Future Technologies Partition¶
Work in Progress
The FTP is under active development. Hardware and software configurations change frequently. Some nodes or features described here may not yet be available or fully functional.
Functionality may be incomplete or subject to change. Thank you for your patience!
Scheduled Downtime: Shutdown of all HPC Clusters: 20.11.2026 – 04.12.2026
From 23 November 2026 until 4 December 2026 (expected), the computing center building (449.3) at KIT Campus North will be completely disconnected from the power grid due to mandatory inspections and maintenance work on the electrical infrastructure. This will affect all HPC systems operated at KIT. Jobs will only be running until noon on Friday, November 20, 2026. The following clusters will be unavailable during this period:
- bwUniCluster 3.0
- HoreKa 2
- HoreKa
- HAICORE@KIT
- Future Technologies Partition
During this period, access to your data will not be possible. Compute jobs will not run while the systems are offline. Workspace expiration and automatic deletion will be suspended for the duration of the downtime; therefore, no ws_extend request is required.
To avoid additional service interruptions in the near future, we will combine this building-related outage with our own maintenance activities, including operating system updates, software upgrades, file system maintenance, completion of the final phase of the power busbar upgrade, and preparations of the rack infrastructure for the planned GPU-based expansion of the HPC systems in 2027. Most importantly, this will include the final phase of bringing the Jade partition of HoreKa 2 into operation.
Once the maintenance work has been completed, the HPC systems will be brought back into operation one cluster at a time. The restart of all systems is planned to be completed by 4 December 2026.
Please be aware that, following the downtime and commissioning of the Jade partition, the second mandatory migration phase is active. This means that HoreKa (legacy) will only be accessible for logging in and accessing the file systems.
The Future Technologies Partition (FTP) is a hardware testbed providing access to a wide range of emerging processor architectures. New hardware is added continuously and more rapidly than is possible for the main production system. The FTP is also integrated into the Continuous Integration infrastructure.
The FTP is suited for users who want to:
- port their code to architectures beyond standard x86, such as ARM, GPU accelerators from different vendors, or dataflow processors,
- future-proof their software by verifying builds across multiple architectures,
- or conduct research on novel hardware.
Hardware¶
The FTP is split into two cluster installations with separate hardware, software stacks, and login nodes: an x86 cluster and an ARM cluster. Access is granted to the FTP as a whole; the appropriate login node depends on the target hardware.
| Manufacturer | Hardware | Cluster |
|---|---|---|
| NVIDIA | Grace-Grace, Grace-Hopper | ARM |
| AMD | EPYC Milan + MI100, MI210, MI250, Instinct MI300A | x86 |
| Intel | Sapphire Rapids (CPU, HBM, PVC), Gaudi 2 & 3 | x86 |
| Ampere | Altra Max | ARM |
| NextSilicon | Maverick-2 | x86 |
| Q.ANT | NPS 2, NPS 2+ | x86 |
Getting Access¶
Access to the FTP is available to all registered HoreKa users. Access can also be granted to users without a HoreKa account on request. In both cases, contact the Support team and describe the intended use (porting, performance evaluation, research, etc.).
Login Nodes¶
Connect to the appropriate login node depending on the target architecture:
- x86 cluster —
ftp-x86-login.scc.kit.edu - ARM cluster —
ftp-a64-login.scc.kit.edu
Batch System Queues¶
Queues are grouped by host architecture and manufacturer.
Queue |
Node type |
Access policy | Minimum resources | Default resources | Maximum resources |
|---|---|---|---|---|---|
amd-milan-mi100 |
AMD Milan + MI100 | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=8025mb | time=24:00:00, nodes=1, ntasks=128, mem=513600mb |
amd-milan-mi210 |
AMD Milan + MI210 | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=8025mb | time=24:00:00, nodes=1, ntasks=128, mem=513600mb |
amd-milan-mi250 |
AMD Milan + MI250 | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=1, mem=8025mb | time=24:00:00, nodes=2, ntasks=128, mem=1027200mb |
amd-milan-mi300 |
AMD Instinct MI300A | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=5356mb | time=24:00:00, nodes=1, ntasks=192, mem=514176mb |
Queue |
Node type |
Access policy | Minimum resources | Default resources | Maximum resources |
|---|---|---|---|---|---|
intel-spr |
Intel Sapphire Rapids | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=5356mb | time=24:00:00, nodes=2, ntasks=224, mem=512960mb |
intel-spr-hbm |
Intel Sapphire Rapids + HBM | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=6180mb | time=24:00:00, nodes=2, ntasks=208, mem=642720mb |
intel-spr-pvc |
Intel Sapphire Rapids + Ponte Vecchio | Shared | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=6420mb | time=24:00:00, nodes=1, ntasks=160, mem=513600mb |
gaudi2 |
Intel Gaudi 2 AI | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem-per-cpu=7150mb | time=24:00:00, nodes=3, ntasks=144, mem=1029600mb |
gaudi3 |
Intel Gaudi 3 AI | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem-per-cpu=8060mb | time=24:00:00, nodes=1, ntasks=288, mem=2321280mb |
Queue |
Node type |
Access policy | Minimum resources | Default resources | Maximum resources |
|---|---|---|---|---|---|
maverick-2 |
NextSilicon Maverick-2 | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=5356mb | time=24:00:00, nodes=2, ntasks=224, mem=512960mb |
qant-nps1 |
Q.ANT NPS 2 | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=5356mb | time=24:00:00, nodes=1, ntasks=1, mem=512960mb |
qant-nps2 |
Q.ANT NPS 2+ | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=2, mem=5356mb | time=24:00:00, nodes=1, ntasks=1, mem=512960mb |
Queue |
Node type |
Access policy | Minimum resources | Default resources | Maximum resources |
|---|---|---|---|---|---|
nvidia100_2 |
ARM-A100 | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=80, mem-per-cpu=6350mb | time=24:00:00, nodes=4, ntasks=80, mem=522400mb |
grace_grace |
Grace-Grace | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=144, mem-per-cpu=3402mb | time=24:00:00, nodes=6, ntasks=144, mem=489960mb |
Queue |
Node type |
Access policy | Minimum resources | Default resources | Maximum resources |
|---|---|---|---|---|---|
dual_a_max |
Ampere Altra Max | Exclusive | nodes=1, ntasks=1 | time=00:30:00, ntasks=256, mem-per-cpu=2035mb | time=24:00:00, nodes=6, ntasks=256, mem=520960mb |
Acknowledgment¶
Use of the Future Technologies Partition is free of charge, but the resource has to be acknowledged on all publications using the following statement.
Acknowledgment statement
This work was performed on the NHR@KIT Future Technologies Partition testbed funded by the Ministry of Science, Research and the Arts Baden-Württemberg and by the Federal Ministry of Education and Research.