This PR applies the performance optimizations from PR #9195 to RlpStream and Rlp.ValueDecoderContext classes, and introduces shared helper code to eliminate duplication between all three RLP processing implementations.

Key Optimizations Applied

1. Branchless Prefix Length Calculation

Replaced conditional logic with bit manipulation for calculating RLP prefix lengths:

// Before: Multiple if/else branches
if (prefix <= 128) return 0;
else if (prefix <= 183) return 1;
// ... more branches

// After: Branchless bit manipulation
uint p = prefixByte;
uint v = p >> 3;
uint r = v >> 4;
// ... calculated without branches

2. Switch Expression Optimizations

Converted if/else chains to switch expressions for better compiler optimization:

// Before: if/else chain in PeekNextRlpLength()
if (prefix <= 128) return 1;
else if (prefix <= 183) return 1 + prefix - 0x80;
// ... more conditions

// After: Switch expression
return prefix switch
{
    <= 128 => 1,
    <= 183 => 1 + prefix - 0x80,
    < 192 => PeekLongStringRlpLength(prefix),
    <= 247 => 1 + prefix - 0xc0,
    _ => PeekLongListRlpLength(prefix),
};

3. Optimized Method Separation

Split complex logic into specialized methods marked with [MethodImpl(MethodImplOptions.NoInlining)] for better cache locality and JIT optimization of hot paths.

4. Shared Helper Code

Created RlpHelpers class to eliminate code duplication:

• CalculatePrefixLength() - Branchless prefix calculation
• DeserializeLengthRef() - Optimized length deserialization
• Shared exception throwing methods

5. Improved Bounds Checking

Replaced range checks with more efficient unsigned comparisons:

// Before: lengthOfLength > 4
// After: (uint)lengthOfLength > 4

6. Optimized SkipItem Implementation

Simplified item skipping to use single method call instead of tuple destructuring:

// Before: 
(int prefix, int content) = PeekPrefixAndContentLength();
SkipBytes(prefix + content);

// After:
SkipBytes(PeekNextRlpLength());

Impact

• Performance: Eliminates branching in hot paths, improves CPU cache utilization
• Code Quality: Reduces duplication by ~50 lines across the three classes
• Maintainability: Centralized shared logic in RlpHelpers
• Compatibility: Zero breaking changes - all existing functionality preserved

Verification

• All 2,299 core unit tests continue to pass
• Custom test harness verifies consistency between all three implementations
• Tested both short string (<56 bytes) and long string (≥56 bytes) encoding paths

The optimizations maintain identical behavior while providing better performance characteristics for RLP processing, which is critical for Ethereum block and transaction parsing.

Fixes #9207.

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.