Hi,My design has multiple kernels communicating through channels forming a chain like this: Kga -> Kconf -> Kene -> Kst It is running correctly in emulation, but some information in the optimization report seems strange to me: Specifically, one kernel (Kga) is compiled as NDRange, although no work-item function is used in any kernel. Is there any function/statement that is causing this kernel to be compiled as NDRange?
Try adding __attribute__((max_global_work_dim(0))) for that kernel, this should ensure it is compiled as a single work-item kernel.source: aocl programming guide, 126.96.36.199 (https://www.altera.com/en_us/pdfs/literature/hb/opencl-sdk/aocl_programming_guide.pdf#page=136)
Thanks for the info.I added that attribute, but unfortunately that kernel is still being compiled as ND-Range. The log shows warnings: "Compiler Warning: Channel ordering barrier insertion in X of Kernel Kga, but kernels may hang as a result" I don´t know how to interpret this warning. Have you experienced something like this before?