Factoring 17^17+2 - HP40GS success, 50G not
in [HP48]
|
From: John H Meyers on 9 Oct 2007 02:31 On Mon, 08 Oct 2007 14:42:06 -0500: > I wish they would omit the 60 second time-out. Since ON would break > the [FACTOR] command anyway, they should have left it to the user to break. Agreed; all models (and all emulators), running at any clock rates, and with or without any intermediate "garbage collection" randomly varying the execution time, should get the same output for the same math functions. "I can not believe that God plays dice with math funtions" (what Einstein might have said about this :) Potentially long operations should be interruptible, or error on inability to complete, or supply a yes/no (true/false) indication of whether they completed; it has only caused embarrassment in comparison with other calculators to have silently quit and delivered time-dependent (and wrong) answers for larger inputs in general, although the specific example given in this thread actually seems to beat the clock all the time, and to give invariably correct results (what did "50G misses the factor 2243" ever mean?") [r->] [OFF]
From: Mike Bryant on 9 Oct 2007 11:34 On Oct 8, 12:42 pm, sc_use...(a)hotmail.com wrote: > On Oct 8, 2:45 pm, "Greg M." <grego...(a)insightbb.com> wrote: > > > > > Bicky-Bock wrote: > > > After re-booting HP 50G number now factored by FACTORS correctly - besides > > > false factoring I noticed no other problems. Serial Nr CNA 61807444. > > > Perhaps your flags settings caused slower operation of FACTOR. > > > I think FACTOR times out after 1 min. if it can't find any factors. > > > Please see a post I wrote awhile back: > > >http://groups.google.com/group/comp.sys.hp48/msg/259cacd1cfc6843e > > > This might not be the problem, though, since my 49g+ only takes ~6 seconds > > to factor 17^17+2, which is nowhere near 60s. > > > Greg M. > > I wish they would omit the 60 second time-out. Since ON would break > the routine anyway, they should have left it to the user to break. > > In your post, you said that the 50g took 532 seconds to factor 2^67-1. > A HW2 TI-89 took 213 seconds. I wonder what kind of algorithm those > guys in Texas are using. Then again, the FACTOR routine is probably > emulated on the ARM machines. > > Of course, PARI can factor this in 15 ms. :) My 50g did it in ~106 sec.
From: Steen Schmidt on 12 Oct 2007 14:25
Mike Bryant wrote: > > In your post, you said that the 50g took 532 seconds to factor > > 2^67-1. A HW2 TI-89 took 213 seconds. I wonder what kind of > > algorithm those guys in Texas are using. Then again, the FACTOR > > routine is probably emulated on the ARM machines. > > > > Of course, PARI can factor this in 15 ms. :) > > My 50g did it in ~106 sec. Mine did it in a bit over 900 ms with most of that stop and start of the Saturn-emulator, reconfiguring etc. (i.e. HPGCC overhead). Cheers, Steen |