lock-free design patterns • Define wait-free algorithms, understand the definition and practical implications • An example of an elegant wait-free algorithm and wait-free design • Some simple benchmarks operations are always completing, but individual operations are never guaranteed to ever complete • Wait free (all threads make progress): Every operation completes in a bounded number of steps regardless progress • Progress is not wait free because a particular operation can fail the CAS loop forever because of competing operations succeeding16 Tools for wait freedom • A wait-free algorithm can not contain

precede(task_c); 17 task_b.precede(task_d); 18 task_c.precede(task_d); 19 20 sched.schedule(); 21 sched.wait(); B A C D 1. Easy to write and express a task graph 2. Allow to implement irregular parallel suspend and resume! 16Why Coroutine • Imaging you want to do two things when you go home… wait, wait, and wait… Suppose each thing is a function Suppose you are single 1. Boil the water 2. Take a shower Synchronous Asynchronous Wait Polling Callback Check Do your things Multitasking! 27Taro’s Programming Model – Example 1 Taro: https://github.com/dian-lun-lin/taro A B Callback Wait Polling D C C++

project. Can be overridden for individual tasks. 1.2.5 retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

Features Django Q Documentation, Release 0.7.11 1.2.5 retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

project. Can be overridden for individual tasks. 1.2.5 retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

project. Can be overridden for individual tasks. 1.2.5 retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

your project. Can be overridden for individual tasks. retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

your project. Can be overridden for individual tasks. retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math

project. Can be overridden for individual tasks. 1.2.5 retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: 1.4.

your project. Can be overridden for individual tasks. retry The number of seconds a broker will wait for a cluster to finish a task, before it’s presented again. Only works with brokers that support tests it is as fast or faster than the Redis broker. You can control the amount of time Disque should wait for completion of a task by configuring the retry setting. Bulk task retrieval is supported via the task_result = result(task_id) # result returns None if the task has not been executed yet # you can wait for it task_result = result(task_id, 200) # but in most cases you will want to use a hook: async('math