Vaaman combines the flexibility of an FPGA with the power of an ARM processor to create something truly unique: a re-configurable computer that adapts to your needs. While traditional computers struggle with the increasing demands of realtime processing and AI workloads, Vaaman takes a different approach. By integrating a powerful six-core ARM processor with an Efinix Trion T20 FPGA with 112k LEs, we’ve created a platform that can dynamically reconfigure itself for optimal performance across diverse computing tasks.
In modern computing, traditional solutions often present developers with an uncomfortable trade-off: they must choose between system flexibility and raw computational performance. These conventional approaches frequently require sacrificing one capability to enhance the other, leading to sub-optimal solutions for complex computing challenges. Vaaman breaks free from this constraint by offering a comprehensive suite of features:
Our platforms are accessible to all software developers without requiring expertise in complex FPGA toolchains or HDL languages. That includes the tools we’ve built for Vaaman, which are straightforward to use:
Gati revolutionizes edge AI deployment by accepting ONNX (Open Neural Network Exchange) deep-learning models as input and automatically re-configuring them to benefit from FPGA acceleration. Gati handles the complex process of translating neural network operations into optimized FPGA configurations so you can focus on your applications, rather than on hardware-implementation details, while still enjoying 25 frames per second on complex object detection models, realtime performance monitoring, and dynamic resource allocation.
Specific Features include:
Suppose you need to connect 40 UART devices to your system. The traditional approach would have you drowning in bridge ICs, soldering for hours, and still wondering if it was all going to work out. Or say you’re looking to combine 10 UART interfaces with 10 SPI interfaces. Conventional wisdom says you’re looking at months of hardware setup, complicated driver development, and a resulting system that might be temperamental at best. That’s where Periplex comes in. This easy-to-use software, which comes with Vaaman, transforms complex challenges like those described above into a simple, five-minute setup process. No additional hardware required!
Implementing this functionality with separate hardware components and chips would require significant time, cost, and complexity. Instead, our system offers:
Vaaman supports wireless configuration. This groundbreaking functionality eliminates the need for physical connections so you can program and reconfigure Vaaman from anywhere on your network, which greatly streamlines both the development process and maintenance of deployed systems.
Vaaman’s unique value lies in its combination of FPGA flexibility, comprehensive peripheral support, and the included Periplex software platform. Thanks to its comprehensive integration of hardware and software, Vaaman supports a broad range of sophisticated applications across multiple domains. Examples include:
Other applications might include edge-AI deployments, realtime video processing, industrial automation, IoT gateway solutions, high-performance computing, custom protocol development, high-speed protocol conversion, hardware prototyping, and other R&D initiatives.
And finally, you can use Vaaman to bring FPGA acceleration to the Raspberry Pi ecosystem thanks to its RPi-compatible GPIOs and available HATs.
| Vaaman | RaspberryPi 5 | Jetson Nano | Alinx AXU2CGB | ULX3S | PolarFire SoC Icicle Kit | Snickerdoodle | Kria KV260 | |
|---|---|---|---|---|---|---|---|---|
| SoC | Rockchip RK3399 | Broadcom BCM2712 | NVIDIA® Tegra® X1 series SoC | Xilinx Zynq UltraScale+ MPSOC XCZU2CG | None | None | None | Zynq™ UltraScale+™ MPSoC |
