Analysis Views in Digital Design Studio
Digital Design Studio 2.0.2: Analysis Views in Digital Design Studio. Check the selected host, account permissions and current support before use.
Overview
Analysis is the shared result-inspection area for RTL and ISA work. Select a project and a saved run before interpreting a chart. Missing input or an unsupported host is not a successful run with zero events or zero cost.
The current views use the selected run’s artifacts where an artifact renderer is available. They do not replace missing results with built-in example content. Older packages can differ; this guide describes the integrated workspace and does not certify a new native release.
Open a Saved Run
Open Analysis from the left rail, choose a saved run, and select an artifact. The run list groups results by module. An artifact chip opens its editor view; the run menu can open the full run or its report.
A saved run normally has an analyze/<run-id>/ directory containing an .alLab descriptor and the artifacts actually produced. The descriptor and viewer metadata are not substitutes for simulation, trace, or engine evidence.
The Explorer Folder view can also show analyze/ when the host returns it. The Analysis panel adds run-aware navigation without hiding that directory from the file tree.
Create or Refresh Results
Use the selected host’s available analysis operation to create a new run; preserve its input identity, engine version, diagnostics and saved output. Cloud jobs require an offered operation, permission and QUOTA. Cloud UVM and the Physical coordinator runtime remain unavailable. Opening a saved run, a setup dialog or a tool catalogue does not submit a job, and a rejected or unsupported action is never a successful zero-valued result.
The tool action must produce the required artifacts. Opening a viewer does not launch a simulation, generate a measured trace, or submit an online job. Preserve source version, target, engine version, run receipt, and artifact identity when comparing results.
What Each View Reads
Open Flow Navigator → Viewers → GDS layout viewer for a saved Cloud layout or a saved layout up to 4 MiB in a signed-in 2.0.2 Windows My PC project. This view reads the saved file and does not require an Analysis run. The Windows Testbench waveform comes from its explicit local simulator run and is separate from server analysis.
For a saved circuit, open SPICE DC operating point from Flow Navigator → Simulation. Its separate R/I and R/I/V panels calculate node voltages and branch currents directly from a saved .cir file.
View | Input and behavior |
|---|---|
Waveform | Recorded VCD for the selected run; no fabricated signal replay when input is missing |
Schematic | The selected module’s RTL structure; its module context must not be confused with a run-specific timing result |
Architecture, Structural, 3D Hierarchy | Design views of the same saved schematic result |
Roofline | A run snapshot or supported module-scoped trace; axes relate arithmetic intensity and achieved performance |
Path | The run’s path-analysis payload, shown as saved metrics and rows |
Memory | The run’s memory payload, shown as saved metrics and rows |
Events / Event Trace | One shared Analysis timeline bound to a selected trace |
Perf | The run’s |
Signoff | Available timing, PPA, and evidence artifacts, with missing sections identified |
System Simulator | The run’s saved simulator payload; no onboarding animation used as execution evidence |
Data Flow | The run’s |
Report | The run descriptor and its available artifact content |
Artifact absence, unreadable content, and unsupported capabilities are explicit states. Choose a different run or produce the required artifact; do not interpret an unavailable panel as a measurement.
Event Trace and ISA
Event Trace lives in Analysis. The ISA Execute group retains a shortcut to this same viewer instead of maintaining a second timeline UI. The selected project and trace determine what is shown.
Without a trace, the view asks for a run or reports the missing input. Timeline zoom, selection, markers, inspection, or export depend on the trace and the operations the host exposes. No demo trace is substituted.
Perf
Perf reads perf.json for the selected run and displays the saved observations as metrics and charts. It shows loading or missing-artifact state instead of an empty live-only panel.
Read labels and units before comparing values. Saved throughput, utilization, latency, or counters are not interchangeable, and a chart cannot establish measurement quality that is absent from its underlying receipt.
Waveform and Source Context
Use waveform navigation to inspect the recorded signal timeline, then return to the corresponding source or module context. Signal names, source locations, and available cross-navigation depend on the run data.
A recorded file can still come from an example or synthetic producer. Verify the producer and source binding before calling it evidence of your RTL. The absence of a built-in viewer fallback alone does not establish simulation correctness.
Signoff and Reports
Timing, PPA, and evidence share one Signoff view. It presents available metrics, requirements, and missing sections. A UI verdict or estimate is a review aid, not independent timing closure, manufacturing approval, or formal signoff.
Reports preserve the run context and expose artifact content. Compare the same target, units, source revision, and methodology; retain the original artifacts rather than relying on a screenshot alone.
Troubleshooting
Symptom | Check |
|---|---|
No runs | Selected project, host listing, and whether an action actually saved results |
Tool missing from a run | The descriptor and the artifact files produced for that run |
Empty Event Trace | Trace selection and host-supported trace access |
Perf unavailable | A readable |
Data Flow has no graph | The artifact’s topology content |
Local action unavailable | Use the selected host’s available analysis operation to create a new run; preserve its input identity, engine version, diagnostics and saved output. Cloud jobs require an offered operation, permission and QUOTA. Cloud UVM and the Physical coordinator runtime remain unavailable. Opening a saved run, a setup dialog or a tool catalogue does not submit a job, and a rejected or unsupported action is never a successful zero-valued result. |
Use the selected host’s available analysis operation to create a new run; preserve its input identity, engine version, diagnostics and saved output. Cloud jobs require an offered operation, permission and QUOTA. Cloud UVM and the Physical coordinator runtime remain unavailable. Opening a saved run, a setup dialog or a tool catalogue does not submit a job, and a rejected or unsupported action is never a successful zero-valued result.
Compare two runs
Match the project, source revision, top module and target.
Compare like units and the same analysis method; keep estimates separate from measured results.
Read warnings and missing sections before comparing a headline metric.
Preserve both complete artifacts with their input identity so a later source change cannot be mistaken for the same run.
Using Digital Design Studio 2.0.2
Use the selected host’s available analysis operation to create a new run; preserve its input identity, engine version, diagnostics and saved output. Cloud jobs require an offered operation, permission and QUOTA. Cloud UVM and the Physical coordinator runtime remain unavailable. Opening a saved run, a setup dialog or a tool catalogue does not submit a job, and a rejected or unsupported action is never a successful zero-valued result.
Cloud UVM and the Physical coordinator runtime remain unavailable. Built-in HLS, QEDA layout/solvers and virtual-FPGA Evolve require the Linux native host; they do not run on Windows or Cloud. Static OpenQASM circuit import is a separate Windows/Linux capability. A result viewer, imported report or AI response is not an engine execution or manufacturing signoff.