refactor: remove bf16 gemm and swiglu kernels and rebuild csrc benchmarks

- delete csrc/kernels/gemm.cu and swiglu.cu and drop their CMake and setup.py registration
- remove the ops wrappers plus backend/linear.py and backend/swiglu.py so Linear and MLP call F.linear directly
- drop the four gemm and swiglu kernel test files and prune the stale cuda_kernels.md sections
- add csrc/bench benchmarks for the remaining kernels: attention decode prefill paged decode paged prefill versus single-launch SDPA references, rotary versus the torch fallback, fp8 quantize and mm_fp8 versus torch baselines
- attention, rotary_emb, and fp8_ops kernels are unchanged

astrai/model/components/mlp.py

@@ -5,7 +5,6 @@ import torch.nn as nn
 import torch.nn.functional as F
 from torch import Tensor
 
-from astrai.extension.backend.swiglu import swiglu
 from astrai.factory import BaseFactory
 from astrai.model.components.linear import Linear
 
@@ -39,7 +38,7 @@ class MLP(nn.Module):
         self.down = Linear(dim_ffn, dim, init_std=down_init_std)
 
     def forward(self, x: Tensor) -> FFNOutput:
-        gated = swiglu(x, self.up.weight, self.gate.weight)
+        gated = self.up(x) * F.silu(self.gate(x))
         out = self.down(gated)
         return {"hidden_states": out, "aux_loss": None, "router_stats": None}