benchmark code

基本解釋

Benchmark code doesn't look like real code
基準代碼看上去并不像實際中的代碼
Even within this limited definition of performance, pitfalls abound in trying to benchmark code accurately.
即使在如此狹義的性能定義之下，精確地進行代碼基準測試仍然有很多困難。
In fact, if we go look at the benchmark code, we will see that the table is created, populated, and used, but statistics are never updated.
事實上，如果我們看一下基準代碼，我們就會發(fā)現(xiàn)，該表已被創(chuàng)建、填充、并使用，但是統(tǒng)計信息沒有被更新。

So what can we say about the internal revenue code using this neutral benchmark?
FORBES: The 'Tax Expenditure' Con Job
He suspects some of the listings were submitted by vendors who ran Linpack (the standard code used to measure supercomputing performance) benchmark before delivering the systems to their customers and then submitted those systems for the list without telling their customers.
FORBES: Supercomputer Manages Fixed Income Risk At JPMorgan