[MUH] Complete all 26 CCCL algorithm tuning headers — full parity with cub/device/dispatch/tuning/

Added 20 missing tuning headers (was 6, now 26):
  P1: radix_sort, reduce_by_key, scan_by_key, select_if, histogram,
      merge, merge_sort, unique_by_key, batched_topk, transform_tile
  P2: segmented_reduce, segmented_scan, segmented_sort,
      segmented_radix_sort, three_way_partition, rle_encode,
      rle_non_trivial_runs
  P3: adjacent_difference, find, find_bound_sorted_values

Updated muh.cuh to include all 26 headers (v0.2.0).
All headers compile clean (g++ -std=c++17), compile_test passes 17/17.
gen_patch.py reads bi100_* structs from all 26 files.

Coverage: muh now has a tuning header for every CCCL tuning_*.cuh file.

muh/include/muh/tuning/tuning_batched_topk.cuh

@@ -0,0 +1,38 @@
+// muh/include/muh/tuning/tuning_batched_topk.cuh — BI-V100 batched_topk tuning
+//
+// Mirrors: cccl_upstream/cub/cub/device/dispatch/tuning/tuning_batched_topk.cuh
+// vllm impact: batched top-k across sequences
+// Competition weight: Output TPS × 16.796
+
+#pragma once
+
+#include "muh/hardware.cuh"
+#include "muh/tuning/common.cuh"
+
+namespace muh::tuning::batched_topk {
+
+struct BatchedTopkPolicy {
+  int threads_per_block;
+  int items_per_thread;
+  BlockLoadAlgorithm load_algorithm;
+};
+
+struct bi100_default {
+  static constexpr int threads = 256;
+  static constexpr int items = 16;
+  static constexpr int load_algo = BLOCK_LOAD_WARP_TRANSPOSE;
+};
+
+struct policy_selector {
+  int key_size;
+
+  constexpr BatchedTopkPolicy operator()(const hardware_capability& hw) const {
+    if (hw.at_least(hardware_capability::vendor_t::iluvatar, 100)) {
+      return {bi100_default::threads, bi100_default::items, BLOCK_LOAD_WARP_TRANSPOSE};
+    }
+    // Fallback
+    return {bi100_default::threads, bi100_default::items, BLOCK_LOAD_WARP_TRANSPOSE};
+  }
+};
+
+} // namespace muh::tuning::batched_topk