Mojo function

generic_fused_qk_rope_bshd_continuous_batch

generic_fused_qk_rope_bshd_continuous_batch[dtype: DType, //, *, interleaved: Bool, target: StringSlice[StaticConstantOrigin]](q_proj: LayoutTensor[dtype, layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], kv_collection: ContinuousBatchingKVCacheCollection[dtype_, kv_params_], freqs_cis: LayoutTensor[dtype, layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], layer_idx: UInt32, output: LayoutTensor[dtype, layout, origin, address_space=address_space, element_layout=element_layout, layout_int_type=layout_int_type, linear_idx_type=linear_idx_type, masked=masked, alignment=alignment], context: DeviceContextPtr = DeviceContextPtr())

Performs a fused RoPE projection for Q and K projections.

We have a manually fused QKV projection with mo.opaque dtypes in our Llama model. Due to a limitation in custom op definitions, we can't declare both a tensor and opaque dtype as output from a custom kernel. This requires us to only note Q_proj as an output from the QKV projection. If we immediately follow the QKV proj kernel with a RoPE kernel applied to K, we'll get a race condition because the graph compiler doesn't know about the dependency between these kernels in the graph definition. Here we fuse the RoPE kernel applied to Q_proj with K_proj, so K_proj RoPE is only executed after QKV completes.