Creating a 16DW TLP from root complex to pci express endpoin

I am trying to create a device driver for a pci-express end-point card which has some ddr memroy. When ever I write a block of data to the endpoint from my device driver , I see TLP's with 1 DW being sent to my endpoint card. How do I get the processor on my motherboard to generate a single TLP with 16DW or more data in it.