About

About Fixstars AIBooster

Fixstars AIBooster installs on GPU servers and records utilisation across GPUs, CPUs, memory, storage, PCIe, networks and running AI functions. Time-series dashboards and flame graphs reveal bottlenecks within training and inference workloads across cloud and on-premises infrastructure.

Performance Intelligence tools analyse optimisation opportunities and provide automated or guided acceleration. AcuiRT converts PyTorch inference models to optimised TensorRT execution, while ZenithTune supports hyperparameter optimisation. Fixstars also provides engineering services for model compression, quantisation and memory-bandwidth improvement.

Timeline

Fixstars adds no-code AIBooster tuning and reports faster hyperparameter search
2026-04-13 - Capability - The command-line workflow lets engineers run hyperparameter tuning without writing Python scripts. The updated optimizer identified improved hyperparameters in about one-sixteenth the search time of the previous version in the disclosed benchmark. Training costs fell by as much as 43 percent for the benchmarked workloads.
Source
Fixstars expands AIBooster diagnostics and performance observability
2026-01-08 - Capability - AcuiRT reports now show layer-level conversion success, failure details, inference-accuracy comparisons and latency breakdowns. Teams can define their own performance metrics and tags to monitor the hardware and software signals relevant to a deployment. The release is designed to expose conversion bottlenecks and performance regressions earlier in the model-deployment cycle.
Source
Fixstars updates AIBooster for edge-inference optimization
2025-10-09 - Product Launch - The update introduced autonomous optimization for edge-inference models and expanded optimization beyond AI training workloads. AIBooster added a managed SaaS plan so teams could run optimization and analysis without maintaining the platform themselves. The release paired automated optimization with GPU-resource analysis intended to shorten model-tuning work for deployment targets.
Source
Fixstars launches AIBooster with Sony Honda Mobility use case
2025-01-08 - Product Launch - AIBooster automates optimization of trained AI models for target hardware. Sony Honda Mobility used the product in autonomous-driving development. The work focused on improving GPU efficiency for the AFEELA vehicle-development programme.
Source