| CPU | 1.8 GHz ARM Cortex A55 (4 core) + A74 (2 core) | Quad core Cortex-A76 (ARM v8) 64-bit SoC @ 2.4GHz | Quad-core ARM A57 @ 1.43 GHz | Dual-core Arm® Cortex®-A53 MPCore™ up to 1.3 GHz; Dual-core Arm Cortex-R5F MPCore up to 533 MHz | None | 600 MHz clock; 1x RV64IMAC core by SiFive; 4x RV64GC core by SiFive | 32-bit dual-core ARM Cortex-A9 w/640 kB cache and 2x 128-bit NEON co-processors | Quad-core Arm® Cortex®-A53 MPCore™ up to 1.5 GHz |
| GPU | ARM Mali-T860MP4 GPU, support OpenGL ES1.1/2.0/3.0, OpenCL1.2, DirectX11.1 | Broadcom VideoCore VII | 128-core Maxwell | Mali™-400 MP2 up to 667 MHz | None | None | None | Mali™-400 MP2 up to 667 MHz |
| Memory | 2 GB LPDDR4-3200 SDRAM | 4, 8, or 16 GB LPDDR4X-4267 SDRAM | 4 GB 64-bit LPDDR4 25.6 GB/s | 2 GB DDR4, 32 bit, Data Speed 2400 Mbps | None | None | None | 4 GB 64-bit DDR4 (non-ECC) |
| Storage | High-Speed eMMC 16, 32, 64GB, 128 GB; MicroSD (TF) Card Slot; PCIe for extended SSD | MicroSD | microSD; 16 GB eMMC 5.1 | 1 GB eMMC, 256 Mb QSPI FLASH; SD Card Slot; PCIe x1 Interface Slot | MicroSD | 1 Gb SPI flash; 8 GB eMMC flash or SD card slot (multiplexed) | None | 16 GB eMMC |
| Hardware Features | ||||||||
| Display | 1x HDMI 2.0 (Micro)，Support maximum 4K @ 60Hz display; 1x MIPI, Support 2560x1600 @ 60fps output with dual channel; DisplayPort over Type-C | 2 × micro-HDMI ports & 2-lane MIPI DSI display port | HDMI; Display port | Mini DP interface, Supports 4K x 2K @ 30 FPS Output | Placeholder for 0.96" SPI color OLED SSD1331 | None | None | DisplayPort 1.2a |
| Audio | 3.5 mm jack with mic | 3.5 mm jack | 3.5 mm jack | 3.5 mm jack | None | None | ||
| Ethernet | 10/100/1000 Mbps Ethernet (Realtek RTL8211E) | Gigabit Ethernet | Gigabit Ethernet | Gigabit Ethernet | Via ESP32, USB | 2x Gigabit Ethernet | 2x Gigabit Ethernet | Gigabit Ethernet |
| Camera | MIPI CSI 2 lanes via FPC connector, support up to 1.5 Gbps/Lane | MIPI CSI 2 lanes | 2x MIPI CSI-2 DPHY lanes | MIPI Camera Interface | None | None | None | 3 MIPI sensor interfaces, USB cameras |
| Wireless | Integrated WiFi Combo Module (6222B-SRC); Wi-Fi 2.4 G + 5G Wi-Fi 2T2R; BT5.0 | 2.4 GHz and 5.0 GHz IEEE 802.11ac wireless, Bluetooth 5.0 | M.2 Key E | None | ESP32-WROOM-32 | None | 150 Mbps 2x2 MIMO 2.4 GHz/5GHz 802.11n 3 Mbps dual-mode Bluetooth 4.0 Classic+EDR/BLE | None |
| PCIe | PCIe x4 Gen 2 Interface via FPC connector | PCIe gen 2.0 1 lane | None | PCIe x1 Interface Slot | None | PCIe x4 Connector | PCIe x1 Interface Slot | None |
| USB | 2x USB 2.0 Host; 1x USB 3.0; 1x USB 3.0 Type-C | 2x USB 3.0 ports & 2x USB 2.0 ports | 4x USB 3.0; 1x USB 2.0 Micro-B | 4x USB 3.0 Host, Speed up to 5.0 Gb/s | 1x USB 2.0 Micro-B | 1x USB 2.0 Micro-B | None | 4x USB 3.0 Host, Speed up to 5.0 Gb/s |
| RTC | Support RTC, on-board backup battery interface | None | None | None | MCP7940N, Low Power Sleep, Wakeup | None | None | None |
| IOs | 1x UART; 2x SPI bus; 2x I²C bus; 1x PCM/I²S; 1x SPDIF; 1x PWM; 1x ADC; 6x GPIO | GPIO, I²C, I²S, and SPI. (UART optional) | GPIO; I²C; I²S; SPI; UART | 2x 40-pin Expansion Connectors for Modules | 56-pin Expansion Connectors for Modules | 40-pin Raspberry Pi Expansion Connector | 179-pin Expansion Connectors for Modules | Pmod 12-pin interface |
| FPGA | Efinix® T120F324 | None | None | Xilinx Zynq XCZU2CG | Lattice ECP5 LFE5U-85F-6BG381C | Microchip’s PolarFire SoC (MPFS250T-FCVG484EES) | Xilinx Zynq-7020 | Kria KV260 |
| System Logic Cells | 112K | N/A | N/A | 103K | 12K – 84K | 254K | 17.6K | 256K |
| FPGA Memory | DRAM Chip DDR3 SDRAM 4 Gbit or 8 Gbit 256Mx16 1.35V/1.5V; 128 Mbit SPI NOR flash memory | N/A | N/A | 2 GB Integrated with CPU | 32 MB SDRAM 166 MHz | 2 GB LPDDR4x 32 | 16 MB XIP NOR + up to 200 GB SDIO NAND via captive microSD card cage | 4 GB (4x 512 Mb x 16 bit) [non-ECC] |
| FPGA IOs | 40-pin GPIO header: supports one 12-pin Pmod and two LVDS lanes or 25 GPIOs | N/A | N/A | None | 56 | 60 | 100 | 40 |
| Total Block RAM | 5.3 Mb | N/A | N/A | 5.3 Mb | 3.7 Mb | 30 Kb | 630 KB | 4 MB |
| DSP Blocks/Slices | 320 blocks | N/A | N/A | 240 blocks | 156 blocks | 786 blocks | 220 blocks | 1.2 K slices |
| MIPI 4-Lane DPHY with Built-in CSI-2 Controller | 1 TX, 1 RX | N/A | N/A | 2x2 Lane MIPI | None | None | None | Multi-camera Support: Up to 8 interfaces |
