基于GPU的数字下变频累积误差控制方法

GPU-based digital down-conversion cumulative error control method

针对数字下变频系统中高速数据处理实时性需求,设计了基于中央处理器(central processing unit,CPU)+图形处理单元(graphic processing unit,GPU)的异构平台数字下变频处理算法,并完成了相应并行程序设计,进行了实际数据验证。鉴于GPU运算采用单精度浮点数格式进行数据处理,进行长时程数据处理过程中,会产生浮点数舍入误差累积,导致结果的不准确不可靠,设计了联合相位循环归零法和无误差变换补偿法,实现了浮点数舍入误差的控制与补偿。数据测试结果表明,所提方法可将相位误差控制在10-11rad以内,且不随时间累积,计算结果准确可靠。

Aiming at the real-time demand of high-speed data processing in digital down-conversion system,a digital down-conversion processing algorithm based on central processing unit(CPU)+graphics processing unit(GPU)is designed for heterogeneous platform,and the corresponding parallel programming is completed,and the actual data is verified.Since GPU computing using single precision floating point number format on data processing,for long time history in the process of data processing,can produce floating-point rounding error accumulation,not accurate and reliable results.A combined phase cycle zero and error-free transformation compensation method is designed,and the floating-point number rounding error control and compensation is realized.Data test results show that the proposed method can control the error within 10-11,and does not cumulative with time,and the calculation results are accurate and reliable.