浮動小数点演算、その問題と制限
******************************

浮動小数点数は、計算機ハードウェアの中では、基数を 2 とする (2進法の)
分数として表現されています。例えば、小数

   0.125

は、 1/10 + 2/100 + 5/1000 という値を持ちますが、これと同様に、2 進法
の分数

   0.001

は 0/2 + 0/4 + 1/8 という値になります。これら二つの分数は同じ値を持っ
ていますが、ただ一つ、最初の分数は基数 10 で記述されており、二番目の分
数は基数 2 で記述されていることが違います。

残念なことに、ほとんどの小数は 2 進法の分数として正確に表わすことがで
きません。その結果、一般に、入力した 10 進の浮動小数点数は、 2 進法の
浮動小数点数で近似された後、実際にマシンに記憶されます。

最初は基数 10 を使うと問題を簡単に理解できます。分数 1/3 を考えてみま
しょう。分数 1/3 は、基数 10 の分数として、以下のように近似することが
できます:

   0.3

さらに正確な近似は、

   0.33

さらに正確な近似は、

   0.333

となり、以後同様です。何個桁数を増やして書こうが、結果は決して厳密な
1/3 にはなりません。しかし、少しづつ正確な近似にはなっていくでしょう。

同様に、基数を 2 とした表現で何桁使おうとも、10 進数の 0.1 は基数を 2
とした小数で正確に表現することはできません。基数 2 では、1/10 は循環小
数 (repeating fraction) となります

   0.0001100110011001100110011001100110011001100110011...

Stop at any finite number of bits, and you get an approximation.

On a typical machine running Python, there are 53 bits of precision
available for a Python float, so the value stored internally when you
enter the decimal number "0.1" is the binary fraction

   0.00011001100110011001100110011001100110011001100110011010

which is close to, but not exactly equal to, 1/10.

It’s easy to forget that the stored value is an approximation to the
original decimal fraction, because of the way that floats are
displayed at the interpreter prompt.  Python only prints a decimal
approximation to the true decimal value of the binary approximation
stored by the machine.  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

これは、ほとんどの人が必要と感じるよりも多すぎる桁数です。なので、
Python は丸めた値を表示することで、桁数を扱いやすい範囲にとどめます

   >>> 0.1
   0.1

It’s important to realize that this is, in a real sense, an illusion:
the value in the machine is not exactly 1/10, you’re simply rounding
the *display* of the true machine value.  This fact becomes apparent
as soon as you try to do arithmetic with these values

   >>> 0.1 + 0.2
   0.30000000000000004

この動作は2進数の浮動小数点にとってはごく自然なものです。これは Python
のバグではありませんし、あなたのコードのバグでもありません。ハードウェ
アの浮動小数点演算をサポートしている全ての言語で同じ種類の問題を見つけ
ることができます (いくつかの言語ではデフォルトの、あるいはどの出力モー
ドを選んでも、この差を **表示** しないかもしれませんが)。

Other surprises follow from this one.  For example, if you try to
round the value 2.675 to two decimal places, you get this

   >>> round(2.675, 2)
   2.67

The documentation for the built-in "round()" function says that it
rounds to the nearest value, rounding ties away from zero.  Since the
decimal fraction 2.675 is exactly halfway between 2.67 and 2.68, you
might expect the result here to be (a binary approximation to) 2.68.
It’s not, because when the decimal string "2.675" is converted to a
binary floating-point number, it’s again replaced with a binary
approximation, whose exact value is

   2.67499999999999982236431605997495353221893310546875

Since this approximation is slightly closer to 2.67 than to 2.68, it’s
rounded down.

