Dontopedia

GPU decode pipeline

From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)

GPU decode pipeline has 13 facts recorded in Dontopedia across 2 references.

13 facts·12 predicates·2 sources

Mostly:uses metal kernels(1), dispatches kernels per layer(1), has number of layers(1)

Maturity scale raw canonical shape-checked rule-derived certified

Other facts (12)

The long tail: predicates that appear too rarely to warrant their own section. Filter or scroll to find a specific one. Each row links to its source.

12 facts
PredicateValueRef
Uses Metal KernelsMetal Kernel[1]
Dispatches Kernels Per Layer8[1]
Has Number of Layers8[1]
Includes Classifier1[1]
Presupposes Metal FrameworkMetal[1]
Total Dispatches65[1]
Causes Dispatch OverheadDispatch[1]
Rdf:typePipeline[2]
Dispatches Kernel Count8[2]
Has Layer Count8[2]
Has Classifier Count1[2]
Total Dispatch Count65[2]

Timeline

Timeline axis is valid_time — when each source says the fact was true in the world, not when Dontopedia learned about it. Retracted rows are kept for provenance; coloured stripes indicate the context kind.

usesMetalKernelsblah/resources/part-46
ex:metal-kernel
dispatchesKernelsPerLayerblah/resources/part-46
8
hasNumberOfLayersblah/resources/part-46
8
includesClassifierblah/resources/part-46
1
presupposesMetalFrameworkblah/resources/part-46
ex:metal
totalDispatchesblah/resources/part-46
65
causesDispatchOverheadblah/resources/part-46
ex:dispatch
typeblah/resources/46
ex:Pipeline
labelblah/resources/46
GPU decode pipeline
dispatchesKernelCountblah/resources/46
8
hasLayerCountblah/resources/46
8
hasClassifierCountblah/resources/46
1
totalDispatchCountblah/resources/46
65

References (2)

2 references
  1. [1]Part 467 facts
    ctx:discord/blah/resources/part-46
  2. [2]466 facts
    ctx:discord/blah/resources/46
    • full textresources-46
      text/plain3 KBdoc:agent/resources-46/c161ccd3-a97a-462d-a712-6bf0ddc94ec8
      Show excerpt
      [2026-03-03 01:04] xenonfun: going to do a further optimization pass: ``` The GPU decode pipeline dispatches 8 Metal kernels per layer × 8 layers + 1 classifier = 65 dispatches. Each dispatch has overhead (~1-3μs for dispatch + memory barr

See also

Keep researching

Missing something or suspicious of what's here? Kick off a research session — a Claude agent will investigate, cite its sources, and file new facts into a dedicated context you can review before accepting into the shared view.