Jason
a3e1996833
FFT engine: merge SHIFT into WRITE (5→4 cycle butterfly, 20% throughput) + barrel-shift twiddle index
...
Opt 1: Eliminated ST_BF_SHIFT state — arithmetic right-shift is pure
bit-selection (zero logic levels), merged into BF_WRITE combinational
add/subtract. Saves LOG2N * N/2 = 5120 cycles per 1024-pt FFT.
Opt 2: Replaced idx_val * tw_stride_reg general multiply with
idx_val << (LOG2N-1-stage) barrel shift. tw_stride_reg is always a
power of 2, so this is mathematically identical and frees a multiplier.
Regression: 18/18 FPGA pass (bit-exact results).
2026-03-20 00:20:59 +02:00
..
2026-03-19 22:59:46 +02:00
2026-03-19 23:54:48 +02:00
2026-03-19 22:59:46 +02:00
2026-03-19 23:54:48 +02:00
2026-03-19 11:31:46 +02:00
2026-03-19 22:59:46 +02:00
2026-03-19 22:59:46 +02:00
2026-03-19 22:59:46 +02:00
2026-03-17 19:38:09 +02:00
2026-03-19 21:34:02 +02:00
2026-03-19 22:59:46 +02:00
2026-03-16 22:24:22 +02:00
2026-03-17 19:38:09 +02:00
2026-03-16 18:14:06 +02:00
2026-03-19 10:31:16 +02:00
2026-03-17 13:48:47 +02:00
2026-03-20 00:20:59 +02:00
2026-03-16 10:25:07 +02:00
2026-03-16 10:25:07 +02:00
2026-03-19 20:39:01 +02:00
2026-03-17 20:11:13 +02:00
2026-03-18 01:28:42 +02:00
2026-03-19 10:31:16 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-19 20:53:29 +02:00
2026-03-16 10:25:07 +02:00
2026-03-17 23:51:04 +02:00
2026-03-19 10:31:16 +02:00
2026-03-19 23:54:48 +02:00
2026-03-19 23:54:48 +02:00
2026-03-18 03:57:26 +02:00
2026-03-19 16:47:59 +02:00
2026-03-19 23:54:48 +02:00
2026-03-17 19:38:09 +02:00
2026-03-17 15:41:06 +02:00
2026-03-19 21:19:07 +02:00
2026-03-16 19:53:40 +02:00
2026-03-16 19:53:40 +02:00
2026-03-19 23:54:48 +02:00
2026-03-16 10:25:07 +02:00