[codex] Lower grouped integer widening to unpack

[Zhendong404]

Summary

• Lower grouped integer widening for gs(2)/gs(4)/gs(8), lane_stride=1, to vsunpack/vzunpack.
• Chain two unpack operations for i8 to i32.
• Add VMI lit coverage for signed/unsigned group layouts.

Validation

• cmake --build build -j8 --target pto-test-opt
• Focused VMI/FileCheck regression checks passed.

Draft note: TileLang codegen probing is being validated separately; the current generated PTODSL still shows a direct BF16-to-F32 vmi.vcvt in the end-to-end per-token case.

[mouliangyu]

P1: test/lit/vmi_new/vmi_to_vpto_group_slot_integer_unpack.pto:69-70 的 CHECK 仍匹配 !pto.vreg<256xi8>/!pto.vreg<128xi16>，但测试输入已改为 signed 类型，实际输出是 !pto.vreg<256xsi8> -> !pto.vreg<128xsi16>。因此 build-and-test 仍失败（1808 个测试中 1 个失败）。请将这些 CHECK 更新为 si8/si16 后重新运行 lit。

[mouliangyu]

已确认上次提出的 signed integer FileCheck 问题修复：检查现已使用 si8/si16，与实际 vsunpack 输出一致。重新检视完整变更集后未发现新的阻塞问题。当前 vpto-sim-validation 失败发生在 git ls-remote 拉取 LLVM 时（Failure when receiving data from the peer），属于基础设施网络错误，建议重跑该 job；build-and-test 仍在运行。

[reedhecre]

A3 板测失败

• 触发方式：merged
• 源码提交：b6960c4e07a0
• 结果汇总：OK 313 / FAIL 2 / SKIP 30
• 日志：/home/zhongxuan/ptoas-board-monitor/runtime/logs/20260818_024910_merged_pr1250.log
• 结果 TSV：/home/zhongxuan/ptoas-board-monitor/runtime/logs/20260818_024910_merged_pr1250.tsv
• 失败阶段：board-validation / exit=1

失败用例

• Ci/ci (run, exit=2)
• Qwen3_14BPrefillA3/out_proj_aiv (run, exit=2)