mirror of
https://github.com/allaunthefox/Research-Stack.git
synced 2026-07-31 03:05:21 +00:00
e5fb0a5f4d
Lean: update Semantics modules, add new numerics/physics data files Hardware: update FPGA bitstreams (tangnano9k_uart_loopback) Infra: k3s-flake tests, netcup-vps configuration, VCN compute substrate Docs: ARCHITECTURE, specs, citation updates
1.3 KiB
1.3 KiB
| 1 | Pattern Name | Primary Substrate | Secondary Substrates | Total Cost | Total Performance | Value Ratio | Use Case | Complexity | Scalability | Status | Notes |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 2 | $10 Morphic Field Computer | MCU (SidonAVM) | USB-C to SFP + 1km Fiber + PCIe Idle + HDMI TMDS + $5 Display | 10 | 36M ops/sec + 6GB/s + 5M spatiotemporal slots | 3.6M ops/$ | Embedded | Medium | Linear | Specified | Optimal balance of cost and performance with spatiotemporal RAM |
| 3 | $30 Pi Cluster | MCU (SidonAVM) | USB-C to SFP + PCIe Idle + HDMI TMDS + $5 Display (x10) | 30 | 360M ops/sec + 60GB/s + 50M spatiotemporal slots | 12M ops/$ | Embedded | Medium | Linear | Theoretical | Linear scaling of optimal configuration |
| 4 | $100 Pi Cluster | MCU (SidonAVM) | USB-C to SFP + PCIe Idle + HDMI TMDS + $5 Display (x100) | 100 | 3.6B ops/sec + 600GB/s + 500M spatiotemporal slots | 36M ops/$ | Embedded | Medium | Linear | Theoretical | Linear scaling for research workloads |
| 5 | PCIe Harvest Only | PCIe Idle | None | 0 | 10M ops/sec + 100MB/s | Infinite | Desktop | Simple | Linear | Specified | Minimal configuration using only harvested cycles |
| 6 | Display Transport Only | HDMI TMDS | None | 0 | 600M ops/sec + 6GB/s | Infinite | Desktop | Complex | Linear | Specified | Display interface as compute substrate |
| 7 | EPYC Fabric Only | EPYC SP5 | None | 10000 | 5.8B ops/sec + 500GB/s | 580K | Data Center | Very Complex | Super-linear | Theoretical | High-end configuration for complex workloads |