15. Kayan Nokta Aritmetiği: Sorunlar ve Sınırlamalar

Floating-point numbers are represented in computer hardware as base 2 (binary) fractions. For example, the decimal fraction 0.625 has value 6/10 + 2/100 + 5/1000, and in the same way the binary fraction 0.101 has value 1/2 + 0/4 + 1/8. These two fractions have identical values, the only real difference being that the first is written in base 10 fractional notation, and the second in base 2.

Ne yazık ki, ondalık kesirlerin çoğu tam olarak ikili kesir olarak gösterilemez. Bunun bir sonucu olarak, genel olarak, girdiğiniz ondalık kayan noktalı sayılar, makinede gerçekte depolanan ikili kayan noktalı sayılar tarafından yalnızca yaklaşık olarak gösterilir.

Problem ilk başta 10 tabanında daha kolay anlaşılabilir. 1/3 kesrini düşünün. Bunu 10 tabanında bir kesir olarak yaklaşık olarak hesaplayabilirsiniz:

0.3

ya da daha iyisi

0.33

ya da daha iyisi

0.333

ve bunun gibi. Kaç basamak yazmak isterseniz isteyin, sonuç hiçbir zaman tam olarak 1/3 olmayacak, ancak 1/3’ün giderek daha iyi bir yaklaşımı olacaktır.

Aynı şekilde, kaç tane 2 tabanı basamağı kullanmak isterseniz isteyin, 0.1 ondalık değeri tam olarak 2 tabanı kesri olarak gösterilemez. Taban 2’de 1/10 sonsuza kadar tekrar eden bir kesirdir

0.0001100110011001100110011001100110011001100110011...

Herhangi bir sonlu bit sayısında durduğunuzda bir yaklaşık değer elde edersiniz. Bugün çoğu makinede, kayan sayılar, pay en anlamlı bitten başlayarak ilk 53 bit kullanılarak ve payda ikinin kuvveti olarak ikili bir kesir kullanılarak yaklaştırılır. 1/10 durumunda ikili kesir, 1/10’un gerçek değerine yakın ancak tam olarak eşit olmayan 3602879701896397 / 2 ** 55 şeklindedir.

Many users are not aware of the approximation because of the way values are displayed. Python only prints a decimal approximation to the true decimal value of the binary approximation stored by the machine. On most machines, if Python were to print the true decimal value of the binary approximation stored for 0.1, it would have to display:

>>> 0.1
0.1000000000000000055511151231257827021181583404541015625

That is more digits than most people find useful, so Python keeps the number of digits manageable by displaying a rounded value instead:

>>> 1 / 10
0.1

Unutmayın, yazdırılan sonuç 1/10’un tam değeri gibi görünse de, saklanan gerçek değer temsil edilebilir olan en yakın ikili kesirdir.

İlginç bir şekilde, aynı en yakın yaklaşık ikili kesri paylaşan birçok farklı ondalık sayı vardır. Örneğin, 0.1 ve 0.10000000000000001 ve 0.1000000000000000055511151231257827021181583404541015625 sayılarının tümü 3602879701896397 / 2 ** 55 ile yaklaştırılır. Bu ondalık değerlerin tümü aynı yaklaşımı paylaştığından, eval(repr(x)) == x değişmezi korunarak bunlardan herhangi biri görüntülenebilir.

Geçmişte, Python komut istemi ve yerleşik repr() işlevi 17 anlamlı basamağa sahip olanı, 0.10000000000000001 ‘i seçerdi. Python 3.1’den itibaren, Python (çoğu sistemde) artık bunlardan en kısa olanı seçebilmekte ve basitçe 0.1 ‘i görüntüleyebilmektedir.

Bunun ikili kayan noktanın doğasında olduğunu unutmayın: bu Python’daki bir hata değildir ve kodunuzda daki bir hata da değildir. Donanımınızın kayan noktalı aritmetiğini destekleyen tüm dillerde aynı şeyi göreceksiniz (bazı diller varsayılan olarak veya tüm çıktı modlarında farkı göstermeyebilir).

For more pleasant output, you may wish to use string formatting to produce a limited number of significant digits:

>>> format(math.pi, '.12g')  # give 12 significant digits
'3.14159265359'

>>> format(math.pi, '.2f')   # give 2 digits after the point
'3.14'

>>> repr(math.pi)
'3.141592653589793'

Bunun gerçek anlamda bir yanılsama olduğunun farkına varmak önemlidir: gerçek makine değerinin görüntüsünü yuvarlıyorsunuz.

