Re: Incredible slow Plot
- To: mathgroup at smc.vnet.net
- Subject: [mg120185] Re: Incredible slow Plot
- From: DrMajorBob <btreat1 at austin.rr.com>
- Date: Tue, 12 Jul 2011 06:58:29 -0400 (EDT)
- References: <iv9etb$dg1$1@smc.vnet.net> <201107111058.GAA08408@smc.vnet.net>
- Reply-to: drmajorbob at yahoo.com
You see, that's very different.
In this code, Evaluate has no effect whatsoever:
Plot[{Evaluate[f1[t]/.sol], Evaluate[f2[t]/.sol]},{t,0,1200}]
because the first argument of Plot is a List, and you're not applying
Evaluate to it. The List is held, not the list elements separately.
In this code, on the other hand:
Plot[Evaluate[f[t]/.sol],{t,0,1200}]
you ARE applying Evaluate to the first argument.
I don't think that explains why the plot is slow, but we still haven't
seen the actual code.
Send me the notebook, if you like.
Bobby
On Mon, 11 Jul 2011 15:25:44 -0500, Iván Lazaro <gaminster at gmail.com>
wrote:
> Well, i'm going to try to clarify it a little, but, as I said, is not
> posible to paste the complete code; the equations are just too big. I
> also made a mistake writing the last email.
>
> So, for example,
>
> {eqns, cond}={f1'[t]==a11*f1[t]+a12*f2[t]+...+a1N*fN[t],...,
> fN'[t]==aN1*f1[t]+aN2*f2[t]+...+aNN*fN[t], f1[0]==t01,...,fN[0]==t0N},
>
> and
>
> f={f1,f2,...,fN}.
>
> If Something is, say 1, then
>
> sol[[1, 1]]:
> f1[t]->InterpolatingFunction[{{0.`,1200.`}},"<>"][t]
>
> and sol[[1, 1, 2]] is just
> InterpolatingFunction[{{0.`,1200.`}},"<>"][t].
>
>
> So, with
>
> sol=NDSolve[{eqns, cond},f,{t,0,1200}][[1]];
> a[t_]=sol[[1, 1, 2]]
> b[t_]=sol[[1, 2, 2]]
>
> i'm just extracting the solution for two of my variables, f1 and f2.
>
> Plot "a" and "b",
>
>
> Plot[{a[t], b[t]},{t,0,1200}],
>
> was fast; however this:
>
> Plot[{Evaluate[f1[t]/.sol], Evaluate[f2[t]/.sol]},{t,0,1200}],
>
> was, somehow, imposible.
>
> 2011/7/11 DrMajorBob <btreat1 at austin.rr.com>:
>>> and that was it. However I don't understand this. Was the problem the
>>> "size" and "amount" of interpolated functions?
>>
>> I don't understand it either. The two methods seem equivalent, but this
>> code
>>
>>> sol=NDSolve[{eqns, cond},f,{t,0,1200}][[1]];
>>> a=sol[[1, Something, 2]]
>>> b=sol[[1, Something+1, 2]]
>>
>> suggests that you're solving for one function f in the first line, and
>> YET,
>> you're extracting two solutions a and b in the next two lines. That's
>> not
>> possible, so you're not showing us the code you actually used. (We know
>> that
>> anyway, since "eqns", "cond", and "Something" are undefined.)
>>
>> I suspect in the real code, the two methods that seem equivalent are NOT
>> equivalent at all.
>>
>> Bobby
>>
>> On Mon, 11 Jul 2011 05:58:03 -0500, Iván Lazaro <gaminster at gmail.com>
>> wrote:
>>
>>> Hi!
>>>
>>> Yes, I tried
>>>
>>> sol=NDSolve[{eqns, cond},f,{t,0,1200}][[1]];
>>> Plot[Evaluate[f[t]/.sol],{t,0,1200}],
>>>
>>> but that was a pain. Thanks to Bobby I managed to solve my speed
>>> problem:
>>>
>>> Instead of
>>>
>>> sol=NDSolve[{eqns, cond},f,{t,0,1200}][[1]];
>>> Plot[Evaluate[f[t]/.sol],{t,0,1200}],
>>>
>>> I selected the specific solutions I needed, and Set them to a variable
>>> that then I plot:
>>>
>>>
>>>
>>> sol=NDSolve[{eqns, cond},f,{t,0,1200}][[1]];
>>> a=sol[[1, Something, 2]]
>>> b=sol[[1, Something+1, 2]]
>>>
>>> Plot[{a[t],b[t]}],{t,0,1200}],
>>>
>>> and that was it. However I don't understand this. Was the problem the
>>> "size" and "amount" of interpolated functions?
>>>
>>
>>
>> --
>> DrMajorBob at yahoo.com
>>
--
DrMajorBob at yahoo.com
- References:
- Re: Incredible slow Plot
- From: Iván Lazaro <gaminster@gmail.com>
- Re: Incredible slow Plot