Alex Zarov
618ecfb0c9
docs: Update README.md
2025-06-11 19:45:50 +10:00
Alex Zarov
c4ca746a60
Merge pull request #4 from Triex:feat--Implement-dynamic-benchmark-summary-with-real-performance-metrics
...
feat: Implement dynamic benchmark summary with real performance metrics
2025-06-11 19:43:10 +10:00
Triex
18097ee5d3
feat: implement dynamic benchmark summary with real performance metrics
...
- Replace mocked performance estimates with actual measured results
- Add `BenchmarkResults` struct to collect live performance data during execution
- Implement honest dynamic summary showing real GFLOPS, timing, and bandwidth
- Add transparent performance assessment based on measured values only
- Display peak performance identification (1160 GFLOPS measured at 512×512)
- Include real memory bandwidth (20.3 GB/s) and latency (1.8 ns) measurements
- Remove misleading static efficiency percentages with live measurement system
- Show clear distinction between measured performance and theoretical estimates
- Provide actionable insights from Apple Accelerate backend performance
Results: 1160 GFLOPS peak measured performance with honest assessment,
eliminating misleading hardcoded comparisons in favor of real benchmark data.
2025-06-11 19:41:51 +10:00
Alex Zarov
2c0ad7fe97
Merge pull request #3 from Triex:feat--BLAS-integration-working---significant-matrix-operation-improvements
...
feat: BLAS integration working - significant matrix operation improvements
2025-06-11 19:33:50 +10:00
Triex
c8eefc8865
feat: BLAS integration working - significant matrix operation improvements
...
Matrix Performance Improvements:
- ✅ Apple Accelerate backend integrated and functional
- ✅ Matrix ops: 1004 GFLOPS (38.6% efficiency) on 1024×1024
- ✅ Significant speedup: 6418ms naive → 2.1ms BLAS
- ✅ Draft implementation with working acceleration
Performance Results (Apple M1, debug build):
- Matrix 256×256: 0.1ms, 561 GFLOPS (21.6% efficiency)
- Matrix 512×512: 0.2ms, 1129 GFLOPS (43.4% efficiency)
- Matrix 1024×1024: 2.1ms, 1004 GFLOPS (38.6% efficiency)
- Matrix 2048×2048: 21.5ms, 799 GFLOPS (30.7% efficiency)
System Integration:
- ✅ Memory bandwidth: 23.5 GB/s
- ✅ Access latency: 1.8ns
- ✅ Apple Silicon detection working
- ✅ BLAS backend selection functional
Web Layer Updates:
- Enhanced /health endpoint with BLAS status
- New /performance endpoint with benchmark data
- Module dependency conflicts resolved
- Hardware acceleration reporting
Implementation Status:
- Matrix operations now use BLAS acceleration
- Foundation ready for transformer development
- DeepSeek V3 model implementation next priority
- Experimental/draft status maintained
This represents significant progress in the experimental foundation - matrix operations now deliver good performance while maintaining the zero-deployment-complexity advantage of Zig.
2025-06-11 19:30:33 +10:00
Alex Zarov
24d94f7c21
Merge pull request #2 from Triex:feat-Enhanced-device-detection-handling-initial-metal
...
Feat-Enhanced-device-detection-handling-initial-metal
2025-06-11 17:50:54 +10:00
Triex
7b81ea27d7
docs: Tidy root README, add hardware notes to experimental/README.md
2025-06-11 17:48:38 +10:00
Triex
0f980354f8
feat: Enhanced device detection handling, added metal initial draft, theoretically-reliable metal mac detection -> experimental implementation
...
✅ Implemented initial Apple Silicon detection using sysctl system calls
✅ Added proper M1/M2/M3/M4 generation detection via CPU brand string
✅ Fixed memory leaks that occured during dev with proper allocator cleanup
✅ Enhanced Metal backend foundation with device capabilities
✅ Added `test_m_series.zig` for hardware verification
🔧 Key Technical Improvements:
- Real hardware detection via `hw.model` (eg; `MacBookPro17,1`)
- CPU brand string parsing for accurate M-series identification
- Unified memory strategy detection (even under Rosetta)
- Apple Neural Engine capability detection
- Memory-safe device info structures
🧪 Verified on Apple Silicon:
- M1 correctly detected (generation 1, no variant)
- 16GB unified memory properly identified
- Builds cleanly with Zig `0.15.0-dev.703+597dd328e`
- No false positives for M1 Pro/Max/Ultra variants
📋 Updated README status to reflect experimental draft implementation
⚠️ Clearly marked as research/development foundation, not production ready
2025-06-11 17:43:04 +10:00
Alex Zarov
68c4c77600
docs: Update README.md
2025-06-08 22:46:46 +10:00
Triex
b1a862818a
docs: Tidy README
2025-06-06 16:39:22 +10:00
Triex
bcee49badf
docs: Tidy experimental README
2025-06-06 16:03:51 +10:00
Triex
8aa2785fad
docs: Tidy experimental README status section
2025-06-06 16:00:24 +10:00
Triex
16fec1d4e9
docs: Update experimental README to reflect current state / performance
2025-06-06 15:58:39 +10:00
Triex
b1c1f2c07f
feat: Update bench + core/root -> latest dev Zig ver
2025-06-06 15:55:07 +10:00
Triex
24b5fcfd02
docs: Tidy experimental/README.md
2025-06-06 15:48:21 +10:00
Alex Zarov
1d7855c807
Merge pull request #1 from Triex:feat--Migrate-experimental-implementation-to-modern-Zig,-achieve-clean-compilation
...
