CharlesKWON
f9ffd4050e
Two stacked optimizations land on the SqlScan hot path. Combined
effect on the 50k-row benchmark:
Before After vs raw
Numeric WHERE 10.2ms 7.8ms 1.15x
String WHERE 10.5ms 7.9ms 1.15x
No WHERE 9.2ms 10.0ms 1.45x
Raw RDD baseline 6.8ms 6.8ms 1.00x
WHERE-predicate paths are now within 15% of the raw Harbour-style
RDD scan loop. The no-WHERE path is unchanged (slight jitter from
the added devirt branch); FieldGet peephole doesn't apply there.
--- Optimization 1: PcOpFieldGet peephole ---
Adds a new pcode opcode `PcOpFieldGet <fieldIdx>` (0x46) that skips
the usual PushSymbol+Function+Frame+FieldGet-RTL+EndProc chain and
calls a direct field getter closure instead. genpc recognizes the
shape `FieldGet(<int-literal>)` during emitCall and emits the
specialized opcode automatically — no SQL-side API change.
Integration:
* hbrt.Thread.FastFieldGetter — hot-path closure set by scan loops.
Non-nil → pcode bypasses dispatch.
Nil → pcode resolves FIELDGET via
the RTL symbol table (correctness
fallback for any other callers).
* compiler/genpc/genpc.go — peephole in emitCall.
* hbrt/pcinterp.go — PcOpFieldGet handler.
This alone cut numeric WHERE from 10.2 → 7.9ms: eliminated roughly
one full Frame/EndProc + RTL dispatch per row × 50k rows.
--- Optimization 2: DBFArea devirtualization ---
SqlScan type-asserts the workarea to *dbf.DBFArea once and runs a
dedicated loop that calls GoTop/EOF/Skip/GetValue directly on the
concrete type. Go's compiler inlines these, skipping the interface
vtable per row. Non-DBF drivers still work via the generic Area
branch.
The FastFieldGetter closure also captures *DBFArea directly in the
DBF branch, so the WHERE predicate side of the hot loop is now
entirely devirtualized: no interface dispatch between the pcode
dispatch loop and the DBF record buffer.
Validation:
- FiveSql2 43/43
- Harbour compat 51/51
- go test ./... ALL PASS
Remaining gap to raw RDD on no-WHERE (~1.45x) is dominated by the
two-column row construction + ArraySlab + flat backing bookkeeping
that the raw loop doesn't do. Going below that requires changing
the SQL engine's result shape — out of scope here.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
120 lines
3.3 KiB
Go
120 lines
3.3 KiB
Go
// Copyright (c) 2026 Charles KWON OhJun (charleskwonohjun@gmail.com)
|
|
// All rights reserved.
|
|
|
|
// Five pcode — stack-based bytecode for FRB interpreter mode.
|
|
// Each opcode maps 1:1 to a Thread method call, making the pcode
|
|
// a direct serialization of what gengo generates as Go code.
