perf(evm): u256 arithmetic optimizations (shift/addmod/barrier/value-range/div-mod) - #458

Merged

zoowii merged 9 commits into

DTVMStack:mainfrom

abmcar:perf/u256-batch1

Apr 28, 2026

Commits on Apr 25, 2026

perf(evm): replace udiv/urem with ushr/and in dynamic shift lowering

abmcar
and
claude
committed
perf(evm): inline ADDMOD fast path to eliminate runtime call

abmcar
and
claude
committed
perf(evm): eliminate dead-carry protectUnsafeValue barriers in u64 fast paths

abmcar
and
claude
committed
perf(evm): add value-range narrowing for u256 arithmetic

abmcar
and
claude
committed
perf(evm): inline general u256 DIV/MOD to eliminate runtime call

abmcar
and
claude
committed
fix(evm): fix register coalescer crash in inline DIV/MOD

abmcar
and
claude
committed
refactor(evm): extract u256 LT helper + document ADDMOD fast-path invariant

abmcar
and
claude
committed
test(evm): add ADDMOD fast-path boundary fixtures
abmcar
committed

Commits on Apr 27, 2026

docs(compiler): add change doc for u256 arithmetic optimizations
abmcar
committed