Skip Lag Day.
ALPCRUN.CH is a next-generation high-performance computing platform engineered to eliminate bottlenecks. Use any programming language, it's real-time optimized, easy to integrate, and built to save you money. Run it in the cloud or on-premise.
// connect → open a session → stream tasks
conn, queue, _ := grpccon.ConnectToQueue(addr, ca, true)
stream, _ := queue.ClientStream(ctx)
session, _ := queue.CreateSession(ctx, &pb.Session{
Attributes: &pb.SessionAttributes{
Priority: proto.Uint32(5),
MinWorkers: proto.Uint32(0),
MaxWorkers: proto.Uint32(50000),
},
})
for _, batch := range tasks {
stream.Send(batch) // results stream back live ←
}Three promises, one platform.
Intelligent priority queueing and a caching-first architecture keep tasks and their data moving, so your system keeps pace without wasting budget.
Real-time optimized
Priority queueing and pull-through caching keep tasks and their data moving quickly, giving you faster response times and smoother workflows under pressure.
Easy to integrate
Plug into your stack with our language bindings, libraries, and SDK, or talk to the gRPC API directly with protobuf serialization. Skip the glue code and endless dependencies: use any language.
Built to save money
A scale-on-demand backend means you only use the resources you actually need. Control costs with optional usage limits per cluster or environment.
The orchestration core.
A queue- and cache-based architecture for submitting workloads, managing distributed workers, and collecting results at scale.
Bidirectional Streaming
Efficient task and result streaming over a single persistent gRPC channel.
Priority Scheduling
Control task execution order with flexible, per-session priority levels.
Distributed Caching
Share data across workers with versioned, TTL-based pull-through caches.
Session Management
Isolated workload contexts with configurable, scoped resources.
Dead Letter Queues
Failed tasks are captured automatically for inspection and retry.
Scalable Architecture
Horizontally scalable services designed for flexible cloud deployment.
Flow Control
Built-in backpressure mechanisms prevent overload before it happens.
Every primitive, fully documented.
Explore the docs →A clean three-tier grid.
Clients submit work, the data layer schedules and caches it, and workers crunch the tasks. Results flow back out through the same channels, so data moves both ways in real time.
Clients
Submit workloads and receive results back over one bidirectional gRPC stream.
Data Layer
Queues and caches route tasks out and results back: scheduling, priority, state & data transfer.
Workers
Pull tasks in, crunch on them, and stream results back up the grid.
No steep learning curve.
You write a client and a service container and we handle the rest. No need to untangle Kubernetes yourself. Let your devs use what they love.
- ▸Use our SDK & language bindings to get moving fast, or call the gRPC API directly: your choice.
- ▸Bring any language with a gRPC binding, and you're off to the races.
- ▸Dedicated self-service clusters: spin up or tear down your own isolated cluster with a single click, in minutes, straight from the dashboard.
- ▸Logging & analytics out of the box: every cluster ships with logging pipelines and usage insights, no setup required.
- ▸Standard container tooling like Docker or Podman manages your image lifecycle on the platform: build, tag, and push as you already do.
# run 100 tasks across an elastic worker pool
$ export METALCORE_QUEUE_MANAGER_ADDR=queue:1337
$ go run main.go
Connecting to queue manager (TLS: true)
Connected successfully
Session created: 01HZXY…
Submitting 100 tasks…
Progress: 47/100 results received
Progress: 100/100 results received
✓ All tasks completed · avg 74.25Run more with less.
ALPCRUN.CH squeezes the most out of your current infrastructure, and in the cloud makes sure every resource is used efficiently or scaled down.
Built for parallel & distributed workloads.
Financial risk, scientific simulation, ML sweeps, media pipelines. If it parallelizes, it crunches here.
Monte Carlo Simulations
- Price exotic derivatives & structured products
- Value at Risk (VaR & CVaR) for portfolios
- Credit risk & default probability modeling
Parameter Sweeps
- Grid, random & Bayesian hyperparameter search
- Sensitivity analysis across thousands of variants
- A/B simulation for features & scenarios
Batch Processing
- Video transcoding & media pipelines
- Large-scale ETL for warehouses
- Genomic alignment & bioinformatics
GPU Computing
- Distributed deep-learning training
- Molecular dynamics for drug design
- Real-time ray tracing & rendering
Scientific Computing
- Astrophysics & N-body simulations
- Weather & atmospheric modeling
- Protein folding & molecular docking
Mission-Critical Compute
- Ballistic trajectory & intercept analysis
- Radar & sonar signal processing
- Satellite orbit propagation
From PoC to production.
We help you learn the ropes, integrate your workloads, and run mission-critical systems, fully managed or inside your own environment.
Professional Services
From dev to PoC and production, we help you learn the platform and get your workloads integrated and set up properly.
Training & Support
Dedicated weekly support calls, onboarding for your 1st and 2nd level teams, and expert 3rd level support in your on-call rotation.
Managed or Self-Hosted
Run it as a fully managed service with developer self-service: spin up and tear down entire clusters with a single click, in minutes. Or self-host on your own cloud or on-premises with full enterprise support.
Advance your HPC story to the next chapter.
Financial risk or pricing, science & engineering simulations, forecasts, backtests, or any parallel and distributed workload: ALPCRUN.CH helps you move faster, integrate easier, and spend smarter.