Skip to content

Decimal

The Decimal<Tag, Scale, TickSize> template represents a fixed-point decimal number with compile-time scaling and optional tick size rounding. It is used to represent precise quantities like price or volume in trading systems, avoiding floating-point errors.

Template Parameters

template <typename Tag, int Scale, int64_t TickSize = 1>
class Decimal;
  • Tag: Empty struct type used to distinguish different Decimal types (e.g. PriceTag, VolumeTag)
  • Scale: Integer scale factor (e.g. 1_000_000 for 6 decimal places)
  • TickSize: Optional tick size used for rounding and discretization

Purpose

  • Provide a safe and precise alternative to double in financial applications
  • Ensure that prices and quantities are stored as integers internally
  • Avoid runtime ambiguity between different uses of decimals (e.g., price vs quantity)

Interface Summary

static constexpr Decimal fromDouble(double val);
static constexpr Decimal fromRaw(int64_t raw);
constexpr double toDouble() const;
constexpr int64_t raw() const;
constexpr Decimal roundToTick() const;

constexpr bool isZero() const;

constexpr Decimal operator+(Decimal d) const;
constexpr Decimal operator-(Decimal d) const;
constexpr Decimal &operator+=(const Decimal &other);
constexpr Decimal &operator-=(const Decimal &other);
constexpr auto operator<=>(const Decimal &) const = default;

Behavior

  • Internally stores the value as a raw integer (int64_t _raw)
  • fromDouble() converts a double to internal integer by multiplying with Scale and rounding
  • toDouble() converts back to double for display or logging
  • roundToTick() aligns the internal value to the nearest multiple of TickSize
  • isZero() checks if the internal value is exactly zero

Notes

  • This type is constexpr-friendly and usable in low-latency paths
  • The Tag type is not used at runtime, but enforces type-safety at compile time
  • All operators and conversions are explicit to avoid accidental misuse

Example Use

struct PriceTag {};
struct VolumeTag {};

using Price = Decimal<PriceTag, 1'000_000>;
using Volume = Decimal<VolumeTag, 1'000_000>;

Price p = Price::fromDouble(12345.67);
Volume v = Volume::fromDouble(0.005);

double pv = p.toDouble();
int64_t raw = v.raw();