okooooooo/EVEngine
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
481 Commits 1 Branch 0 Tags
e95248dbe56237aa22eabbdb2ab9cdebcf8d3d07
Commit Graph
5 Commits
Author SHA1 Message Date
西风逍遥游 cff9bb4d0c tensor: AITemplate-style compile pipeline, gpgpu Sequence, mini-LLM demo and weight-only quantization 
Refactor the tensor module around AITemplate-style static compilation:

- Optimizer: DCE, constant folding, elementwise chain fusion, matmul/conv
  bias+activation epilogue fusion, liveness-based static memory planning,
  topological group execution order.
- KernelGen: shape-baked GLSL per fused group; two-pass softmax/layernorm;
  shared-memory SDPA; naive + 16x16 tiled matmul templates.
- GpuBackend: arena buffer allocation per memory plan, matmul autotuning,
  single command-buffer submission via the new gpgpu::Sequence (uploads,
  dispatches and downloads recorded once per run, Kompute-style).
- shaderc in-process GLSL->SPIR-V on Windows (no glslc.exe dependency),
  with glslc spawn fallback.
- CPU reference kernels shared by eager and the compiled interpreter;
  no-Vulkan fallback keeps compiled graphs runnable headless.

Neural-network coverage: softmax/logsoftmax, layernorm/rmsnorm (with
weight/bias), gelu/silu, conv1d/conv2d, pool, embedding, concat/slice,
axis reduces, argmax, cast, fused SDPA (+masked), resize2d, batched
matmul, seeded random.

Weight-only quantization (activations stay fp32):
- fp16 (IEEE half), fp8 (e4m3), fp4 (e2m1), int8, int4 with per-group
  block scales (MXFP-style: scale = maxAbs / format max magnitude).
- Quantized weights capture into graph Const nodes; CPU interpreter
  dequantizes at materialization; GPU matmul/embedding kernels dequantize
  inline from packed uint buffers (naive matmul variant).
- TF::quantizeWeight API + script binding.

Mini-LLM validation (TinyStories-gpt2-3M, 8 layers d=64):
- test/tensor_llm.cpp builds the full GPT-2 graph from engine ops and
  matches numpy reference: probs within 3e-7, greedy generation 24/24
  identical (~100 ms/token CPU), compiled path exact.
- tensor.llm.quantizedDtypes: fp16/int8/fp8/int4 all keep 16/16 greedy
  top-1 (fp4 16/16 with proper block scaling), weights 27.6 MB -> 3.3-13 MB.
- scripts/export_tiny_gpt2.py converts a PyTorch .bin checkpoint (no torch
  dependency) plus GPT-2 tokenizer into engine assets + numpy reference.

Tests: gpgpu 8/8, tensor 31/31 (original suite) + tensor.llm.* + tensor.quant.*
Assets in test/assets/mini_llm are gitignored (regenerate via the script).
 2026-08-18 17:42:38 +08:00
西风逍遥游 a940088b70 Merge branch 'main' of https://github.com/EVEngine/EVEngine 2026-08-10 22:17:58 +08:00
西风逍遥游 fff24466c6 docs: implement GPGPU backend abstraction plan 
Introduce a comprehensive implementation plan for GPGPU backend abstraction, detailing the transition to backend-agnostic interfaces for ComputeShader, GpuBuffer, and Gpgpu. The plan outlines architecture, global constraints, file structure, and specific tasks for integrating Vulkan while maintaining API compatibility. This change aims to enhance modularity and facilitate future backend support.
 2026-08-10 21:46:33 +08:00
西风逍遥游andCursor Agent 5210135fff feat(gpgpu): bridge ECS systems to compute shaders (#20) 
Add ShaderSystem so ECS batch updates can run as GLSL compute kernels:
pack float component fields into SSBOs, dispatch, and write back.

Script: eve.ShaderSystem extends eve.System with bindFields/update.
C++: gpgpu::ShaderSystem + EcsGpu.h View pack/unpack helpers.

Co-authored-by: Cursor Agent <cursoragent@cursor.com>
 2026-08-10 21:20:44 +08:00
sunxfancyandCursor 87145e5ce2 feat: enhance system and thread modules with new functionalities and documentation updates 
- Added the system module to provide OS, CPU, RAM, wall clock, sleep, clipboard, battery, and GPU information.
- Enhanced the thread module to support thread pools, tasks, and named channels for inter-thread communication.
- Updated documentation to reflect the new features in the system and thread modules, including usage examples and API references.
- Modified CMakeLists.txt to include the new system and thread modules, ensuring proper build configurations.
- Updated load.nut script to initialize the new system and thread functionalities.

Co-authored-by: Cursor <cursoragent@cursor.com>
 2026-08-07 00:06:25 +08:00