Filter math accelerator (FMAC) RM0440
490/2126 RM0440 Rev 4
Figure 43. Circular input buffer
The X2 buffer can only be used in vector mode (that is not circular), and needs to be pre-
loaded, except if the contents of the buffer do not change from one operation to the next. For
filter functions, the X2 buffer is used to store the filter coefficients.
When operating as a circular buffer, the space allocated to the buffer (x1_buf_size) should
generally be bigger than the number of elements in use for the current calculation, so that
there are always new values available in the buffer. Figure 43 illustrates the layout of the
buffer for a filter operation. While calculating an output sample y[n], the unit uses a set of
N+1 input samples, x[n-N] to x[n]. When this is finished, the unit starts the calculation of
y[n+1], using the set of input samples x[n-N+1] to x[n+1]. The least-recent input sample, x[n-
N], drops out of the input set, and a new sample, x[n+1], is added to it.
The processor, or DMA controller, must ensure that the new sample x[n+1] is available in
the buffer space when required. If not, the buffer is flagged as empty, which stalls the
execution of the unit until a new sample is added. No underflow condition is signaled on the
X1 buffer.
x[n]
x[n+1]
Write pointer
x[n-1]
x[n-2]
x[n-3]
x[n-4]
x[n-5]
x[n-6]
x[n-N]
x[n+2]
x[n+3]
x[n+4]
These values in use for
calculating y[n]
Next values already
loaded
Available buffer space
Available buffer space
x1_base
x1_buf_size
MSv45870V1
To Next Page
Loading...
Need help?
General
