官方的 MASTR 标志 MASTR
菜单

Solana:从报价到结算

Solana priority fees: the compute limit is part of the price

Separate the base network fee, prioritisation, Jito tips and application charges.

更新于 2026 年 9 月 30 日 · MASTR Labs

文章正文和技术图表为英语,导航提供 7 种语言。

参考指南
  1. Compute is a budget
  2. Be specific about transaction versions
  3. Why a larger budget can waste money
  4. How to read a trade ticket
  5. 具体示例
  6. 来源与原始资料

Compute is a budget

Solana measures execution work in compute units. A transaction’s compute limit caps the work it can request; the compute-unit price expresses a prioritisation bid. In the compute-budget model used by legacy and v0 messages, the prioritisation fee is the requested limit multiplied by the micro-lamport price, divided by 1,000,000 and rounded up to lamports. It is not simply a bill for the units ultimately consumed. 1

Be specific about transaction versions

Solana’s current fee documentation distinguishes the V1 format, where the prioritisation fee is an absolute lamport amount in message configuration. A reader should therefore verify the transaction format before applying a formula copied from an older tutorial. The base network fee, prioritisation fee and application-level charges are different line items. A Jito tip is another distinct payment, not another name for compute pricing.

Why a larger budget can waste money

In the compute-unit-price model, increasing the requested limit increases the prioritisation fee at an unchanged unit price. Too small a limit can cause execution to fail; too large a limit can pay for headroom the transaction never uses. Simulation can inform an estimate, but it is a preview under observed conditions rather than a guarantee of future execution.

How to read a trade ticket

Compare the transaction’s expected output with the complete cost breakdown. Distinguish fees from token-account deposits that may be recoverable, and from token-specific transfer fees. A higher priority setting can improve inclusion chances without improving the swap’s exchange rate or guaranteeing success. Interfaces should show the final amounts before authorisation, not ask users to infer them from labels such as ‘Turbo’.

Read each cost separately
解释性图表。点击查看完整尺寸。 Credits ↗ Legacy/v0 example: 200,000 × 1,000 / 1,000,000 = 200 lamports.

具体示例

Under the compute-unit-price model, a 200,000-CU limit at 1,000 micro-lamports per CU costs 200 lamports in prioritisation fees. A 400,000-CU limit at the same price costs 400 lamports, even if execution only uses 180,000 CU. Base and other costs are separate.

来源与原始资料

  1. Solana: fee structure and version differences
  2. Solana: transactions

继续阅读

Solana:从报价到结算 →

MASTR

支持独立研究

这里的调查、原始证据和指南均可免费阅读。自愿捐赠帮助支付研究成本,让 MASTR 能够继续提供工具。

打开钱包