# Stop Paying by the Letter: Accelerating Full-Chip RTL Simulation with Host-Side Tokenized Logging

DevFeed: [Stop Paying by the Letter: Accelerating Full-Chip RTL Simulation with Host-Side Tokenized Logging](<https://devfeed.tech/articles/stop-paying-by-the-letter-accelerating-full-chip-rtl-simulation-with-host-side-tokenized-logging-93695.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.27356>)

Author: Daniel Keller, Rohit Kumar

Published: 2026-10-21T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Chip design](<https://devfeed.tech/topics/chip-design.md>), [Logging](<https://devfeed.tech/topics/logging.md>), [Simulation and Design](<https://devfeed.tech/topics/simulation-and-design.md>), [io interface](<https://devfeed.tech/topics/io-interface.md>), [Encoding](<https://devfeed.tech/topics/encoding.md>)

Tags: [architecture](<https://devfeed.tech/tags/architecture.md>), [bus](<https://devfeed.tech/tags/bus.md>), [chip](<https://devfeed.tech/tags/chip.md>), [cpu-time](<https://devfeed.tech/tags/cpu-time.md>), [cs-ar](<https://devfeed.tech/tags/cs-ar.md>), [cs-pf](<https://devfeed.tech/tags/cs-pf.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [dsp](<https://devfeed.tech/tags/dsp.md>), [elf](<https://devfeed.tech/tags/elf.md>), [firmware](<https://devfeed.tech/tags/firmware.md>), [logging](<https://devfeed.tech/tags/logging.md>), [memory](<https://devfeed.tech/tags/memory.md>), [pairs](<https://devfeed.tech/tags/pairs.md>), [regression](<https://devfeed.tech/tags/regression.md>), [rtl](<https://devfeed.tech/tags/rtl.md>), [saving](<https://devfeed.tech/tags/saving.md>), [simulation](<https://devfeed.tech/tags/simulation.md>), [simulator](<https://devfeed.tech/tags/simulator.md>), [traces](<https://devfeed.tech/tags/traces.md>)

## AI overview

The proposed host-side tokenized logging architecture reduces printf's cost in cycle-accurate RTL simulation without firmware call-site edits. It sends link-time ELF format-string addresses and arguments through a fixed 32-bit slot protocol; a SystemVerilog monitor reconstructs bit-exact strings with zero simulated-time cost. This targets both MMIO bus stalls, which dominate cross-fabric calls, and on-core formatting, which dominates local subsystem calls. Across 858 matched full-chip test pairs, output parity was 100% with zero simulator memory overhead. The authors report 26.18% fewer CPU hours per regression pass, median test CPU time reductions of 41.51% on Suite A and 21.33% on Suite B, plus smaller traces and waveforms. These findings are reported for the evaluated Arm, RISC-V, and DSP systems.

## Source excerpt

arXiv:2610.27356v1 Announce Type: cross Abstract: Firmware diagnostic logging (printf) is an essential debug instrument in full-chip SoC design verification, yet it imposes a heavy tax on cycle-accurate RTL simulation: formatting a 30-character, 1-argument message on a simulated CPU and serializing it through an MMIO aperture costs 3.04 to 15.23 $\mu$s per call (15.27 $\mu$s unweighted / 15.08 $\mu$s print-weighted critical-path mean on Suite A). Unlike embedded tokenizers that target flash footprint and wire bytes with offline decoding, RTL simulation is governed by a two-regime cost function---cross-fabric SoC paths are dominated by MMIO bus stalls (84.7% of call latency) whereas local subsystem paths are dominated by on-core vsprintf formatting (61.7%)---and requires live, zero-simulation-time string reconstruction to drive UVM scoreboards. We present a host-side in-simulation tokenized logging architecture for heterogeneous multi-core SoCs that eliminates both bottlenecks with zero firmware call-site edits. Using link-time ELF format-string virtual addresses as tokens over a fixed 32-bit word-aligned slot protocol across Arm, RISC-V, and DSP cores, on-core logging collapses to a single-pass specifier scan and $(1+S)$ MMIO word writes. This removes all on-core radix conversion ($C_{\mathrm{fmt}} \to 0$), 85.0%--94.5% of deployed bus transactions, 84.6%--93.3% of MMIO bus-stall cycles, and 63.5%--90.9% of total on-core logging cycles ($2.74\times$--$11.03\times$), while a SystemVerilog monitor reconstructs bit-exact strings in zero simulated cycles ($\Delta t_{\mathrm{sim}} = 0$). Across 858 matched full-chip test pairs (1,716 paired runs across 1,972 runs), we achieve 100% output parity with zero simulator memory overhead, saving 1,314.85 CPU hours (-26.18%) per regression pass, cutting median test CPU time by 41.51% (Suite A) and 21.33% (Suite B), and shrinking traces by 42.84% and waveforms by 28.82%.