Search found 737 matches

by Thomas Okken
Thu Oct 07, 2021 2:26 am
Forum: Discuss!
Topic: 42s or 41x? Or wait for 43S
Replies: 10
Views: 994

Re: 42s or 41x? Or wait for 43S

Would really like the DM42L that they beta demo'd several years ago but I guess it is an idea set aside... That looked like a proof of concept, just something to verify that their low-power ARM platform was capable of running Free42 with acceptable performance. They used what looked like the same h...
by Thomas Okken
Sun Oct 03, 2021 12:26 pm
Forum: Usage tips, tricks and problem reports
Topic: Big stack and Σ+
Replies: 1
Views: 206

Re: Big stack and Σ+

Confirmed! That's a bug.
I'll fix it in the next release.
by Thomas Okken
Sun Sep 12, 2021 7:56 pm
Forum: Usage tips, tricks and problem reports
Topic: Big stack queries
Replies: 2
Views: 315

Re: Big stack queries

2. When pushing ENTER after data entry, the data is entered into the x register, without lifting the stack and duplicating the data, as HP42 normally would. This means that the function of the ENTER key is now duplicating the effect of pushing EXIT. Is this intended behavior? Yes. Note that this is...
by Thomas Okken
Tue Sep 07, 2021 9:52 pm
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

N.B. I don't know how to check Werner's test cases. The code, as it is, is not valid (PGMINT "Jn" but LBL "JNX") and it is not clear to me what the parameter N is supposed to be set to.
by Thomas Okken
Tue Sep 07, 2021 7:38 pm
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

Strangely, when I rewrite the routines to have but a single global label, as follows: [...] It now takes 4 million evaluations of the integrand as well for ACC=1E-4. Strange? It should be exactly the same. Are you sure the label is the only difference? I don't see how that could affect anything abo...
by Thomas Okken
Mon Sep 06, 2021 5:53 pm
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

The problem isn't in SOLVE. With ACC=1e-4, the root is found in 22 evaluations of BESSEL. The problem is that INTEG takes a long time: JNX is evaluated 4,218,730 times before the root is found.
by Thomas Okken
Sun Sep 05, 2021 3:06 pm
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

Actually, this came up in this forum recently, and I think the case under discussion that time didn't involve the integral and absolute error estimates spiraling towards zero, but rather, non-improvement, i.e. the estimate of the integral was just bouncing around and not improving at all. Either way...
by Thomas Okken
Sun Sep 05, 2021 5:07 am
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

Ah, that looks familiar. The function you're solving is an integral, so in the process of approaching a root, INTEG is having to evaluate an integral that is close to zero. And that is something the Free42 INTEG doesn't handle well, because such integrals appear not to converge, because the estimate...
by Thomas Okken
Sat Sep 04, 2021 7:53 pm
Forum: DM42
Topic: DM42/Free42 - which algorithm used in SOLVE ?
Replies: 16
Views: 1073

Re: DM42/Free42 - which algorithm used in SOLVE ?

SOLVE in Free42 uses the secant method, until it has two guesses that evaluate to values with opposite signs. It then proceeds by using Ridders root refinement, falling back on the secant method, or bisection as a last resort, if the function is not well-behaved. If you could share instructions how ...
by Thomas Okken
Tue Aug 31, 2021 9:54 pm
Forum: DM42
Topic: Will there be a full 2 part update this August?
Replies: 7
Views: 780

Re: Will there be a full 2 part update this August?

So we can assume that when they started to work on the August update the version of Free42 that qualified was 3.05. And they didn't wanted to restart their work to include a newer version later on, as they had also to work on a new firmware for the DM1x/DM41(L) in addition to the new product they t...