One illusion may beget another. For example, since 0.1 is not exactly 1/10, summing three values of 0.1 may not yield exactly 0.3, either:

>>> 0.1 + 0.1 + 0.1 == 0.3
False

Also, since the 0.1 cannot get any closer to the exact value of 1/10 and 0.3 cannot get any closer to the exact value of 3/10, then pre-rounding with round() function cannot help:

>>> round(0.1, 1) + round(0.1, 1) + round(0.1, 1) == round(0.3, 1)
False

Though the numbers cannot be made closer to their intended exact values, the math.isclose() function can be useful for comparing inexact values:

>>> math.isclose(0.1 + 0.1 + 0.1, 0.3)
True

Alternatively, the round() function can be used to compare rough approximations:

>>> round(math.pi, ndigits=2) == round(22 / 7, ndigits=2)
True

Binary floating-point arithmetic holds many surprises like this. The problem with “0.1” is explained in precise detail below, in the “Representation Error” section. See Examples of Floating Point Problems for a pleasant summary of how binary floating-point works and the kinds of problems commonly encountered in practice. Also see The Perils of Floating Point for a more complete account of other common surprises.

Söylendiği üzere, “kolay cevaplar yoktur.” Yine de, kayan nokta konusunda gereksiz yere temkinli olmayın! Python kayan nokta işlemlerindeki hatalar kayan nokta donanımından miras alınır ve çoğu makinede işlem başına 2**53’te 1 parçadan fazla değildir. Bu, çoğu görev için fazlasıyla yeterlidir, ancak bunun ondalık aritmetik olmadığını ve her kayan nokta işleminin yeni bir yuvarlama hatasına maruz kalabileceğini aklınızda bulundurmanız gerekir.

Patolojik durumlar mevcut olsa da, kayan noktalı aritmetiğin sıradan kullanımı için, nihai sonuçlarınızın görüntüsünü beklediğiniz ondalık basamak sayısına yuvarlarsanız, sonunda beklediğiniz sonucu görürsünüz. str() genellikle yeterlidir ve daha ince kontrol için Format String Syntax içindeki str.format() yönteminin biçim belirleyicilerine bakın.

Tam ondalık gösterim gerektiren durumlar için, muhasebe uygulamaları ve yüksek hassasiyetli uygulamalar için uygun ondalık aritmetiği uygulayan decimal modülünü kullanmayı deneyin.

Kesin aritmetiğin bir başka biçimi, rasyonel sayılara dayalı aritmetik uygulayan fractions modülü tarafından desteklenir (böylece 1/3 gibi sayılar tam olarak temsil edilebilir).

If you are a heavy user of floating-point operations you should take a look at the NumPy package and many other packages for mathematical and statistical operations supplied by the SciPy project. See <https://scipy.org>.

Python provides tools that may help on those rare occasions when you really do want to know the exact value of a float. The float.as_integer_ratio() method expresses the value of a float as a fraction:

>>> x = 3.14159
>>> x.as_integer_ratio()
(3537115888337719, 1125899906842624)

Since the ratio is exact, it can be used to losslessly recreate the original value:

>>> x == 3537115888337719 / 1125899906842624
True

The float.hex() method expresses a float in hexadecimal (base 16), again giving the exact value stored by your computer:

>>> x.hex()
'0x1.921f9f01b866ep+1'

This precise hexadecimal representation can be used to reconstruct the float value exactly:

>>> x == float.fromhex('0x1.921f9f01b866ep+1')
True

Temsil tam olduğundan, değerleri Python’un farklı sürümleri arasında güvenilir bir şekilde taşımak (platform bağımsızlığı) ve aynı biçimi destekleyen diğer dillerle (Java ve C99 gibi) veri alışverişi yapmak için kullanışlıdır.

