May 2003 - Fractdev - mailman.xmission.com

Phew! No "conjbug"
by Tim Wegner 17 Jun '03

17 Jun '03

Jonathan, there's no "conjbug". See my long message in the other forum. In researching this I noticed that the Xfractint function: void FPUcplxlog_CMPLX *x, _CMPLX *z) { double mod,zx,zy; mod = sqrt(x->x*x->x + x->y*x->y); zx = log(mod); zy = atan2(x->y,x->x); z->x = zx; z->y = zy; } could be speeded up by getting rid of the sqrt() at the cost of a multiply: void FPUcplxlog_CMPLX *x, _CMPLX *z) { double zx,zy; zx = .5*log(x->x*x->x + x->y*x->y); zy = atan2(x->y,x->x); z->x = zx; z->y = zy; } There's no downside that I can see, the two are mathematically equivalent, and the faster version is just as readable. I didn't benchmark the different but I'll bet it's measurable. Tim

2 3

Xfractint boundary tracing
by Jonathan Osuch 17 Jun '03

17 Jun '03

Tim, Do you have any recollection of any platform differences that would cause problems with enumerated data types? The Xfractint boundary tracing ends up in an infinite loop (actually there are several different loops) because the coming_from variable (of enumerated type direction) is ending up with values that are out of the bounds of the type. Jonathan

4 20

Power bug
by Tim Wegner 03 Jun '03

03 Jun '03

The function ComplexPower() in mpmath_c.c apparently has a bug. The bug appears to be in the Linux also - I tried runninmg it, but haven't checked the code. If so, someone did an incredible job of translating the assembler! I instrumented StkPwr in parser.c to print out the arguments and the result. I found this wehnen running Jim White's example: Arg1: 0.166666666666667 0.000000000000000 Arg2: -5.000000000000000 0.000000000000000 Result: 1.132467200411351 -0.653830243005915 Arg1: 0.166666666666667 0.000000000000000 Arg2: -5.000000000000000 0.000000000000000 Result: 1.132467200411351 0.653830243005915 In two subsequent calls, the sign of the imaginary part changed despite the arguments being the same!! The problem is in ComplexPower() in mpmath_c.c. I was able to make the symtoms go away by replacing the FPUcplxlog function with C: /* FPUcplxlog(&xx, &cLog); */ cLog.x = .5*log(xx.x*xx.x + xx.y*xx.y); cLog.y = atan2(xx.y,xx.x); That's as far as I got. WIll take this up again soon. Tim

2 6

Non deterministic log
by Tim Wegner 31 May '03

31 May '03

Jonathan, I thought I understood the "conjbug", but I discovered I don't completely. I am convinced that it is no big deal, but I don't understand why FPUcplxlog() is not behaving deterministically. When (-5,0) is passed to ComplexPower() in mpmath_c.c, sometimes the imaginary part is PI and sometimes -PI. As I explained elsewhere, the -5 argument is right on the edge of a discontinuity, so these two values aren't really different, but since I have verified that the argument is EXACTLY the same in both cases, I don't see why the same result is not returned. I'll try to make a little test case that duplicates the problem. I wonder if there could be some FPU status flag that gets set by some multiplies that happen first in one case but not the other. I'll try to make a little test program that illustrates the problem. If you have a chance, look at FPUcplxlog and tell me if you see anything suspect in the logic that determines the sign of the imaginary component. Tim

2 2

indent issues
by Jonathan Osuch 19 May '03

19 May '03

Tim, Running indent again on the same source produces more changes. It looks like some comments are getting put all the way to the right and running indent again moves them to the right some more. The man page has a comment about this. I'll see if there is anything else. Jonathan

2 4

Virtual video test results
by Tim Wegner 09 May '03

09 May '03

Jonathan asked: > Any results? HP-Pavilion with NVIDIA Vanta This works OK with virtual=no, but with virtual=yes, even non-virtual SVGA modes have problems going back to text. Something as simple as SF4 ESC gives a scrambled screen. BTW, it is possible that the virtual feature works, but since it messes up the text screens, I have no way to set it up to try it. I'm inclined to think that the virtual windows parameters should be settable even if the current mode is not a virtual mode. There could be words on the screen to the effect that these parameters only take effect in virtual mode. Dell Optiplex with ATI 3D Rage Pro - virtual screen modes works fine. Sony VAIO laptop with Rage Mobility 1 M AGP - works fine. Old no-name 120 Mhz Pentium with Diamond VIPER V550. Same behavior as HP - OK with virtual=no, scrambled screen otherwise. Summary. ATI video boards are OK, but Nvidia boards have problems. I believe the Diamond viper uses an Nvidia chipset. Since virtual=no works OK, I suggest you release this patch and let our loyal public help tabulate what works and what doesn't. We'll get more feedback that way. Or at least, move the proposed patched version to a public area. Tim

4 45