Is P = 32 being smaller than P = 16 at equal cycles a real effect or a synthesis artefact?

research/chip-macp32cells

Re-synthesises the 16- and 32-lane MAC engines (same cycles) with three seeds and three Yosys flows: the sign of the cell-count gap depends on the flow, with swings of 1-10%, so chip-systolic-p32's 0.7% ranking is not reliable (artefact-sensitive, stylised, generic-gate counts).

4 tests

What it found

chip-systolic-p32 found the 32-lane MAC engine smaller than the 16-lane one although both take the same cycles. Re-synthesised with three model seeds and three Yosys/abc flows at F = 16 and 32, the sign depends on the flow: P = 32 is smaller in all 6 cases under two flows and larger in all 6 under a four-gate library, by 47 to 699 cells (1–10%). That swing is far above the 0.7% margin used to rank the best design earlier, so that ranking should be discarded. Three seeds, one tool version, generic-gate counts only; the mechanism is not traced.

Paper, code and results

More in LLM chip design loops

At which Yosys stage does the P = 32 versus P = 16 cell gap appear?Counts cells of the 16- and 32-lane MAC engines after the word-level, bit-level and abc stages: the 32-lane engine is bigger at the word level, about 500 cells smaller after techmap, so the chip-macp32cells gap is set before abc (localised, mechanism unexplained, stylised).
Which Yosys pass flips the P = 32 versus P = 16 cell gap?Runs the Yosys fine passes one at a time on the 16- and 32-lane MAC engines: the P = 32 versus P = 16 sign flips (+2,100 to about -120 cells) at the first opt -fast after techmap, before any abc, and chip-p32netlist's pre-abc label is corrected (Yosys 0.33's -run :abc includes abc -fast). Stylised.
How few bits can a MAC lane use to classify real handwritten digits before accuracy falls off a cliff?Weight and activation bit-width sweep of a 64-16-10 quantised MLP on the scikit-learn digits data (5 splits), with the MAC lane's gate count from Yosys and its RTL checked against NumPy: 4-bit weights and activations lose about 2 points of accuracy, 5 and 5 about 0.5, below 3 bits it collapses.
Can per-channel power-of-two scales make the per-neuron quantiser an integer-exact MAC lane, and what accuracy does it keep at 2 to 4 weight bits?Giving each neuron of the digits MLP its own power-of-two weight scale turns the per-channel rescale into one or two shift stages and keeps the lane integer-exact: at W2A4 the MSE-chosen exponents reach 70.5% (against 62% for per-tensor clipping), but at W3A4 they give 87.9%, below clipping's 90.8%.
Does clipping, per-channel scaling or quantisation-aware training move the 3-bit cliff of an MLP lane on real digits?MSE-optimal weight clipping lifts the 3-bit-weight digits MLP from 81% to 91% test accuracy at the same MAC-lane hardware and integer-exact; per-channel scales and QAT help at 4 bits (95.7% at W4A4, within 1.1 points of float) but QAT fails at 2 bits.
What does an all-open RTL flow report for small designs, and can it prove them equivalent?Reproducible open RTL flow (Yosys, Icarus, Verilator): gate count and depth for adders, a multiplier and a FIFO, SAT equivalence of two adders plus a broken-adder control. Gate counts, not silicon.

All 40 LLM chip design loops projects →

Other research topics

Proper scoring & elicitation28 Effort & peer prediction9 Verification games & slashing19 Detection & tolerances14 Markets as routers13 Prediction-market design9 Paying contributors8 Decentralised training systems23 Digital twins & sim-to-real164 LLM chip design loops40 Robot manufacturing loops53 Ambient AI assistant46 Self-replicating manufacturing34 Simulation-first design64 Engineered clothing45 Task-specific AI accelerators40 Algorithmic chip-design loops42 Open-source purpose-built robots44 The software-engineering loop43 Adversarial human imitation41 Open-source residential construction41 Autonomous multi-angle capture34 Sensing, imaging and 3D31 Energy supply and dense energy storage39 Robot societies among people31 Agentic co-design of low-cost robots27 Laundry-hanging robots25 Machine learning on qualitative data23 Robots on the building site27 Civilizations & group psychology1 History & human sciences1

Produced with AI assistance (Claude). Results are from the code in this repository and are simulated or stylised unless the abstract says otherwise; nothing here implies affiliation with or endorsement by any lab, researcher or company cited. MIT licensed.