Another helpful tool is the sum() function which helps mitigate loss-of-precision during summation. It uses extended precision for intermediate rounding steps as values are added onto a running total. That can make a difference in overall accuracy so that the errors do not accumulate to the point where they affect the final total:

>>> 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 + 0.1 == 1.0
False
>>> sum([0.1] * 10) == 1.0
True

The math.fsum() goes further and tracks all of the “lost digits” as values are added onto a running total so that the result has only a single rounding. This is slower than sum() but will be more accurate in uncommon cases where large magnitude inputs mostly cancel each other out leaving a final sum near zero:

>>> arr = [-0.10430216751806065, -266310978.67179024, 143401161448607.16,
...        -143401161400469.7, 266262841.31058735, -0.003244936839808227]
>>> float(sum(map(Fraction, arr)))   # Exact summation with single rounding
8.042173697819788e-13
>>> math.fsum(arr)                   # Single rounding
8.042173697819788e-13
>>> sum(arr)                         # Multiple roundings in extended precision
8.042178034628478e-13
>>> total = 0.0
>>> for x in arr:
...     total += x                   # Multiple roundings in standard precision
...
>>> total                            # Straight addition has no correct digits!
-0.0051575902860057365

15.1. Temsil Hatası

Bu bölüm “0.1” örneğini ayrıntılı olarak açıklamakta ve bu gibi durumların tam analizini kendiniz nasıl yapabileceğinizi göstermektedir. İkili kayan nokta gösterimine temel düzeyde aşina olunduğu varsayılmaktadır.

Temsil hatası, bazı (aslında çoğu) ondalık kesirlerin tam olarak ikili (taban 2) kesirler olarak temsil edilemeyeceği gerçeğini ifade eder. Bu, Python’un (veya Perl, C, C++, Java, Fortran ve diğerlerinin) genellikle beklediğiniz tam ondalık sayıyı göstermemesinin başlıca nedenidir.

Why is that? 1/10 is not exactly representable as a binary fraction. Since at least 2000, almost all machines use IEEE 754 binary floating-point arithmetic, and almost all platforms map Python floats to IEEE 754 binary64 “double precision” values. IEEE 754 binary64 values contain 53 bits of precision, so on input the computer strives to convert 0.1 to the closest fraction it can of the form J/2**N where J is an integer containing exactly 53 bits. Rewriting

1 / 10 ~= J / (2**N)

şu şekilde

J ~= 2**N / 10

and recalling that J has exactly 53 bits (is >= 2**52 but < 2**53), the best value for N is 56:

>>> 2**52 <=  2**56 // 10  < 2**53
True

That is, 56 is the only value for N that leaves J with exactly 53 bits. The best possible value for J is then that quotient rounded:

>>> q, r = divmod(2**56, 10)
>>> r
6

Since the remainder is more than half of 10, the best approximation is obtained by rounding up:

>>> q+1
7205759403792794

Therefore the best possible approximation to 1/10 in IEEE 754 double precision is:

7205759403792794 / 2 ** 56

Hem pay hem de paydayı ikiye böldüğünüzde kesir şuna indirgenir:

3602879701896397 / 2 ** 55

Aslında bölümü yukarı yuvarladığımız için değerin 1/10’dan biraz daha büyük olduğuna dikkat edin; yukarı yuvarlamamış olsaydık, bölüm 1/10’dan biraz daha küçük olurdu. Ancak hiçbir durumda tam olarak 1/10 olamaz!

So the computer never “sees” 1/10: what it sees is the exact fraction given above, the best IEEE 754 double approximation it can get:

>>> 0.1 * 2 ** 55
3602879701896397.0

If we multiply that fraction by 10**55, we can see the value out to 55 decimal digits:

>>> 3602879701896397 * 10 ** 55 // 2 ** 55
1000000000000000055511151231257827021181583404541015625

meaning that the exact number stored in the computer is equal to the decimal value 0.1000000000000000055511151231257827021181583404541015625. Instead of displaying the full decimal value, many languages (including older versions of Python), round the result to 17 significant digits:

>>> format(0.1, '.17f')
'0.10000000000000001'

The fractions and decimal modules make these calculations easy:

>>> from decimal import Decimal
>>> from fractions import Fraction

>>> Fraction.from_float(0.1)
Fraction(3602879701896397, 36028797018963968)

>>> (0.1).as_integer_ratio()
(3602879701896397, 36028797018963968)

>>> Decimal.from_float(0.1)
Decimal('0.1000000000000000055511151231257827021181583404541015625')

>>> format(Decimal.from_float(0.1), '.17')
'0.10000000000000001'