GigaIO and SourceCode have teamed up to bring the world an ultra-portable yet supercomputer-class device for AI needs. The Gryf weighs 55 pounds or less and packs into a TSA-compliant carry-on suitcase.
Despite the small form factor, Gryf can accommodate data collection and processing on a scale that would otherwise require sending data offsite. This revolutionary development for use cases requires rapid processing and analytics transformation.
Gryf is a suitcase-sized supercomputer that supports disassembling and reassembling its GPUs. The user can customize the computer's hardware configuration in the field on the fly. You can create the best hardware configuration for one assigned workload and then change it for the next.
Each Gryf has multiple slots populated with compute, storage, accelerator, and network sleds depending on the workload. The suitcase-sized supercomputer has six sled slots for inserting and removing modules as needed.
For AI or ML workloads, for example, you can deploy two compute sleds, one accelerator sled, two storage sleds, and one network sled. Are you moving towards a storage project? Change the configuration to include one compute sled and five storage sleds instead.
Following are the specifications and capabilities for each type of sled and the Gryf platform itself:
| CPU | (1) AMD EPYC 7003 series, 7313, 16 cores/32 threads, 155W, 3Ghz base/3.7Ghz turbo |
| System memory | (4) 64 GB DDR4 DIMMs (256 GB total), 3,200Mhz max |
| OS storage | (1) 512GB NVMe-M.2 SSD |
| OS support | Linux Rocky 8/9 or Ubuntu 20/24 |
| Networking | (2) QSFP56/QSFP28/QSFP+ 100GbE, copper/optical |
| BMC/IPMI | Via platform RJ45 |
| Accelerator slot | Single/double wide PCIe-FHFL form factor, up to 350W |
| Accelerator | (1) Nvidia L40S-48GB (other options to be eligible) |
| Storage | (8) 30TB NVMe-E1.L SSD (246TB total) |
| Ports | (2) QSFP56-100GbE, copper/optical |
| Row 1 – Cell 0 | (6) SFP28-25GbE, copper/optical |
| Sledge slots | (6) Anywhere for Compute Sledge, Accelerator Sledge, Storage Sledge, Network Sledge |
| AI Memory Fabric | Internal: PCIe 256Gb/s board-to-board |
| Row 2 – Cell 0 | Expansion: (8) FabreX Mini-SAS-HD-32G 32Gb/s (256Gb/s total) for data offload to Gryf daisy chaining or home base |
| Management network | (5) RJ45-10GbE/1GbE for FabreX Fabric Manager and Out-of-Band (OOB) Baseboard Management Controller (BMC) / IPMI |
| Software | The GigaIO FabreX Fabric Manager is preloaded. |
| Row 5 – Cell 0 | Linux and Libraries: GPUDirect RDMA (GDR), Libfabric, and NVMe-oF |
| power | Dual AC/DC 2,500W 1+1 power supplies |
| Row 7 – Cell 0 | IEC-320-C13 Power Inlet, 100-240 VAC @ 50 to 60Hz |
| fan | (6) 60mm fans are dynamically optimized for the system workload. |
| Fan filters | Removable 45 ppi filters |
| Dimensions | With Wheels: 9.00″ x 14.00″ x 24.50″ (228.6mm x 355.6mm x 622.3mm) |
| Row 11 – Cell 0 | Without Wheels: 9.00″ x 14.00″ x 22.00″ (228.6mm x 355.6mm x 558.8mm) |
| The weight | Less than 55 lbs (24.95 kg) maximum |
| Enclosure | Rugged carbon fiber, detachable top with folding handle, detachable bottom with wheels |
| Ambient Operating Temperature: | 10°C to 32°C (50°F to 90°F) |
| Eligibility for service | Field Replaceable Units (FRUs): Sledges, Power Supplies, Fan Trays with Fan Filtration, Case Top and Bottom Covers |
| compliance | FCC Class A, CE |
According to GigaIO and SourceCode, a single Gryf can be configured to process petabytes of information. Using GigaIO's FabreX memory fabric, the Gryf can also be stacked with up to four other Gryfs for even greater workloads.
Back in the data center, the FabreX memory fabric allows the Gryf to connect to the main computer, a GigaPod, for more essential processing and analysis tasks. Instead of waiting for days to transmit data over Internet connections, engineers collect and begin processing the data where it was collected, then transport it to the data center.
Citing the needs of Department of Defense customers, GigaIO CEO Alan Benjamin identified the need to collect and process data where it happened in the field.
“This is true for our Department of Defense customers, who have emphasized the critical need for timely and actionable intelligence in the field. Gryf's novel architecture, made possible by FabreX, our AI memory fabric, will help these customers provides the advanced compute, storage, and GPU capabilities they demand in today's sensor-rich environment.”
The companies did not disclose pricing for the Gryf, but the mobile supercomputer is available for purchase now.