Implementing and Evaluating an Heterogeneous, Scalable, Tridiagonal Linear System Solver with OpenCL to Target FPGAs, GPUs, and CPUs

Table 2

FPGA hardware utilisation for the OpenCL static portion, TDMA, and truncated SPIKE kernels, and total utilisation for each implementation (static  kernel).

Resource	OpenCL static (%)	TDMA		Truncated SPIKE
		Kernel (%)	Total (%)	Kernel (%)	Total (%)
ALUTs	13	14	27	18	31
FFs	13	10	23	13	23
BRAMs	16	52	68	49	65
DSPs	10	16	26	27	37