SC26 · The International Conference for High Performance Computing, Networking, Storage, and Analysis

SC26, Chicago
Booth 2960.

We’re excited to exhibit at SC26 in Chicago, and look forward to exchanging ideas with people across high performance computing, AI and research. Our theme this year is AskDona Assessment. Stop by and say hello!

Dates
November 16-19, 2026
Venue
McCormick Place West, Chicago
Booth
2960

From the entrance
to Booth 2960.

Booth 2960 · McCormick Place West

Take a look
around our booth.

Drag to look around the booth.

Open full screen ↗︎

How AskDona
supports Fugaku.

AskDona gives AI agents the specialized knowledge of the supercomputer Fugaku that general-purpose agents lack, so the code they write fits the machine it runs on.

Researcher
“I want to port my GPU code to run fast on Fugaku. Simulate CO2 + water fracturing in a 50 m rock block for geothermal power.”

AI Agent

Claude Code, Codex, etc.

  • Generates the code
  • Submits and runs jobs
  • Connects to data storage

Runs on commercial or open-weight LLMs.

The supercomputer Fugaku at RIKEN

Supercomputer Fugaku

RIKEN Center for Computational Science

Runs the code the agent writes

AskDonaMCP

The context layer

Specialized knowledge of Fugaku that general AI agents lack:

  • 16,000+ pages of Fugaku manuals
  • Environments, compilers, libraries
  • Job scripts and tuning information

Data Archives,
Storage

Accelerating AI for Science

AskDona Assessment

AI that checks your documents
against regulatory criteria.

AskDona is GFLOPS’s evidence-based AI assessment platform. For every item on a checklist or requirements list, it checks your own documents (policies, procedures, reports and records), makes a first-pass judgment, and returns the verdict, the reasoning, the supporting passage and its source. Your experts make the final decision.

  1. 01 Your documents
    & criteria
    Policies, procedures and reports, plus your regulatory criteria
  2. 02 AskDona
    assessment
    AI checks every item against your documents
  3. 03 Evidence-based
    judgment
    AI provides preliminary assessment. Then your team decides.

FDAISO 13485EU MDRQMSRMDSAPPMDA

Why teams trust it

Three principles
behind every result.

Evidence-first

Every judgment traced to its source.

Each judgment comes with its reasoning, the supporting passage and the source document, so reviewers check the basis instead of trusting the answer.

Human-in-the-loop

AI that supports human judgment.

AI makes the first pass on every item; your experts make the final call, focusing on the items flagged for attention.

Deterministic

Same criteria. Same evidence. Same result.

Byte-identical judgments, every run, so a recorded result can be reproduced for validation and audit.

Read our determinism study →

What teams assess

Gaps, risks
and controls.

Gaps

Gaps between current practice and the requirements set by regulations, standards and internal policies

  • Regulatory
  • Standards
  • Internal policies

Risks

Risks related to product safety, cybersecurity and operations

  • Product safety
  • Cybersecurity
  • Operational

Controls

Adequacy and effectiveness of design controls, security controls and change controls

  • Design
  • Security
  • Change

Learn more about AskDona Assessment → Enterprise-level security →

ブース2960で、
お待ちしています。
See you at
Booth 2960.

会期中のデモやご相談も、お気軽にどうぞ。We're happy to give a demo or talk through your use case.