Re: [Ql-Users] Arithmetic precision
On 10 May 2011, at 19:20, Dave Walker wrote: Dilwyn, Writing such routines from scratch is likely to be a non-trivial task! I would agree with this. The c68 library has a whole set of support routines written in assembler for this purpose that conform to the IEEE standard for Double Precision floating point. I suspect it would not be that hard to put some Basic keywords at the front-end of them to make them accessible from SuperBasic. They have the advantage that they have been put through a formal test suite to check that they give the correct answers which might be important? If you could give an indication of how you expect to use them then I could have a quick look to see if I think the c68 routines could be adapted to that purpose. It would be nice to know what operations you want to carry out if it is more than simply arithmetic operations. Possibly important is how you expect to load up and extract the values without losing precision. Of course if you wanted to use them from C then they are ready to go - you just need to use the 'long double' data type. I would think it MUCH easier to use C68 to do the arithmetic. A small program could be started by EW from a BASIC program. Dave Walker Tel: +44 (0)1707 652791 Mob: +44 (0)7999 218953 Web: http://www.itimpi.com Skype: itimpi -Original Message- From: ql-users-boun...@lists.q-v-d.com [mailto:ql-users- boun...@lists.q-v-d.com] On Behalf Of Dilwyn Jones Sent: 10 May 2011 18:14 To: ql-users@lists.q-v-d.com Subject: [Ql-Users] Arithmetic precision Does anyone know of any routines for the QL which can handle double precision floating point maths? When asked, all I could think of was Helmut AInger's math toolkit, but I haven't looked at it yet. Alternatively, if anyone knows how to write such a routine (a BASIC demo would do) or has a sample listing showing how to achieve this I'd be very grateful. At this stage, accuracy is more important than speed. Dilwyn Jones George ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] Arithmetic precision
Dilwyn, Writing such routines from scratch is likely to be a non-trivial task! The c68 library has a whole set of support routines written in assembler for this purpose that conform to the IEEE standard for Double Precision floating point. I suspect it would not be that hard to put some Basic keywords at the front-end of them to make them accessible from SuperBasic. They have the advantage that they have been put through a formal test suite to check that they give the correct answers which might be important? If you could give an indication of how you expect to use them then I could have a quick look to see if I think the c68 routines could be adapted to that purpose. It would be nice to know what operations you want to carry out if it is more than simply arithmetic operations. Possibly important is how you expect to load up and extract the values without losing precision. Of course if you wanted to use them from C then they are ready to go - you just need to use the 'long double' data type. Dave Walker Tel: +44 (0)1707 652791 Mob: +44 (0)7999 218953 Web: http://www.itimpi.com Skype: itimpi -Original Message- From: ql-users-boun...@lists.q-v-d.com [mailto:ql-users- boun...@lists.q-v-d.com] On Behalf Of Dilwyn Jones Sent: 10 May 2011 18:14 To: ql-users@lists.q-v-d.com Subject: [Ql-Users] Arithmetic precision Does anyone know of any routines for the QL which can handle double precision floating point maths? When asked, all I could think of was Helmut AInger's math toolkit, but I haven't looked at it yet. Alternatively, if anyone knows how to write such a routine (a BASIC demo would do) or has a sample listing showing how to achieve this I'd be very grateful. At this stage, accuracy is more important than speed. Dilwyn Jones ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm
Re: [Ql-Users] Arithmetic precision
Dave Walker wrote: Dilwyn, Writing such routines from scratch is likely to be a non-trivial task! The c68 library has a whole set of support routines written in assembler for this purpose that conform to the IEEE standard for Double Precision floating point. I suspect it would not be that hard to put some Basic keywords at the front-end of them to make them accessible from SuperBasic. They have the advantage that they have been put through a formal test suite to check that they give the correct answers which might be important? If you could give an indication of how you expect to use them then I could have a quick look to see if I think the c68 routines could be adapted to that purpose. It would be nice to know what operations you want to carry out if it is more than simply arithmetic operations. Possibly important is how you expect to load up and extract the values without losing precision. Of course if you wanted to use them from C then they are ready to go - you just need to use the 'long double' data type. Dave Walker I have been trying to help Steve Poole with a program he is writing. I do not fully understand the need for such routines, although he has tried to explain (without much success!) what he is trying to achieve. The routines would need to be usable from BASIC, perhaps as BASIC extensions. Never having looked at anything like this, I simply didn't know what it would involve. Hence my request for information. Dilwyn Jones ___ QL-Users Mailing List http://www.q-v-d.demon.co.uk/smsqe.htm