[Chi Lo (chilo-ms)]

Description

Add opt-in fixed-placement memory diagnostics with session.enable_memory_backing_report=1 (default off).

• Capture initializer payloads before kernel construction and after local PrePack, before recursive subgraph finalization.
• Model local-graph tensor and declared L2 workspace lifetimes using allocation-plan alias roots and release actions. Sequential workspaces do not overlap; parallel/multi-stream execution conservatively overlaps modeled roots.
• Keep L1 prepack/initialization-scratch estimates separate from Run payloads. Reuse finalization's L2 declarations without querying kernels again.
• Expose allocator-wide current observations separately from modeled payloads. Unsupported zero-filled statistics remain unavailable, not proven zero.
• Log at initialization and after ordinary execution. Keep physical initialization/Run peaks explicitly unavailable and list uncovered terms.

This is the first diagnostic adapter, not complete session backing accounting or a no-OOM guarantee. It changes no allocation, placement, contract-selection, or budget policy and adds no device allocation or explicit synchronization. Control-flow/subgraph reports are not composed into a session peak.

Validation

Windows CPU Release, Ninja/MSVC 2022, using the existing isolated incremental build with the Windows memory leak checker disabled.

From the configured CPU Release build directory:

cmake --build . --target onnxruntime_test_all --parallel 10

Passed.

& '.\onnxruntime_test_all.exe' '--gtest_filter=SessionMemoryReportTest.*:SessionStateTest.*Workspace*:OpKernelTest.DefaultDeclareWorkspaceRequirementsClearsOutput:InferenceSessionTests.NoTimeout' --gtest_color=no

Passed: 22 tests, no skips. Eight new cases cover sequential/parallel scratch, input/initializer aliases, source release and packed destinations, dynamic shapes, known-zero versus unavailable workspace, reporting disabled, and a real report-enabled CPU Run with and without an arena. The runtime case verifies allocator-retained bytes are not added to the modeled payload peak and unsupported allocator statistics stay unavailable.

From the source worktree:

lintrunner -r origin/chilo/gqa-level2-envelope --take CLANGFORMAT --output oneline
git diff --check

Both passed. No new CUDA, training-enabled, minimal-build, or cross-stream runtime validation was performed for this slice.

Motivation and Context

Implements the first fixed-placement reporting slice from #32944, stacked on #33240's L1/L2 reconciliation. The stack is #33316 -> #33240 -> this PR.

The report deliberately separates conditional logical lifetime estimates from physical backing observations. Retained pools, alignment/fragmentation, packed ownership/sharing, constructor/preparation scratch, caller I/O/KV capacities, private library/EP allocations, capture, retained outputs, subgraph residency, and concurrency still prevent a complete physical peak.

The remaining sequence is to complete those backing adapters, introduce enforceable pre-allocation session budgets, then add generic contract selection and CPU-region placement. CUDA workspace/activation-pattern integration remains in #32071, WebGPU region integration in #32720, and the placement design in #33372; this PR does not duplicate them.