https://github.com/wolftech-innovations/project_lupine

A x64 CPU for general use
https://github.com/wolftech-innovations/project_lupine

Last synced: 4 months ago
JSON representation

A x64 CPU for general use

• Host: GitHub
• URL: https://github.com/wolftech-innovations/project_lupine
• Owner: WolfTech-Innovations
• Created: 2024-12-18T00:19:01.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2024-12-19T07:53:27.000Z (over 1 year ago)
• Last Synced: 2025-04-06T08:18:46.051Z (about 1 year ago)
• Size: 11.7 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# CPU Blueprint: Codename "Lupine"

## Overview

- **Architecture**: x86-64 (AMD64)

- **Core Count**: 6 physical cores (no hyper-threading for simplicity).

- **Manufacturing Process**: 12nm FinFET.

- **Clock Speed**: Base 3.2 GHz, Boost 4.0 GHz.

- **Cache Hierarchy**:

  - **L1 Cache**: 64 KB per core (32 KB Instruction, 32 KB Data).

  - **L2 Cache**: 512 KB per core.

  - **L3 Cache**: 12 MB shared.

- **TDP**: 65W.

- **Memory Support**: Dual-channel DDR4-3200.

- **Integrated Features**:

  - PCIe Gen 3 x16 controller.

  - On-die memory controller.

  - Basic I/O management (SATA, USB controllers).

---

## Block Diagram

### Core Layout

1. **Execution Core**:

   - 4-wide decode pipeline.

   - 3 ALUs and 2 FPUs per core.

   - 256-bit AVX support.

   - Out-of-order execution with 192 instruction queue.

2. **Pipeline**:

   - Fetch: 2 stages.

   - Decode: 2 stages.

   - Execute: 4 stages.

   - Commit: 2 stages.

3. **Cache Access**:

   - L1 Data Cache: 2-cycle latency.

   - L2 Cache: 12-cycle latency.

   - L3 Cache: 40-cycle latency (shared).

### Interconnect

- **Ring Bus**:

  - Links all six cores with the L3 cache.

  - Bandwidth: 32 GB/s.

### Power Management

- **Voltage Planes**:

  - Per-core dynamic voltage scaling.

  - Sleep states: C1 (idle), C6 (deep sleep).

### I/O and Integration

- PCIe Controller:

  - 16 lanes for GPU or other peripherals.

  - 4 additional lanes for NVMe storage.

- Memory Controller:

  - DDR4 with ECC support.

---

## Die Layout

1. **Core Cluster**:

   - Six identical cores arranged in a 2x3 grid.

2. **Cache Structure**:

   - L3 cache surrounds the core cluster for minimal latency.

3. **I/O and Memory Controllers**:

   - Positioned at the periphery for signal routing efficiency.

4. **Power and Clock Distribution**:

   - Centralized clock generator with tree distribution.

   - Dedicated power planes for cores and uncore components.

---

## Assembly Instructions

### Material Requirements

- **Silicon Die**: 12nm process node with FinFET transistors.

- **Substrate**: Organic PCB for packaging.

- **Pins**: 1151-pin LGA.

### Manufacturing Steps

1. Design Mask: Layout according to the above block diagram.

2. Photolithography: Use EUV for sub-7nm feature etching.

3. Layer Assembly: Stack transistor layers for each core, cache, and interconnect.

4. Testing and Validation: Run benchmarks and stress tests to ensure reliability.

---

## Notes

- **Expandable Features**: Support for AVX-512 can be added in future iterations.

- **Debugging Tools**: Include JTAG ports for debugging and firmware updates.

- **Cooling Requirements**: Recommend a basic air cooler for 65W TDP, or liquid cooling for overclocking scenarios.