M4 Pro
From Dontopedia, the open, paraconsistent wiki. (Last updated 2026-06-06.)
M4 Pro has 32 facts recorded in Dontopedia across 10 references, with 1 live disagreement.
32 facts·23 predicates·10 sources·1 in dispute
Mostly:rdf:type(5), cannot benefit from(2), has activity duration(1)
Maturity scale
raw canonical shape-checked rule-derived certifiedInbound mentions (7)
Other subjects in dontopedia point AT this entity as a value. These are inverse relationships — e.g. "X motherOf this subject" — and answer questions the forward facts can't. Grouped by predicate.
fasterThanFaster Than(1)
- M3 Ultra
ex:m3-ultra
hardwareHardware(1)
- Benchmark 1024x512
ex:benchmark-1024x512
hardwarePlatformHardware Platform(1)
- M4 Pro F32 Result
ex:m4-pro-f32-result
outperformsM4ProAt1024SquaredOutperforms M4 Pro At1024 Squared(1)
- M3 Ultra
ex:m3-ultra
projectsHardwareProjects Hardware(1)
- Performance Projection Table
ex:performance-projection-table
shouldPullAheadShould Pull Ahead(1)
- M3 Ultra
ex:m3-ultra
wouldFlipComparisonWould Flip Comparison(1)
- M3 Ultra Parallel
ex:m3-ultra-parallel
Other facts (28)
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.
28 facts
| Predicate | Value | Ref |
|---|---|---|
| Rdf:type | Device | [6] |
| Rdf:type | Hardware Model | [7] |
| Rdf:type | Hardware Processor | [8] |
| Rdf:type | Hardware | [9] |
| Rdf:type | Processor | [10] |
| Cannot Benefit From | Multi Cluster Parallelism | [3] |
| Cannot Benefit From | Multi Cluster Parallelism | [7] |
| Has Activity Duration | 30 days | [1] |
| Has Threads Count | 12 | [2] |
| Superior in Per Unit Speed | M3 Ultra | [3] |
| Has Wider Amx Units | True | [3] |
| Is Faster Per Amx Unit Than | M3 Ultra | [3] |
| Has Newer Microarchitecture | Amx Unit | [3] |
| Has Number of Pclusters | 1 | [3] |
| Is Faster Per Amx Unit | 165 | [3] |
| Duration29 1s | 1024 Squared | [4] |
| Solid Performance | 29.1s | [4] |
| Slower Than M3 Ultra at Same Task | M3 Ultra | [5] |
| Has Duration of Activity | 30 days | [6] |
| Per Amx Unit Performance | ~165 GFLOP/s | [7] |
| Faster Per Amx Unit Than | M3 Ultra | [7] |
| Microarchitecture | newer | [7] |
| Amx Unit Width | wider | [7] |
| Has Property | Rumored 2x Fma Throughput | [7] |
| Comparison Target | M3 | [7] |
| Fused F64 Performance At1 M | 19 | [7] |
| P Cluster Count | 1 | [7] |
| Thread Count | 12 | [9] |
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.
—
hasActivityDurationblah/watt-activation/part-19
30 days
—
hasThreadsCountblah/watt-activation/part-599
12
—
superiorInPerUnitSpeedblah/watt-activation/part-596
ex:m3-ultra
—
hasWiderAmxUnitsblah/watt-activation/part-596
ex:true
—
cannotBenefitFromblah/watt-activation/part-596
ex:multi-cluster-parallelism
—
isFasterPerAmxUnitThanblah/watt-activation/part-596
ex:m3-ultra
—
hasNewerMicroarchitectureblah/watt-activation/part-596
ex:amx-unit
—
hasNumberOfPclustersblah/watt-activation/part-596
1
—
isFasterPerAmxUnitblah/watt-activation/part-596
165
—
duration29-1sblah/watt-activation/part-601
ex:1024-squared
—
solidPerformanceblah/watt-activation/part-601
29.1s
—
slowerThanM3UltraAtSameTaskblah/watt-activation/part-602
ex:m3-ultra
—
typeblah/watt-activation/19
ex:Device
—
labelblah/watt-activation/19
M4 Pro
—
hasDurationOfActivityblah/watt-activation/19
30 days
—
labelblah/watt-activation/593
M4 Pro
—
typeblah/watt-activation/593
ex:HardwareModel
—
perAmxUnitPerformanceblah/watt-activation/593
~165 GFLOP/s
—
fasterPerAmxUnitThanblah/watt-activation/593
ex:m3-ultra
—
microarchitectureblah/watt-activation/593
newer
—
amxUnitWidthblah/watt-activation/593
wider
—
hasPropertyblah/watt-activation/593
ex:rumored-2x-fma-throughput
—
comparisonTargetblah/watt-activation/593
ex:m3
—
fusedF64PerformanceAt1Mblah/watt-activation/593
19
—
pClusterCountblah/watt-activation/593
1
—
cannotBenefitFromblah/watt-activation/593
ex:multi-cluster-parallelism
—
typeblah/watt-activation/594
ex:HardwareProcessor
—
labelblah/watt-activation/594
M4 Pro
—
typeblah/watt-activation/596
ex:Hardware
—
labelblah/watt-activation/596
M4 Pro
—
threadCountblah/watt-activation/596
12
—
typeblah/watt-activation/599
ex:Processor
References (10)
10 references
ctx:discord/blah/watt-activation/part-19ctx:discord/blah/watt-activation/part-599ctx:discord/blah/watt-activation/part-596ctx:discord/blah/watt-activation/part-601ctx:discord/blah/watt-activation/part-602ctx:discord/blah/watt-activation/19- full textwatt-activation-19text/plain2 KB
doc:agent/watt-activation-19/e74bc25c-aab8-43ac-90e0-2f036b5a9627Show excerpt
[2026-03-05 22:21] xenonfun: Both started from the same checkpoint, so same baseline: - Start checkpoint ./philosophy_model_fresh/checkpoint_iter_9198.npz - Baseline on same eval slice/settings: val_loss=5.355859, val_ppl=211.85 So …
ctx:discord/blah/watt-activation/593- full textwatt-activation-593text/plain3 KB
doc:agent/watt-activation-593/5c383b0f-5e7c-4f28-b8e6-6c00caa3ef98Show excerpt
[2026-04-04 17:10] xenonfun: ⏺ Tiled is a massive win. Now let me wire it into the production code and commit. ``` ┌────────┬─────────────┬─────────────┬─────────────┐ │ N │ Fused f32 │ Tiled f32 │ Improvement │ ├────────┼───…
ctx:discord/blah/watt-activation/594- full textwatt-activation-594text/plain3 KB
doc:agent/watt-activation-594/d9974ad5-bce6-44fb-a387-861b63507258Show excerpt
[2026-04-04 17:21] xenonfun: ``` The headline numbers ┌──────────────────────┬─────────────────┬───────────────┬─────────┐ │ │ Scalar (before) │ Par tiled AMX │ Speedup │ ├──────────────────────┼───────────────…
ctx:discord/blah/watt-activation/596- full textwatt-activation-596text/plain3 KB
doc:agent/watt-activation-596/83561bf3-adfc-4a74-8293-71c29d442bdfShow excerpt
[2026-04-04 18:11] xenonfun: ⏺ scalar_product went from 10,789 µs to 142 µs — 76× speedup. The precomputed scalar pairs reduced ~1024 iterations to ~32. Overall Metropolis per-site: 4.8× speedup, from 21 µs/site to 4.3 µs/site. For a …
ctx:discord/blah/watt-activation/599
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.