Will this model run on RISC-V?
RVStack AI helps developers check, optimise, benchmark and package AI models for RISC-V software environments.
Architecture
Eight stages, upload to deployable bundle.
The pipeline answers one question in stages and stops at the first stage that says no. An unimplemented operator is a cheaper answer than a failed build.
Memory and latency are estimates until the benchmark stage runs on a real target. The page labels them that way, and a benchmark figure will carry the environment it was produced on.
Model upload
A model file enters the pipeline with its intended target profile.
Format check
Identify the format, its version and the toolchain that exported it.
Operator scan
Enumerate operators and flag those with no RISC-V implementation.
Quantisation
Report stored precision and whether a smaller one is viable.
Runtime selection
Choose a runtime for the format and operator set, or report that none fits.
RISC-V test
Build runtime and model for the target and confirm they load and execute.
Benchmark
Record timings and memory from a real run, with the environment named.
Deployment export
Emit model, runtime and configuration as one installable bundle.
Modules
What each stage is made of.
Ten modules. Seven are at research stage and three are scheduled behind them.
Model intake
Accept a model file and identify its format, version and export toolchain.
Operator scan
List every operator the graph uses and flag those with no RISC-V implementation.
Quantisation check
Report the precision the model is stored at and whether a smaller one is viable.
Runtime selection
Recommend a runtime for the format and operator set, or state that none is suitable.
Memory estimate
Estimate weights, activations and peak working set against a profile budget.
Latency estimate
Estimate inference cost against a profile, stated as an estimate rather than a measurement.
RISC-V build test
Build the runtime and the model for the target and confirm it loads and executes.
Benchmark report
Record real timings from an emulated or physical run, with the environment named.
Deployment bundle
Package model, runtime and configuration as one installable artefact.
Store listing
Publish the bundle with its scan results, licence and tested date attached.
Model targets
Formats and categories under assessment.
Four formats and five model categories. The level column will change per row as the work progresses.
| Model target | Type | Level | What is being assessed |
|---|---|---|---|
| ONNX | format | Research | Broadest operator coverage to assess; the first format under investigation. |
| TensorFlow Lite | format | Research | Mobile-oriented and already quantisation-aware, which suits a constrained profile. |
| GGUF | format | Research | Quantised language models; the runtime question is vector-extension support. |
| PyTorch export | format | Research | Requires an export step before any RISC-V question can be asked. |
| Speech models | category | Research | Recognition and synthesis at sizes a mobile power budget allows. |
| Vision models | category | Research | Classification and detection; sensitive to available memory bandwidth. |
| Text models | category | Research | Small language models, dominated by quantisation and memory. |
| Translation models | category | Research | Sequence models with a wide operator surface to check. |
| Recommendation models | category | Research | Embedding-heavy; storage and lookup cost matter more than compute. |
Not every model runs on every architecture. The scanner identifies which will and which will not, before development time is committed.
RISC-V