if you have:
for(d1s - d1e)
for(d2s - d2e)
for(d3s - d3e)
...
for(dns - dne)

that means you actually execute the whole thing M=(d1e - d1s)*(d2e - d2s)*(d3e - d3s)*...*(dne - dns) times

So that means you can make a single loop 1..M and now you're left with the problem of extracting from the counter all the n counters you had initially.

If all the intervals you considered initially were equal(let's say the interval was I long) you could just write the new counter in base I and you get the initial indexes,but otherwise...it would probably be a bit more complicated.

You probably should convince yourself if the overhead in additional calculations is worth the effort. Good luck :)