2026-03-01 - 2026-08-29
Overview
13 Issues closed from 1 user
Closed
#10 DGR-026: Provision exact split-GGUF artifacts outside /home
Closed
#12 DGR-028: Implement numbered patch-stack apply and verification
Closed
#13 DGR-029: Create the native CMake skeleton and deterministic CPU lane
Closed
#8 DGR-024: Implement real generated-gRPC protocol harness
Closed
#4 DGR-020: Run the controlled whole-model GGUF baseline
Closed
#3 DGR-019: Lock alpha and beta performance contracts
Closed
#9 DGR-025: Define exact artifact and runtime recipe identity
Closed
#7 DGR-023: Make Python and C++ protobuf generation reproducible
Closed
#11 DGR-027: Add exact llama.cpp provenance manifest and fetch workspace
Closed
#6 DGR-022: Define Shard lifecycle and structured status RPCs
Closed
#5 DGR-021: Define the versioned named-tensor stream envelope
Closed
#1 DGR-017: Reconcile and clean the superseded DGR backlog
Closed
#2 DGR-018: Define canonical Ralph and Gitea metadata schema
55 Issues created by 1 user
Opened
#2 DGR-018: Define canonical Ralph and Gitea metadata schema
Opened
#7 DGR-023: Make Python and C++ protobuf generation reproducible
Opened
#8 DGR-024: Implement real generated-gRPC protocol harness
Opened
#4 DGR-020: Run the controlled whole-model GGUF baseline
Opened
#3 DGR-019: Lock alpha and beta performance contracts
Opened
#5 DGR-021: Define the versioned named-tensor stream envelope
Opened
#1 DGR-017: Reconcile and clean the superseded DGR backlog
Opened
#6 DGR-022: Define Shard lifecycle and structured status RPCs
Opened
#9 DGR-025: Define exact artifact and runtime recipe identity
Opened
#13 DGR-029: Create the native CMake skeleton and deterministic CPU lane
Opened
#14 DGR-030: Add accelerator build presets and native CI matrix
Opened
#15 DGR-031: Introduce the project-owned ShardEngine interface
Opened
#11 DGR-027: Add exact llama.cpp provenance manifest and fetch workspace
Opened
#10 DGR-026: Provision exact split-GGUF artifacts outside /home
Opened
#12 DGR-028: Implement numbered patch-stack apply and verification
Opened
#20 DGR-036: Prove dense fixture and real-model range parity
Opened
#22 DGR-038: Implement isolated shard-local Hot KV State
Opened
#17 DGR-033: Build a standalone fake C++ gRPC Shard worker
Opened
#18 DGR-034: Implement dense-Llama range-aware GGUF ownership
Opened
#16 DGR-032: Implement deterministic fake ShardEngine
Opened
#19 DGR-035: Implement dense architecture boundary input/output
Opened
#21 DGR-037: Bind llama.cpp to the standalone worker
Opened
#25 DGR-041: Register native Shard capabilities without redesigning Meshnet
Opened
#26 DGR-042: Carry native frames through direct and existing relay seams
Opened
#28 DGR-044: Pin the DeepSeek V4 Flash target contract
Opened
#24 DGR-040: Add node-side native worker supervision
Opened
#23 DGR-039: Pass local two-process dense acceptance
Opened
#29 DGR-045: Inventory V4 GGUF tensors and layer ownership
Opened
#30 DGR-046: Define the V4 typed architecture boundary schema
Opened
#27 DGR-043: Expose GGUF compatibility and measured cost inputs to existing routing
Opened
#32 DGR-048: Carry token-ID sideband through the first three hash-routed layers
Opened
#31 DGR-047: Adapt the upstream V4 mHC boundary for ranged ownership
Opened
#34 DGR-050: Validate upstream V4 MoE and hash-routing execution under ranged ownership
Opened
#36 DGR-052: Pass local V4 whole-model versus ranged parity
Opened
#35 DGR-051: Assemble the DeepSeek V4 Flash ShardEngine adapter
Opened
#33 DGR-049: Keep V4 attention and auxiliary state shard-local
Opened
#38 DGR-054: Enforce the DeepSeek V4 Flash alpha gate
Opened
#37 DGR-053: Certify a real 2–4-stage V4 route
Opened
#40 DGR-056: Enforce bounded admission, queues, and backpressure
Opened
#41 DGR-057: Benchmark batching saturation and fairness
Opened
#42 DGR-058: Harden worker and stream failure semantics
Opened
#43 DGR-059: Implement alpha-safe route recovery
Opened
#44 DGR-060: Certify V4 long-context state correctness
Opened
#39 DGR-055: Implement compatible continuous decode batching
Opened
#51 DGR-067: Certify the backend capability matrix
Opened
#49 DGR-065: Define V4 MTP ownership and distributed state contract
Opened
#45 DGR-061: Certify existing routing with 10+ GGUF stage candidates
Opened
#48 DGR-064: Add negotiated activation compression
Opened
#47 DGR-063: Profile and optimize the critical seam
Opened
#50 DGR-066: Implement and benchmark V4 MTP
Opened
#46 DGR-062: Certify a dynamic real 10+ stage V4 scenario
Opened
#52 DGR-068: Package reproducible native worker releases
Opened
#55 DGR-071: Establish upstream pin, patch, and certification maintenance
Opened
#54 DGR-070: Enforce the DeepSeek V4 Flash beta gate
Opened
#53 DGR-069: Prepare narrow upstream llama.cpp collaboration patches