Regenerate latency plots/diagrams for post-Phase-2c model

Allreduce + pe2pe + ipcq + pe_view auto-regenerated by test sweeps
running against the new chunk-streaming wire timing (per-flit
wormhole) — absolute numbers shift upward to reflect bottleneck-link
transit charged once per flit (instead of the previous cut-through
subtraction at HBM CTRL).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

docs/diagrams/allreduce_latency_plots/buffer_kind_sweep.csv

@@ -1,13 +1,13 @@
-buffer_kind,sip_topology,n_sips,n_elem,bytes_per_pe,latency_ns
-hbm,torus_2d,6,128,256,1858.0399999999827
-hbm,torus_2d,6,1024,2048,2389.0399999999827
-hbm,torus_2d,6,8192,16384,6673.039999999986
-hbm,torus_2d,6,32768,65536,21361.03999999992
-sram,torus_2d,6,128,256,1774.0399999999827
-sram,torus_2d,6,1024,2048,2389.0399999999827
-sram,torus_2d,6,8192,16384,7345.039999999986
-sram,torus_2d,6,32768,65536,24337.039999999935
-tcm,torus_2d,6,128,256,1678.0399999999827
-tcm,torus_2d,6,1024,2048,1957.0399999999827
-tcm,torus_2d,6,8192,16384,4225.039999999986
-tcm,torus_2d,6,32768,65536,12001.03999999992
+buffer_kind,sip_topology,n_sips,n_elem,bytes_per_pe,latency_ns
+hbm,torus_2d,6,128,256,2144.0399999999754
+hbm,torus_2d,6,1024,2048,2908.74499999995
+hbm,torus_2d,6,8192,16384,8851.185000000081
+hbm,torus_2d,6,32768,65536,29225.265000008752
+sram,torus_2d,6,128,256,2060.0399999999754
+sram,torus_2d,6,1024,2048,2908.74499999995
+sram,torus_2d,6,8192,16384,9523.185000000081
+sram,torus_2d,6,32768,65536,32201.265000008752
+tcm,torus_2d,6,128,256,1964.0399999999754
+tcm,torus_2d,6,1024,2048,2476.74499999995
+tcm,torus_2d,6,8192,16384,6403.185000000081
+tcm,torus_2d,6,32768,65536,19865.265000008738