Track manipulation, locomotion and embodied systems.
Search papers, labs, and topics across Lattice.
Analog in-memory computing (IMC) enables energy-efficient LLM inference, but limited-precision ADC output quantization coupled with input operand rounding creates complex, compounding errors that standard digital clipping methods fail to mitigate. To resolve this, IMC-CLINIC formulates an analytical surrogate of IMC matrix multiplication error that jointly models input operand clipping, accumulated bias, and ADC quantization. The framework delivers 6.5–11.5 percentage point improvements in average zero-shot accuracy over grid-search baselines while speeding up calibration time by 10–12x and operating within 1% of the certified global optimum.
Analog LLM accelerators can claw back over 11 percentage points in zero-shot accuracy while calibrating 12x faster by analytically co-optimizing ADC output quantization alongside weight and activation clipping.
Analog in-memory computing (IMC) offers a promising path toward energy-efficient large language model (LLM) inference by executing matrix multiplications (MatMul) directly within memory arrays in the analog domain. Its efficiency, however, comes with an additional source of error: limited-precision analog-to-digital converters (ADCs) quantize accumulated analog partial sums, introducing output-side error distinct from conventional activation and weight quantization at the MatMul inputs. Clipping can mitigate both operand and ADC quantization errors, but the optimal clipping factors must jointly balance activation rounding and clipping, weight rounding and clipping, and ADC quantization. Existing clipping methods, designed for digital quantization, do not explicitly optimize these coupled sources of IMC error and often rely on costly search-based calibration. We introduce IMC-CLINIC (Coupled Loss-Informed Newton Iterations for Clipping), a clipping calibration framework based on an analytical surrogate for IMC MatMul output error. The surrogate jointly models operand quantization, accumulated clipping-induced bias, and ADC quantization, enabling efficient evaluation of its gradient and approximate curvature from a small calibration set. IMC-CLINIC jointly optimizes activation and weight clipping factors using a safeguarded Newton-type method. Across multiple models and datasets, it improves average zero-shot accuracy by 6.5-11.5 percentage points over the grid search baseline while reducing calibration time by factors of 10.0-12.1. Its analytical surrogate closely tracks empirical IMC output error, and its optimizer is certified within 1% of the global optimum under the loss objective across all projections on two representative models.