Search found 196 matches
- Fri Mar 22, 2024 7:42 pm
- Forum: Software Library
- Topic: Happy Pi Day!!!
- Replies: 8
- Views: 565
Re: Happy Pi Day!!!
Ron Knapp later published an enhanced routine, in PPC Journal V8N6, hardcoded for 1000 digits, that ran in 8h30 on a ‘41. Allegedly, no one has ever been able to beat that. Using Ron’s original idea, I optimised (and corrected..) his routine, and estimate (as my 41 is long dead, but I have a DM41) t...
- Fri Mar 15, 2024 6:45 pm
- Forum: DM32 BETA Bugs
- Topic: DM32 Numerical Precision for Transcendental Functions
- Replies: 14
- Views: 923
Re: DM32 Numerical Precision for Transcendental Functions
If the DM32 and DM42 are both using the Intel decimal floating-point library, why is the value of SIN(31°) different? (1 ulp, but still)
Werner
Werner
- Tue Mar 12, 2024 9:07 am
- Forum: DM32 BETA Bugs
- Topic: DM32 Numerical Precision for Transcendental Functions
- Replies: 14
- Views: 923
Re: DM32 Numerical Precision for Transcendental Functions
The trig values of the DM32 are not (always) correctly rounded, it seems. Your example of SIN(31°) isn't, for instance. The C47/WP34S etc use 39-digit intermediate precision to get them right, I guess the DM32 does not. That would explain the higher number of 'misses'.
Cheers, Werner
Cheers, Werner
- Mon Mar 11, 2024 11:08 am
- Forum: DM32 BETA Bugs
- Topic: DM32 Numerical Precision for Transcendental Functions
- Replies: 14
- Views: 923
Re: DM32 Numerical Precision for Transcendental Functions
(update: note to self: read a bit more and take the time to understand a post before posting stupid replies..) Why would you expect the identity sin(x)^2+cos(x)^2=1 to hold? You're not working with SIN(31°) and COS(31°), you're working with their 34-digit (rounded or truncated) approximations. It's ...
- Wed Jan 03, 2024 11:33 am
- Forum: Discuss!
- Topic: Wishing You a Happy 2024!
- Replies: 20
- Views: 5286
Re: Wishing You a Happy 2024!
One more to wish you and everyone else here a happy, prosperous, healthy 2024!
Werner
Werner
- Mon Nov 20, 2023 9:38 am
- Forum: Software Library
- Topic: QuickSort using recursion with LSTO command
- Replies: 2
- Views: 842
- Fri Aug 18, 2023 8:53 am
- Forum: Software Library
- Topic: GCD - Greatest Common Divisor
- Replies: 6
- Views: 2112
Re: GCD - Greatest Common Divisor
Thanks, Joe. I noticed you used only the bare bones Rv, X<>Y and LASTX, so this'll work on any HP calc
And what a bummer that the 15C of all models does not have the MOD function...
Cheers, Werner
- Fri Aug 18, 2023 8:49 am
- Forum: Software Library
- Topic: GCD - Greatest Common Divisor
- Replies: 6
- Views: 2112
Re: GCD - Greatest Common Divisor
Of course, for the 42S and its compatible successors, this is the ultimate LCM/GCD program (compatible with the 41, even): 00 { 32-Byte Prgm } 01▸LBL "LCM" 02 RCL ST Y 03 RCL ST Y 04 XEQ 03 05 ÷ 06 × 07 RTN 08▸LBL "GCD" 09▸LBL 03 10 MOD 11 LASTX 12 X<>Y 13 X≠0? 14 GTO 03 15 + 16 ...
- Sat Jul 08, 2023 8:49 am
- Forum: Software Library
- Topic: Summer challenge
- Replies: 16
- Views: 4725
- Fri Jul 07, 2023 2:43 pm
- Forum: Software Library
- Topic: Summer challenge
- Replies: 16
- Views: 4725
Re: Summer challenge
I don't like programs that include their own input routines, so here's a different approach, in 42S-compatible code: X Y Z T In: y (x2,y2) (x1,y1) Out: x y b a 00 { 35-Byte Prgm } X Y Z T 01▸LBL "INTP" @ y p2 p1 02 X<> ST Z 03 STO- ST Y @ p1 p2-p1 y 04 X<>Y 05 COMPLEX @ y2-y1 x2-x1 p1 y 06...