It's Windows XP sp2. It's laptop, I checked the CPU is a Intel Centrino Mobile M 740, which is 1.73GHz and it supports only up to SSE2. The system bus is 533MHz, and the L2 Cache is 2MB, and L1 data chase 32KB, L1 instruction cache 32KB. Processor classification: CPU Type 0, Family 6, Model D, Stepping 8, Revision 20. I have 2GB RAM and virtual memory page size 2.302GB, as recommended by the system. C Drive total available: 7.074GB. I thought everybody has the same behavior. For me, it is very common now, if I have been using the session for a long time, doing some non-heavy duty stuff, then when I want to do a heavy duty work, I reboot my PC to get a clear state.

Hi JacquesC, Thanks a lot! This is very skillful! May I confirm that you got two answers and they are basically the same? I've got ans is much longer than ans2. I guess the difference is that ans2 is of order 1, and ans is of order 2, which makes it very long. May I say that (ans2)^(2*a*xbar/sigma2) is the correct final value? Two more questions: 1. I don't understand the "-" in "ans2 := evalc(MultiSeries:-asympt(evalc(ans), v,1)):"? 2. I cannot see the error of "sorting exponents is still there"? Is this result correct or wrong? Thanks again!

When I did: ------------- asympt(evalc(abs(CCP1)), v) assuming positive; It complained: Error, (in assuming) when calling 'asympt'. Received: 'unable to compute series'. ------------- What's wrong? Please help me!

Thanks a lot JacquesC!

Hi dharr, It's a very complicated expression. And I observe it has very fast decay, which leads me to my suspicion that as x -> infinity, f(x) < exp(-a*x)*b/(x+c)^(n). If I can find a, b, and c. Then I will be able to truncate the function and estimate the truncation error. I am currently using that function as an integrand in my numerical integration. By truncation, I will be able to make it a finite integral. Any more thoughts? Thanks a lot!

When you do: C((3), optimize), where (3) is referring to an equation defined as above. It gave an error message: "Error, (in PrintTarget) assigning to a long list, please use Arrays", what's wrong? Thanks!

Does anybody know how to let the C code generator precisely translate the greeks I used in 2D math into their English name? For example, the "epsilon" I used in the 2D math form has been translated into weird variable names for weird reasons...

Hi all, What if the in the above code generation, "g*ee" and "g*v" are indeed vector dot products? That's to say, g, ee, v are vectors of same length? How to do the C code generation in this case?

When you do C code generation, you will get Error, (in codegen/C) unable to translate i there is no way to get rid of it?

Thanks! This demonstrates Maple's capability. I didn't expect this trick before at all. What a great learning experience. I just looked at my expression, in fact I have two Gamma functions in fractional form, which is Gamma(a+b*I)/Gamma(a+b*I+n+1), and "a" and "b" are any real numbers. Any more simplifications that can be done? Thanks!