Loading...
Searching...
No Matches
Collaboration diagram for ConcatKernels:
Classes | |
| class | ConcatScalar< TT, LCNT, H, INP_W, OUT_W > |
| Scalar implementation, ConcatScalar<f,5,4,32,144> takes 5858 cycles (~850 for output window) More... | |
| class | ConcatFloat< TT, LCNT, H, INP_W, OUT_W > |
| Vector implementation, requires INP_W%4=0, OUT_W%4=0. ConcatFloat<f,5,4,32,144> takes 715 cycles (~300 for output window). More... | |
| class | ConcatInt8< TT, LCNT, H, INP_W, OUT_W > |
| Vector implementation for int8_t, requires INP_W%16=0, OUT_W%16=0, ConcatInt8<f,5,4,32,144> takes 283 cycles (~same with output window). More... | |
| class | ConcatFloatStream< TT, H, INP_W1, INP_W2, OUT_W > |
| Scalar implementation for stream concat, ConcatFloatStream<f,4,32,32,64> takes ~1000 cycles. More... | |
| class | ConcatFloatStreamWithStall< TT, H, INP_W1, INP_W2, OUT_W > |
| Scalar implementation for stream concat,. More... | |
| class | ConcatFloatStreamSequentially< TT > |
| Scalar implementation for stream concat, ConcatFloatStreamSequentially<f,4,32,32,64> takes ~1000 cycles. More... | |
| class | ConcatFloatPktStream< TT, LCNT, H, INP_W, OUT_W > |
| Scalar implementation for stream concat, ConcatFloatPktStream<f,4,32,32,64> takes cycles. More... | |
| class | ConcatInt8Stream< TT, H, INP_W1, INP_W2, OUT_W > |
| Vector implementation for stream concat with int8, ConcatInt8Stream<f,4,32,32,64> takes cycles. More... | |
| class | ConcatTwo32bitStreams< TT, LCNT, H, INP_W, OUT_W > |
| Scalar implementation for concatenating 2 chunked streams, ConcatTwo32bitStreams<f,4,32,32,64> takes ~1000 cycles. More... | |