|
|
//
|
|
// Format: [opcode:1byte] [operands:variable]
|
|
// Strings: [len:uint16 LE] [bytes]
|
|
// Numbers: int64 = 8 bytes LE, float64 = 8 bytes LE
|
|
|
|
package hbrt
|
|
|
|
// Opcode definitions
|
|
const (
|
|
// Stack operations
|
|
PcOpNop byte = 0x00
|
|
PcOpPushNil byte = 0x01
|
|
PcOpPushTrue byte = 0x02
|
|
PcOpPushFalse byte = 0x03
|
|
PcOpPushInt byte = 0x04 // + int64 LE
|
|
PcOpPushDouble byte = 0x05 // + float64 LE (8 bytes)
|
|
PcOpPushString byte = 0x06 // + uint16 len + bytes
|
|
PcOpPushLocal byte = 0x07 // + uint16 index
|
|
PcOpPopLocal byte = 0x08 // + uint16 index
|
|
PcOpPop byte = 0x09
|
|
PcOpDup byte = 0x0A
|
|
|
|
// Arithmetic
|
|
PcOpPlus byte = 0x10
|
|
PcOpMinus byte = 0x11
|
|
PcOpMult byte = 0x12
|
|
PcOpDivide byte = 0x13
|
|
PcOpMod byte = 0x14
|
|
PcOpPower byte = 0x15
|
|
PcOpNegate byte = 0x16
|
|
|
|
// Comparison
|
|
PcOpEqual byte = 0x20
|
|
PcOpNotEqual byte = 0x21
|
|
PcOpLess byte = 0x22
|
|
PcOpGreater byte = 0x23
|
|
PcOpLessEq byte = 0x24
|
|
PcOpGreaterEq byte = 0x25
|
|
PcOpInString byte = 0x26
|
|
|
|
// Logical
|
|
PcOpAnd byte = 0x28
|
|
PcOpOr byte = 0x29
|
|
PcOpNot byte = 0x2A
|
|
|
|
// String
|
|
PcOpConcat byte = 0x2C // same as Plus for strings
|
|
|
|
// Flow control
|
|
PcOpJump byte = 0x30 // + int32 LE (relative offset)
|
|
PcOpJumpFalse byte = 0x31 // + int32 LE
|
|
PcOpJumpTrue byte = 0x32 // + int32 LE
|
|
PcOpReturn byte = 0x33
|
|
PcOpRetValue byte = 0x34
|
|
|
|
// Frame
|
|
PcOpFrame byte = 0x38 // + uint16 params + uint16 locals
|
|
PcOpEndProc byte = 0x39
|
|
|
|
// Function calls
|
|
PcOpPushSymbol byte = 0x40 // + uint16 string len + name
|
|
PcOpPushNilArg byte = 0x41 // push NIL for function self
|
|
PcOpFunction byte = 0x42 // + uint16 nArgs
|
|
PcOpDo byte = 0x43 // + uint16 nArgs
|
|
|
|
// Workarea field access — skips PushSymbol + Function dispatch
|
|
// for `FieldGet(n)` where n is a literal. Emitted by genpc as a
|
|
// peephole optimization. Operand: uint16 1-based field position.
|
|
PcOpFieldGet byte = 0x46
|
|
|
|
// Self / OOP
|
|
PcOpPushSelf byte = 0x48
|
|
PcOpPushSelfField byte = 0x49 // + uint16 len + name
|
|
PcOpSetSelfField byte = 0x4A // + uint16 len + name
|
|
PcOpSend byte = 0x4B // + uint16 len + name + uint16 nArgs
|
|
|
|
// Array / Hash
|
|
PcOpArrayGen byte = 0x50 // + uint16 count
|
|
PcOpHashGen byte = 0x51 // + uint16 count
|
|
PcOpArrayPush byte = 0x52
|
|
PcOpArrayPop byte = 0x53
|
|
|
|
// Block
|
|
PcOpPushBlock byte = 0x58 // + uint32 codeLen + pcode bytes + uint16 nDetached
|
|
|
|
// Local operations
|
|
PcOpLocalAddInt byte = 0x60 // + uint16 index + int32 value
|
|
PcOpInc byte = 0x61
|
|
PcOpDec byte = 0x62
|
|
|
|
// Special
|
|
PcOpPopLogical byte = 0x70 // pop and store logical result
|
|
PcOpPushBool byte = 0x71 // + 1 byte (0 or 1)
|
|
|
|
// Line info (for debugging)
|
|
PcOpLine byte = 0xFE // + uint16 lineNo
|
|
PcOpHalt byte = 0xFF
|
|
)
|
|
|
|
// PcodeFunc represents a pcode-compiled function.
|
|
type PcodeFunc struct {
|
|
Name string
|
|
Code []byte // bytecode
|
|
Params int // number of parameters
|
|
Locals int // number of locals
|
|
}
|
|
|
|
// PcodeModule represents a compiled pcode module (multiple functions).
|
|
type PcodeModule struct {
|
|
Name string
|
|
Funcs map[string]*PcodeFunc
|
|
Strings []string // string constant pool
|
|
}
|