What samples have you used? This will determine how you analyse your data. For instance, do you have a large cohort of tumour samples for which you want to find common CNVs? Or do you have patient matched tumour-normal sample pairs that you'd like to find somatic CNVs in?

In either case, you are probably going to find some nice R programs that will help you. For instance, DNAcopy is a very good package that will perform copy number and segmentation analysis for a variety of sample types.

I have case control schizophrenia samples and am looking for rare CNVs comparing cases and controls. The samples were genotype with PsychChip Illumina microarray which has about 500K snps. 

Muhammad

If you are using iScan Illumina platform then Genome Studio and Blue fuse are best solutions for CNV analysis.. We use this for HumanCytoSNP 300K array genotyping. 

Thank you Musharraf Jelani. 

dchip software

Altaf, dchip is very hard software for users and for reliable results to be obtained. Best, of my opinion, is GoldenHelix SVS but it is not free as dchip. 

PennCNV y QuantiSNP are free and good software.

PennCNV y QuantiSNP is really good for data from Illumina SNP chip. I have used for omniexpress chip.