An apparatus and method for dynamic allocation of multiple processor cores in a computer running an operating system includes providing a program operable to halt a core from processing instructions from a respective ready queue. A next step includes establishing a maximum count of tokens available to allow the program execution, wherein the maximum count is less than a total number of cores. A next step includes obtaining tokens by the cores, wherein at least one core will not be able to obtain a token. If a token has been obtained by a core, executing the program by that core, or otherwise not executing the program by that core and remaining active to process instructions from the respective ready queue of that core.