Re: V8 slow like a snail

*To*: mathgroup at smc.vnet.net*Subject*: [mg127726] Re: V8 slow like a snail*From*: "Dr. Wolfgang Hintze" <weh at snafu.de>*Date*: Fri, 17 Aug 2012 03:43:24 -0400 (EDT)*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com*Delivered-to*: l-mathgroup@wolfram.com*Delivered-to*: mathgroup-newout@smc.vnet.net*Delivered-to*: mathgroup-newsend@smc.vnet.net*References*: <20120815073325.F37376842@smc.vnet.net> <k0i2mv$1e0$1@smc.vnet.net>

On 16 Aug., 08:07, Murray Eisenberg <mur... at math.umass.edu> wrote: > You said for which values of m you timed this, but what values of n? > > On Aug 15, 2012, at 3:33 AM, Dr. Wolfgang Hintze <w... at snafu.de> wrote: > > > f1[n_, m_] := > > Integrate[n t^m Exp[-n t] (Exp[t] - 1)^(n - 1), {t, 0, \[Infinity]}, > > Assumptions -> {{n, m} \[Element] Integers, m >= 0, n > 0}] > > --- > Murray Eisenberg = mur... at math.umass.edu > Mathematics & Statistics Dept. > Lederle Graduate Research Tower phone 413 549-1020 (H) > University of Massachusetts 413 545-28= 59 (W) > 710 North Pleasant Street fax = 413 545-1801 > Amherst, MA 01003-9305 Murray, no value of n was specified, as I am interested in symbolic calculation, and in the time the two versions need. I'll show the example m=1 Version 8: Timing[Integrate[n*t*Exp[(-n)*t]*(Exp[t] - 1)^(n - 1), {t, 0, Infinity}]] {30.840999999999994, ConditionalExpression[HarmonicNumber[n], Re[n] > -1]} So v8 is able to perform the symbolic calculation and needs in 30.84 seconds. Using the Assumptions -> {{n, m} \[Element] Integers, m >= 0, n > 0} leads to the same result in 30.43 seconds Now version 5.2: Timing[Integrate[n*t*Exp[(-n)*t]*(Exp[t] - 1)^(n - 1), {t, 0, Infinity}]] {0.5 Second, n If[Re[n] > -1, HarmonicNumber[n]/n, Integrate[E^(-n t) (-1 + E^t)^(-1 + n) t, {t, 0, \[Infinity]}, Assumptions -> Re[n] <= -1]]} so it does the job with the same result (including a suggestion about the assumptions) as v8 but 60 times faster! With the Assumptions: {0.109 Second, HarmonicNumber[n]} hence v5.2 now is even 30/0.11 = 270 times faster than version 8. My point is that I'm observing factors of 60 or 270 longer calculation times for version 8 in comparison to version 5.2 The integrals of my example are by no means exotic. They are just the m-th moments of the maximum of n independent exponentially distributed random variables. My second example I put already in my first reply to the newsgroup (I repeat it here for ease of reference): Version 8: Timing[f = FullSimplify[Integrate[Log[Abs[(Tan[x] + Sqrt[7])/(Tan[x] - Sqrt[7])]], {x, Pi/3, Pi/2}]]] {3191.391, Integrate[Log[Abs[(Sqrt[7] + Tan[x])/(Sqrt[7] - Tan[x])]], {x, Pi/3, Pi/2}]} Hence no result (and no comments) in about one hour. Version 5.2 Timing[f = FullSimplify[Integrate[ Log[Abs[(Tan[x] + Sqrt[7])/(Tan[x] - Sqrt[7])]], {x, Pi/3, Pi/ 2}]]] {32.667*Second, (1*(-9*ArcCot[Sqrt[7]]*Log[2] - (Pi - 6*ArcCot[Sqrt[7]])* Log[-Sqrt[3] + Sqrt[7]] + Pi*Log[Sqrt[3] + Sqrt[7]] + 3*I*(-PolyLog[2, (-Sqrt[3] + Sqrt[7])/(-I + Sqrt[7])] + PolyLog[2, (-Sqrt[3] + Sqrt[7])/(I + Sqrt[7])] + PolyLog[2, (-I + Sqrt[7])/(Sqrt[3] + Sqrt[7])] - PolyLog[2, (I + Sqrt[7])/(Sqrt[3] + Sqrt[7])])))/6} so v 5.2 gives the result in terms of known functions and needs only half a minute. Can the difference be larger? Again the integral was not selected by some "malicious" procedure, instead it appeared just recently in my Spanish group of recreational mathematics (es.ciencia. matematicas). Best regards, Wolfgang

**References**:**V8 slow like a snail***From:*"Dr. Wolfgang Hintze" <weh@snafu.de>