feat(ptodsl): Add tile debug print support

[Crystal-wzy]

Add PTODSL support for the native pto.tprint debug-print operation and
expose it through pto.tile.print. The operation is modeled as a device-side
side effect, emits no numeric result, and is routed through the native EmitC
TPRINT lowering path.

Summary

• Add _ops.tprint(src, *, tmp=None) and export it through pto.tile.print
  for tile debug printing, including the optional scratch tmp operand
• Reject pto.tile.print on the VPTO backend with an explicit unsupported
  diagnostic, while allowing EmitC modules to keep the native pto.tprint
  operation
• Treat TPrintOp as a native operation in tile expansion and template
  metadata insertion, so it is not expanded through TileLib templates
• Add JIT compile coverage for EmitC pto.tile.print lowering and the VPTO
  unsupported-backend diagnostic
• Add an A5 TileLib ST case for pto.tile.print using an EmitC kernel and a
  GM -> tile -> print -> GM round trip
• Add a VPTO lit negative case that checks the explicit unsupported diagnostic
• Document pto.tile.print in the PTODSL user guide, including side-effect
  behavior, tmp handling, dtype coverage, backend support, and A5 constraints

Testing

• ptodsl/tests/test_jit_compile.py adds EmitC lowering and VPTO
  unsupported-backend checks for pto.tile.print
• test/lit/vpto/tprint_vpto_unsupported.pto covers the VPTO unsupported
  diagnostic for pto.tprint
• test/tilelib-st/a5/tprint/case.py covers the A5 EmitC
  pto.tile.print path with an f32 round trip

[Zhendong404]

tprint只能在emitc后端用吗？vpto后端的实现方案是什么？

[Crystal-wzy]

目前 tprint 只支持 EmitC。EmitC 已有 PTOPrintToTPRINT lowering，能直接生成 pto-isa TPRINT 调用；VPTO 侧目前没有 TPrintOp/TPRINT 的 emission/runtime debug print 实现。因此这版不做 TileLib 模板展开，并在 VPTO backend下遇到 pto.tprint 时显式报 unsupported，避免静默走错路径。后续 VPTO 支持需要先补 VPTO lowering/runtime 的 print side-effect 方案。