c9eab3c7e0
diag: rewrite multi-tile test with explicit per-segment compile and reference
2026-05-26 15:39:39 +00:00
7f983fb855
diag: add direct segment 0 test to compare with run_segment
2026-05-26 15:37:06 +00:00
2a5f9dc6e3
fix: simplify run_segment to use full hd V tensor (was incorrectly splitting by pv_n_tile)
2026-05-26 15:34:57 +00:00
aa2df1a202
diag: test n_comp=96 with sink bias directly
2026-05-26 15:33:38 +00:00
25b236fe00
diag: test D5c multi-tile with no sink bias to isolate issue
2026-05-26 15:31:38 +00:00
a3989929de
diag: per-segment reference comparison for D5c multi-tile
2026-05-26 15:29:52 +00:00
bbc29945e8
diag: add per-segment debug output for D5c multi-tile
2026-05-26 15:28:19 +00:00
e64392f1ac
D5c: add apply_sink_bias flag (independent of n_comp)
...
For all-SWA segments (n_comp=0), sink bias still needs to be applied
to all positions. The apply_sink_bias flag controls compilation of
the sink bias code path, independent of n_comp offset.
2026-05-26 15:26:52 +00:00
60b6f8ee79
fix: use separate kernels for segments with/without compressed KV (n_comp is compile-time)
2026-05-26 15:23:21 +00:00
2efd15c852
fix: correct swa_len_local calculation per segment for D5c multi-tile
2026-05-26 15:22:03 +00:00
3abcc7ff09
D5c: multi-tile test using Python KV merge with sink bias
2026-05-26 15:20:45 +00:00