Feat--Migrate-experimental-implementation-to-modern-Zig,-achieve-clean-compilation
2025-06-06 15:34:31 +10:00
Triex
31ef81000f
feat: Migrate experimental implementation to modern Zig, achieve clean compilation (private repo dump -> /experimental)
...
- Port HTTP server, and appropriate points across core etc from old API to Zig `0.15.0-dev` patterns
- Fix mutability, unused variables, and API compatibility issues
- Validate SIMD tensor operations and backend architecture
- Foundation now compiles cleanly and produces working binary
2025-06-06 15:31:21 +10:00
Triex
5ff856c018
docs: Tidy
2025-06-05 05:17:21 +10:00
Triex
90627c6d54
docs: Tidy spacing/line breaks
2025-06-05 04:11:55 +10:00
Triex
a7f0a5391b
docs: Tidy
2025-06-05 04:10:56 +10:00
Triex
9c25e23773
docs: Improved README, add additional references
2025-06-05 04:08:59 +10:00
Triex
9aedaae1d5
docs: Tidy list items @ README
2025-06-04 11:41:41 +10:00
Triex
69c1bab49e
docs: Further tidy initial proposal idea
2025-06-04 11:38:26 +10:00
Triex
e480e15e5f
docs: Tidy README
2025-06-04 11:36:38 +10:00
Triex
5c57ab1f8c
docs: Move why section up to top of root README proposal/architecture notes - more cohesive flow
2025-05-23 04:23:39 +10:00
Alex Zarov
21654d7511
docs: Tidy title
2025-05-23 04:09:48 +10:00
Triex
ef39c76b2d
docs: Tidy title
2025-05-23 04:08:57 +10:00
Triex
59c80bf948
docs: Enhanced draft code, table of contents + copy
2025-05-23 03:54:47 +10:00
Triex
715d0d2e6d
docs: Tidy introduction
2025-05-23 03:33:03 +10:00
Triex
3af7848785
docs: Initial architecture notes for Zig implementation
2025-05-23 03:29:53 +10:00
Triex
a1895012dd
feat: Initial MacBook optimisation draft for DeepSeek V3 inference > moving to Zig instead
2025-05-23 01:53:02 +10:00
Xingkai Yu
4cc6253d5c
Merge pull request #666 from codinglover222/deepseek-doc-fix
...
fix an args description.
2025-04-09 09:50:40 +08:00
Huang Panpan
57d7bd45df
Merge pull request #736 from shihaobai/main
...
Docs: add LightLLM as supported engine
2025-04-08 22:18:33 +08:00
Xingkai Yu
88d6547df2
Merge pull request #816 from KPCOFGS/main
...
Update README.md
2025-04-08 17:27:09 +08:00
Xingkai Yu
741b06ebca
Merge pull request #720 from xiaokongkong/main
...
modify the explanation of MLA
2025-04-08 17:20:37 +08:00
Shixian Sheng
a5d2ad229e
Update README.md
2025-03-26 08:58:35 -04:00
DeepSeekDDM
a878eada08
Delete DeepSeek_V3.pdf
2025-03-16 23:42:21 +08:00
DeepSeekDDM
98e67a71f4
Update paper link
2025-03-16 23:41:52 +08:00
shihaobai
408e6e188a
Update README.md
...
polish
2025-03-03 20:16:37 +08:00
shihaobai
73f2954fa8
polish
2025-03-03 20:10:18 +08:00
shihaobai
1ab09c8780
Docs: add LightLLM as supported engine
2025-03-03 19:23:08 +08:00
huxuedan
d29a967601
modify the explanation of MLA
2025-02-26 17:07:39 +08:00
DeepSeekDDM
592fd5daf8
Delete CITATION.cff
2025-02-24 11:50:20 +08:00
DeepSeekDDM
c9353aba6c
Update bib info
2025-02-24 11:25:44 +08:00
Huang Panpan
f09f5fa321
Merge pull request #616 from Konano/chore-readme
...
chore: update README.md to improve layout
2025-02-18 18:04:06 +08:00
oyzh
4a65fd9221
fix an args description.
2025-02-15 11:02:28 +08:00
Xingkai Yu
1398800ebf
fix scores mask
2025-02-14 20:26:45 +08:00
Konano
f07bccc49e
fix: resolve center alignment issue in preview
2025-02-14 12:12:16 +08:00
Konano
0866cab5f9
chore: update README.md to improve layout and image attributes
2025-02-14 12:02:10 +08:00
Konano
e15f67af1c
chore: update README.md to improve layout and image attributes
2025-02-08 18:28:40 +08:00