If you’re in a situation where you care which way your decimal
halfway-cases are rounded, you should consider using the "decimal"
module. Incidentally, the "decimal" module also provides a nice way to
「see」 the exact value that’s stored in any particular Python float

   >>> from decimal import Decimal
   >>> Decimal(2.675)
   Decimal('2.67499999999999982236431605997495353221893310546875')

Another consequence is that since 0.1 is not exactly 1/10, summing ten
values of 0.1 may not yield exactly 1.0, either:

   >>> sum = 0.0
   >>> for i in range(10):
   ...     sum += 0.1
   ...
   >>> sum
   0.9999999999999999

2 進の浮動小数点数に対する算術演算は、このような意外性をたくさん持って
います。 「0.1」 に関する問題は、以下の 「表現エラー」 の章で詳細に説
明します。 2 進法の浮動小数点演算にともなうその他のよく知られた意外な
事象に関しては The Perils of Floating Point を参照してください。

As that says near the end, 「there are no easy answers.」  Still,
don’t be unduly wary of floating-point!  The errors in Python float
operations are inherited from the floating-point hardware, and on most
machines are on the order of no more than 1 part in 2**53 per
operation.  That’s more than adequate for most tasks, but you do need
to keep in mind that it’s not decimal arithmetic, and that every float
operation can suffer a new rounding error.

While pathological cases do exist, for most casual use of floating-
point arithmetic you’ll see the result you expect in the end if you
simply round the display of your final results to the number of
decimal digits you expect.  For fine control over how a float is
displayed see the "str.format()" method’s format specifiers in 書式指
定文字列の文法.


表現エラー
==========

この章では、」0.1」 の例について詳細に説明し、このようなケースに対して
どのようにすれば正確な分析を自分で行えるかを示します。ここでは、 2 進
法表現の浮動小数点数についての基礎的な知識があるものとして話を進めます
。

*Representation error* refers to the fact that some (most, actually)
decimal fractions cannot be represented exactly as binary (base 2)
fractions. This is the chief reason why Python (or Perl, C, C++, Java,
Fortran, and many others) often won’t display the exact decimal number
you expect:

   >>> 0.1 + 0.2
   0.30000000000000004

Why is that?  1/10 and 2/10 are not exactly representable as a binary
fraction. Almost all machines today (July 2010) use IEEE-754 floating
point arithmetic, and almost all platforms map Python floats to
IEEE-754 「double precision」.  754 doubles 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)

を書き直すと

   J ~= 2**N / 10

となります。 *J* は厳密に 53 ビットの精度を持っている (">= 2**52" だが
"< 2**53" ) ことを思い出すと、 *N* として最適な値は 56 になります:

   >>> 2**52
   4503599627370496
   >>> 2**53
   9007199254740992
   >>> 2**56/10
   7205759403792793

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

剰余が 10 の半分以上なので、最良の近似は切り上げて丸めたものになります
。

   >>> q+1
   7205759403792794

Therefore the best possible approximation to 1/10 in 754 double
precision is that over 2**56, or

   7205759403792794 / 72057594037927936

丸めたときに切り上げたので、この値は実際には 1/10 より少し大きいことに
注目してください。 もし切り捨てをした場合は、商は 1/10 よりもわずかに
小さくなります。どちらにしろ *厳密な* 1/10 ではありません！

つまり、計算機は 1/10 を 「理解する」 ことは決してありません。計算機が
理解できるのは、上記のような厳密な分数であり、 754 の倍精度浮動小数点
数で得られるもっともよい近似は以下になります:

   >>> .1 * 2**56
   7205759403792794.0

If we multiply that fraction by 10**30, we can see the (truncated)
value of its 30 most significant decimal digits:

   >>> 7205759403792794 * 10**30 // 2**56
   100000000000000005551115123125L

meaning that the exact number stored in the computer is approximately
equal to the decimal value 0.100000000000000005551115123125.  In
versions prior to Python 2.7 and Python 3.1, Python rounded this value
to 17 significant digits, giving 『0.10000000000000001』.  In current
versions, Python displays a value based on the shortest decimal
fraction that rounds correctly back to the true binary value,
resulting simply in 『0.1』.
