Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: bart <bc@freeuk.com>
Newsgroups: comp.lang.c
Subject: Re: Python recompile
Date: Tue, 11 Mar 2025 21:37:36 +0000
Organization: A noiseless patient Spider
Lines: 37
Message-ID: <vqqaf0$267fp$2@dont-email.me>
References: <vq1qas$j22$1@gallifrey.nk.ca> <vq6j5h$1qosf$1@dont-email.me>
 <20250304092827.708@kylheku.com> <vq7g1p$1vmg5$1@dont-email.me>
 <vq94dt$2boso$1@dont-email.me> <vqcsk7$23bfo$1@paganini.bofh.team>
 <vqefn1$3flpt$1@dont-email.me> <vqeu5c$3imil$1@dont-email.me>
 <vqeun4$3iqbq$1@dont-email.me> <vqfcbe$3lkkc$1@dont-email.me>
 <871pv861ht.fsf@nosuchdomain.example.com> <20250308192940.00001351@yahoo.com>
 <vqi1ge$8jg8$1@dont-email.me> <vqmgjv$3a2il$1@paganini.bofh.team>
 <vqn4dn$1eb9s$1@dont-email.me> <vqo3ss$3hkas$1@paganini.bofh.team>
 <vqph2e$203bs$2@dont-email.me> <vqpjh7$210q9$1@dont-email.me>
 <vqpo1s$222s0$1@dont-email.me> <vqpqo6$23197$1@dont-email.me>
 <vqpsvc$23gc1$1@dont-email.me> <20250311201757.000045e2@yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Injection-Date: Tue, 11 Mar 2025 22:37:37 +0100 (CET)
Injection-Info: dont-email.me; posting-host="e0b1512e31c88b1cbe05a7f62cb3028b";
	logging-data="2301433"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19fM7xWviTAXHinSbDg7XeH"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:J5Fc7FfjgfExI2/FX7tG+tY76jM=
In-Reply-To: <20250311201757.000045e2@yahoo.com>
Content-Language: en-GB
Bytes: 3265

On 11/03/2025 18:17, Michael S wrote:
> On Tue, 11 Mar 2025 17:47:24 +0000
> bart <bc@freeuk.com> wrote:
> 
>> Performance is roughly on a par with my library.
>>
> 
> You can not appreciate performance of performace of GMP until you use
> it with BIG numbers. Not 100 or 1,000 digits, although for 1,000
> digits it likely beats your library by a good margin. Try to multiply
> numbers with 100,000 or more digits. Then you will see really massive
> difference.

I meant performance of mini-gmp. But I just tried it on a couple of 
tests, and mini-gmp was 80% faster, when both are optimised the same way.

But both are magnitudes slower than the full GMP, at least with bigger 
precision. (The dlls you linked didn't work; the one that David Brown 
vaguely linked to, which I eventually found, did work.) However I needed 
all these years ago.

> As to floating-point, GMP used to have it, but long ago decided that it
> is out of scope. Support for floating-point is still here for historical
> reasons, but for new software they recommend to use MPFR.


There are special issues with arbitrary precision floating point. For 
example, if, you do 1.0/3.0, it scan spend forever working out 0.3333... 
until it runs out of memory, or just takes too long.

I only support basic arithmetic, with configurable caps on precision. 
Doing stuff like trig functions, for which I used to use Taylor Series, 
would be challenging (in finding something that converges to a specific 
precision in a reasonable time).

But the whole thing is just a bit of fun.