[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [csmith-dev] Csmith usage - basic question

Hi Ravi,

1. Does csmith has any standard recommendation for range of seed to used
for effective test case generation . i.e Min range =< seed <= Max range

Csmith uses the rand48 library, I would imagine because its algorithm is well-defined, making seeds portable across platforms. 32 bits of any seed you pass should be useful.

2. How effective does checksum helps to keep a tab on runtime errors ,
can you explain how checksum calculation is happening, so that I can
understand its effectiveness?

It's just a CRC32. It should be very close to 100% effective in detecting errors in global variables.


The only purpose of the checksum is to create a concise summary of a program's execution. If you're at all worried about it being less than perfectly effective, the --check-global option causes the generated code to just print the values of all globals before the program terminates.