问题在于乘法指令的结果可能会有其操作数的两倍之长。
The problem is that the result of a multiply instruction can be twice as long as its operands.
由于乘法指令的操作差异很大，因此SPUintrinsic并不会像对其他操作一样将它们连接在一起。
Because the multiplication instructions operate so differently, the SPU intrinsics do not coalesce them as much they do for other operations.
为了解决这一问题，乘法指令只使用每个32位的最低有效的16 位以便结果能够放进整个 32 位寄存器内。
To combat the problem, multiplication instructions only use the least-significant 16 bits of every 32-bit value so that the result will fit in the full 32-bit register.
现在，SPU上的整数相乘指令多少有点问题，特别是有符号乘法(使用mpy指令)。
Now, the integer multiply instructions on the SPU are somewhat problematic, especially signed multiplication (using the mpy instruction).
文章以某数值协处理器的设计为例，研究了微指令格式的确定以及微程序代码的编写，并给出了编制的乘法微代码的实例。
Combined with the design of a math coprocessor, the paper dis-cusses how to determine microinstruction format and write microgram code. Multiplication code is presented as an example.
文章以某数值协处理器的设计为例，研究了微指令格式的确定以及微程序代码的编写，并给出了编制的乘法微代码的实例。
Combined with the design of a math coprocessor, the paper dis-cusses how to determine microinstruction format and write microgram code. Multiplication code is presented as an example.
应用推荐