fix(CRITICAL): remove xformers patches — Sub168 proves base ixformer attention works at 11.9 TPS, our patches reduced to 2.6 TPS

Root cause of Sub 520 output_tps=2.6 (vs Sub 168 output_tps=11.9):
- patch_xformers_sdpa_seq.py replaces ixformer flash attention with
  pure PyTorch O(L^2) matmul+softmax serial implementation
- 32 full attention layers x every token = 4.6x slower

Sub 168 (base image) proof:
- output_tps_avg=11.9, output_tps_p50=13.0, output_tps_p90=18.1
- XFormers backend used WITHOUT any patches
- ixformer flash_attn works correctly on BI-V100

This commit: skip xformers patches in patch_ops.sh
Expected: output_tps should recover to ~11.9 (Sub 168 level)

qwen3_6_scripts/patch_ops.sh

@@ -164,12 +164,13 @@ if [ -d "$_SITE" ]; then
 fi
 
 # ===========================================================
-# 5. XFormers patches — head_dim=256 bypass for BI-V100
-#    Comp 168 also had xformers patches (base uses xformers for attention)
+# 5. XFormers patches — DISABLED
+#    Sub 168 (base image) achieved output_tps=11.9 WITHOUT any xformers patches.
+#    Sub 520 applied these patches → output_tps=2.6 (4.6x slower!)
+#    The patches replace ixformer flash attention with pure PyTorch O(L²) matmul.
+#    Base image ixformer attention works correctly — proven by Sub 168.
 # ===========================================================
-python3 ./patch_xformers_sdpa_seq.py 2>&1 || true
-python3 ./patch_xformers_sdpa_batch.py 2>&1 || true
-echo "[patch_ops] xformers patches applied"
+echo "[patch_ops] xformers patches SKIPPED — base ixformer attention works (Sub 168 proof)"
 
 # ===========================================================
 # 6. model_runner patch (prefix_cache_hit fix)