The paper presents the parallelizing programming environment CoDe-X introducing hardware/software codesign strategies on two levels of partitioning for datadriven Xputer-based accelerators. CoDe-X performs both, in the first level a profiling-driven host/accelerator partitioning for performance optimization, and in a second level a resource-driven sequential/structural partitioning of the accelerator source code to optimize the utilization of its reconfigurable resources. CoDe-X accepts a C dialect also including optional data-procedural language features, which can be included to achieve highest possible acceleration factors provided by the Xputer hardware.