perf: tune bf16 gemv and add opt-in fused swiglu

- deepen common-shape BF16 GEMV tuning with warp-row tiling for LLaMA/Qwen2/GPT-NeoX/OPT decode projections
- add fused BF16 up/gate SwiGLU CUDA primitive with ASTRAI_SWIGLU=0/1/auto dispatch
- keep the unfused linear backend as the default path; auto enables no shape until per-architecture checkpoint gates pass
- fall back to the linear/torch chain when kernels are absent, on CPU, in training, or outside supported M/K/dtype shapes
- add gemv/swiglu benchmark scripts, dispatch and parity tests, and kernel documentation

Benchmark: NVIDIA L20 (sm_89), CUDA 12.8, PyTorch 2.11.0+cu128, idle GPU. AstrAI 1B config (24 layers, hidden 1536, vocab 100000), BF16, prompt 128, 32 greedy decode tokens, CUDA graphs enabled, A/B in separate interleaved processes (3 rounds, 8 trials each, medians). Default vs ASTRAI_SWIGLU=1 per generate call: batch 1 134.8->129.1 ms (+4.44%), batch 2 136.2->130.9 ms (+4.06%), batch 4 145.5->140.3 ms (+3.66%). Greedy output identical at batch 1, differs at batch 2/4, so auto stays unfused by default; kernelless fallback verified bit-identical greedy.

csrc/CMakeLists.txt

@@ -62,6 +62,7 @@ set(KERNEL_NAMES
     attn_paged_decode
     attn_paged_prefill
     bf16_gemv
+    bf16_swiglu
     rotary_emb
 )
 set(KERNEL_SRCS
@@ -70,6 +71,7 @@ set(KERNEL_SRCS
     attention/paged_decode.cu
     attention/paged_prefill.cu
     gemv/bf16_gemv.cu
+    gemv/bf16_swiglu.cu
     rotary_emb.cu
 )