Re: Shuffling 10^8 numbers: memory problems
- To: mathgroup at smc.vnet.net
- Subject: [mg53300] Re: Shuffling 10^8 numbers: memory problems
- From: George Szpiro <george at netvision.net.il>
- Date: Wed, 5 Jan 2005 01:21:14 -0500 (EST)
- References: <crdm08$ra5$1@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
I realized that the problem probably starts when I write the results to the
file. My computer has sufficient RAM to read the whole input file. I tried
writing to the same file that I read from, but that did not solve the
problem.
Question: Is there a way to read the data that I am done with, i.e.,
LL[[j]] , to the output file and then delete it from memory? Since shuffling
proceeds sequentially in the program below, there may be a way to do this
every time 10,000 entries have been shuffled: write them to the output file,
and free up memory. How?
Thanks and best regards,
George
----- Original Message -----
From: "George Szpiro" <george at netvision.net.il>
To: mathgroup at smc.vnet.net
Subject: [mg53300] Shuffling 10^8 numbers: memory problems
>
> The problem of quadratic time seems to have been solved by using
>
> aa = Developer`ToPackedArray[aa];
>
> after reading in the data.
>
> But now the Mathematica kernel hits the limit again at 2*10^7 data: it
> shuts down
> MemoryInUse[] shows about 82 MB were used, but my machine has 704MB RAM
> installed! What is happening?
>
> Is there a way to work around that one? I need to shuffle five times as
> many integer numbers, i.e., I would need about 420 MB memory.
>
> Would it maybe help to start my machine in Safe Mode, so that no other
> processes run, and thus get the use of nearly the full 704 MB memory?
> (By the way, the 10^8 numbers are integers that are not larger than
> about 1000 or 5000.)
>
> Thanks,
> George
>
> The program I used:
>
> Clear[SSS,aa];
> SSS=20000000; (*I need to use 100000000*)
>
> aa=ReadList["c:GG.doc",Number,SSS];
> aa = Developer`ToPackedArray[aa];
>
> SetAttributes[shuffleSet, HoldFirst];
> shuffleSet[LL_] := Module[{n=SSS, rand},
> Do[
> rand = Random[Integer, {j,n}];
> LL[[{j,rand}]] = LL[[{rand,j}]],
> {j,n}];
> ]
>
> Timing[shuffleSet[aa]]
>
> <<Utilities`CleanSlate`
>
> OpenWrite["c:ShuffledGG.doc"];
> Put[OutputForm[aa],"c:ShuffledGG.doc"]
> Close["c:ShuffledGG.doc"];
>
>
>