phase 4: transformer core kernels

CUDA kernels (csrc/):
- common.cuh: shared warp_reduce_sum/max, block_reduce_sum/max
- normalization/rmsnorm.cu: RMSNorm (F32 + BF16)
- normalization/layernorm.cu: LayerNorm with Welford (F32 + BF16)
- activation/activations.cu: GELU tanh-approx + SiLU (F32 + BF16)
- reduce/softmax.cu: safe softmax, 3-pass (F32 + BF16)
- embedding/embedding.cu: gather lookup (F32 + BF16)
- embedding/rope.cu: RoPE in-place + precomputed cos/sin cache (F32 + BF16)

Rust wrappers (xserv-kernels/src/):
- rmsnorm.rs, layernorm.rs, activation.rs, softmax.rs, embedding.rs, rope.rs
- RopeCache struct with GPU-side precomputation

Tests: 12 new tests (ops_test.rs), all passing with good precision:
- F32: max_err 1e-6 ~ 1e-9
- BF16: max_err 2e-3 ~ 7e-3
Total: 29 kernel tests + 27 prior = 56 tests passing

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

crates/xserv-kernels/src/embedding.rs

@@ -0,0 +1,51 @@
+use std::ffi::c_void;
+use xserv_cuda::GpuBuffer;
+use xserv_tensor::{DType, Device, Tensor};
+
+unsafe extern "C" {
+    fn launch_embedding_f32(table: *const c_void, token_ids: *const c_void, out: *mut c_void,
+                            num_tokens: i32, hidden_size: i32, stream: *mut c_void);
+    fn launch_embedding_bf16(table: *const c_void, token_ids: *const c_void, out: *mut c_void,
+                             num_tokens: i32, hidden_size: i32, stream: *mut c_void);
+}
+
+/// Embedding lookup: table[token_ids[i]] for each i.
+/// table: [vocab_size, hidden_size], token_ids: [num_tokens] (i32 on CPU)
+pub fn embedding(table: &Tensor, token_ids: &[u32]) -> Tensor {
+    assert_eq!(table.ndim(), 2);
+    assert!(table.is_contiguous());
+    assert!(matches!(table.device(), Device::Cuda(_)));
+
+    let hidden_size = table.shape()[1];
+    let num_tokens = token_ids.len();
+
+    // Upload token_ids to GPU
+    let ids_bytes = unsafe {
+        std::slice::from_raw_parts(
+            token_ids.as_ptr() as *const u8,
+            num_tokens * std::mem::size_of::<u32>(),
+        )
+    };
+    let mut ids_gpu = GpuBuffer::alloc(ids_bytes.len()).expect("alloc token_ids");
+    ids_gpu.copy_from_host(ids_bytes).unwrap();
+
+    let out = Tensor::zeros(&[num_tokens, hidden_size], table.dtype(), table.device());
+
+    unsafe {
+        match table.dtype() {
+            DType::F32 => launch_embedding_f32(
+                table.data_ptr() as _, ids_gpu.as_ptr() as _,
+                out.data_ptr() as *mut c_void,
+                num_tokens as i32, hidden_size as i32, std::ptr::null_mut(),
+            ),
+            DType::BF16 => launch_embedding_bf16(
+                table.data_ptr() as _, ids_gpu.as_ptr() as _,
+                out.data_ptr() as *mut c_void,
+                num_tokens as i32, hidden_size as i32, std::ptr::null_mut(),
+            ),
+            _ => panic!("unsupported dtype for embedding"),
+        }
+    }
+    xserv_cuda::device::synchronize().unwrap();
+    out
+}