| LVDS (TX, RX) | 20 TX, 20 RX | N/A | N/A | None | 28 | None | None | None |
| PLLs | 7 | N/A | N/A | None | None | None | 4 | None |
| Price | $180 | $75 | $212 | $251 | $252 | $599 | $245 | $300 |
Vaaman supports various open-source operating systems, including:
Community-supported operating systems and additional distributions are continuously being developed and will be available soon.
We invite you to explore our comprehensive documentation, which includes in-depth technical guides, API references, and samplme projects that cover everything from getting started to advanced implementations, ensuring you have all the resources needed to develop innovative solutions. Before shipping products to Mouser for fulflilment, we will also publish Vaaman schematics, open-hardware KiCAD design files for all HATs, enclosure design files for all boards, and a verilog FPGA sample project.
If you have any questions, please feel free to reach out using the appropriate Ask a question link below.
We have fully validated both our design and and our manufacturing process:
After our production run is complete, we will package everything up and send it along to Crowd Supply’s fulfillment partner, Mouser Electronics, who will handle distribution to backers worldwide. You can learn more about Crowd Supply’s fulfillment service under Ordering, Paying, and Shipping in their guide.
Hardware manufacturing at scale is a journey filled with challenges. From potential disruptions in supply chains to quality assurance, from component shortages to unforeseen natural disasters, there are numerous hurdles that can impact the successful delivery of a product. And of course software development comes with its own set of risks that demand a meticulously development strategy.
Our foremost commitment is to ensure the quality of Vaaman’s hardware and software products. In order prevent fulfillment delays, we’ve made every effort to account for possible supply-chain delays, quality assurance challenges, and other unexpected obstacles. For its part, our software is continuously tested, broken, fixed, improved, and polished. And that extends not just to core functionality but to wireless connectivity, board support packages, mobile applications, back-end device management, data security, and baseboard functionality, to name just a few peripheral areas of software development.
Whatever happens, we will maintain transparency and responsiveness by posting regular project updates throughout this journey. Backers and subscribers will be kept informed about our progress, our direction, the challenges we encounter, and the steps we take to overcome them. We deeply value your trust and support!
"Open-hardware board looks to accelerate on-device machine learning and artificial intelligence via an on-board FPGA."
"Vaaman’s design emphasizes speed and performance with a fast 300-MBps link between its FPGA and Rockchip RK3399 CPU..."
"The Vaaman is an RK3399-powered SBC that also includes an Efinix Trion T120 FPGA onboard. The RK3399 SoC can be used for powerful yet energy-efficient computing where as the Efinix Trion T120 FPGA can be used for customizable hardware acceleration."
Produced by Vicharak in Surat, India.
Sold and shipped by Crowd Supply.
Single Board Computer with RK3399 processor, Efinix Trion T120 FPGA, 2 GB LPDDR4, and 512 MB DDR3. Comes with 5 ft Type-C PD cable and USB JTAG FPGA programmer
Vaaman NVMe PCI Gen 2 HAT with a flex cable included
Vaaman SATA SSD HAT with a flex cable included. Supports two SATA SSDs simultaneously.
Vaaman FPGA LVDS GPIO Expander supporting up to 80 GPIOs.
Use Vaaman as a software-defined radio with this powerful SDR hat.
· Vicharak_In · vicharak-in · vicharak.in
We are pioneering reconfigurable computing in order to address complex challenges more quickly and easily. Our unique hardware seeks to redefine existing architectures and capabilities so we can move beyond century-old notions of computing.
