Question: Large system of ODEs

Hi! I have to solve a very large system of ODEs for a set of functions that can be indexed with two integers, n and l, call them f[n,l](x), where n ranges from 1 to an order 1000 number and l ranges from 1 to an order 10 number. 

What is the best strategy in dsolve, keeping in mind I will only need to have quick access to the values of the first few such functions, say f[0,0](x) and f[0,1](x) for any given x in a given range. I dont need to store the rest, and trying to store all of them actually will give me a warning about length of output exceeds limit of 1000000. So, having maple output the result of dsolve in a procedural form seems to be what I want. However in doing so, for evey value of x i call the solving procedure maple will solve the system each time, thus taking a long time. 

My dilema is: to save storage space and memory I want to use dsolve output as procedures. To save time, when accessing the functions I need (only the first few in the large set of unknown functions that obey the ODE) for a given value of x, I would like maple to have those already stored instead of computing them with each invocation. So, what I need is a somehow split behavior of output. For some of my unknown functions to act as if i use range in dsolve, so that I have access at the values without further computation, and for the rest of the solutions (most of the unknown functions) just keep them in procedural form. How do I achieve this? 

Below is a schematic of my probem:

ODEs:={Set of about 10000 coupled ODEs for functions labeled f[n,l] n ranging from 1..1000 and l from 1..10}

dsolve(ODEs,numeric, output=???)#What options to put in dsolve such that I have quick access to f[0,0](x) and f[0,1](x) in a given range (say xmin=0.1 and xmax=10 for example) but not store the rest.