e+045
3.20409e+046
1.74192e+047
1.15988e+048
9.44172e+048
2.90004e+049
1.10774e+050
5.25129e+050
3.08328e+051
2.23781e+052
2.0038e+053
6.46245e+053
2.6166e+054
1.32724e+055
8.41626e+055
6.65818e+056
6.55813e+057
2.22081e+058
9.53143e+058
5.17312e+059
3.5428e+060
3.05499e+061
3.31e+062
1.17692e+063
5.35427e+063
3.10942e+064
2.29984e+065
2.16166e+066
2.5763e+067
9.6185e+067
4.63837e+068
2.88222e+069
2.30234e+070
2.35877e+071
3.09235e+072
1.21224e+073
6.19657e+073
4.12e+074
3.55438e+075
3.96923e+076
5.72402e+077
2.35608e+078
1.27661e+079
9.08217e+079
8.46212e+080
1.03003e+082
1.63394e+083
7.06178e+083
4.05592e+084
3.08748e+085
3.10682e+086
4.12204e+087
7.19272e+088
3.26408e+089
1.9872e+090
1.6186e+091
1.75904e+092
2.54389e+093
4.88283e+094
2.32664e+095
1.50147e+096
1.30857e+097
1.53588e+098
2.42106e+099
5.11179e+100
2.55752e+101
1.74949e+102
1.63146e+103
2.06805e+104
3.55334e+105
8.25268e+106
4.33541e+107
3.14361e+108
3.13674e+109
4.29424e+110
8.04244e+111
2.05466e+113
1.13335e+114
8.71101e+114
9.3004e+115
1.3751e+117
2.80712e+118
7.88871e+119
4.56899e+120
3.72246e+121
4.25253e+122
6.79051e+123
1.51098e+125
4.67083e+126
2.84052e+127
2.45309e+128
2.99857e+129
5.17123e+130
1.25422e+132
4.26486e+133
2.72331e+134
2.49298e+135
3.26064e+136
6.07305e+137
1.60551e+139
6.00533e+140
4.02641e+141
3.90702e+142
5.46782e+143
1.09987e+145
3.16939e+146
1.30404e+148
9.1804e+148
9.44268e+149
1.41399e+151
3.07183e+152
9.64847e+153
4.36683e+155
3.22795e+156
3.51938e+157
5.639e+158
1.32305e+160
4.52964e+161
2.25509e+163
1.7503e+164
2.02283e+165
3.468e+166
8.78773e+167
3.27937e+169
1.79591e+171
1.4636e+172
1.79297e+173
3.2891e+174
9.00118e+175
3.66134e+177
2.20559e+179
1.88735e+180
2.45081e+181
4.81058e+182
1.42182e+184
6.30391e+185
4.17723e+187
3.75324e+188
5.16617e+189
1.08503e+191
3.46346e+192
1.6738e+194
1.22004e+196
1.15101e+197
1.67938e+198
3.77402e+199
1.30106e+201
6.85358e+202
5.49517e+204
5.44349e+205
8.41883e+206
2.02437e+208
7.53716e+209
4.32767e+211
3.8169e+213
3.97005e+214
6.50843e+215
1.67455e+217
6.73348e+218
4.21417e+220
4.08848e+222
4.46515e+223
7.7593e+224
2.13613e+226
9.27669e+227
6.32838e+229
6.75359e+231
7.74459e+232
1.42656e+234
4.20223e+235
1.97092e+237
1.46553e+239
1.7204e+241
2.07149e+242
4.04464e+243
1.27484e+245
6.45753e+246
5.23382e+248
6.75844e+250
8.54454e+251
1.76845e+253
5.96417e+254
3.26276e+256
2.88247e+258
4.09435e+260
5.43521e+261
1.19241e+263
4.30296e+264
2.5423e+266
2.44812e+268
3.82512e+270
5.33171e+271
1.23989e+273
4.78748e+274
3.05485e+276
3.20643e+278
5.51097e+280
8.06562e+281
1.9882e+283
8.21425e+284
5.66077e+286
6.4764e+288
1.22442e+291
1.88162e+292
4.91653e+293
2.17346e+295
1.61764e+297
2.01728e+299
4.19525e+301
6.76935e+302
1.87491e+304
8.86863e+305
7.12872e+307
1.#INF
1.#INF
1.#INF
1.#INF
... lots more infinitys

Press any key to continue . . .

[jonnin]

What is happening is the *= statement is growing the value at a rapid rate. The power function is OK, the values from it are small, but if you just do this:

y = 2;
for(x = 1; x < 1000; x++)
y *= 2;

it will be bigger than you can store in short order, what you have is similar to this.

[badguy]

i know that but the problem says you must print out every (n=380) result which goes to infinite

[jonnin]

then you need to use a large number class. Its like trying to park a semi with a trailer in a house garage... it will not fit.

An 8 byte floating point number holds up to 1.7 X 10^308 with 15 digits or so of precision. If you need something bigger than this, you must use those classes.

Note that long double is at least as big as a double, and a double is 8 byte floating point. A long double may be 10 bytes, but it may just be 8 bytes, in other words, depends on your compiler.

If this is an assignment, in an intro level class, odds are your job is just to understand this limitation and not to fix it.

[badguy]

thanks