# Per thread buffers: What are they really?

**URL:** <https://forums.percona.com/t/per-thread-buffers-what-are-they-really/493>\
**Category:** Other MySQL® Questions\
**Created:** [October 18, 2007, 12:21pm UTC](https://forums.percona.com/t/per-thread-buffers-what-are-they-really/493 "2007-10-18T12:21:26Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![dwellman](https://avatars.discourse-cdn.com/v4/letter/d/ecc23a/32.png) [@dwellman](https://forums.percona.com/u/dwellman)\
**Post date:** [October 18, 2007, 12:21pm UTC](https://forums.percona.com/t/per-thread-buffers-what-are-they-really/493/1 "2007-10-18T12:21:26Z")

</div>

Basically commenting on [http://www.mysqlperformanceblog.com/2006/05/17/mysql-server-](http://www.mysqlperformanceblog.com/2006/05/17/mysql-server-) memory-usage

When computing maximum memory usage for a thread I have one source that says:

another that says:

yet another:

Regardless, all attest the importance of keeping per thread memory allocation reasonable; hoverver, there is no way to even generically estimate what it could/should be?

---

<div class="post-metadata">

**Author:** ![xxclixxx](https://avatars.discourse-cdn.com/v4/letter/x/67e7ee/32.png) [@xxclixxx](https://forums.percona.com/u/xxclixxx)\
**Post date:** [October 22, 2007, 1:22pm UTC](https://forums.percona.com/t/per-thread-buffers-what-are-they-really/493/2 "2007-10-22T13:22:04Z")

</div>

I’m not sure how it does it, but a tool I found tells you at the top what the max possible usage would be with the settings you have:

[URL][http://rackerhacker.com/mysqltuner/[/URL]](http://rackerhacker.com/mysqltuner/%5B/URL%5D)

---

<div class="post-metadata">

**Author:** ![dwellman](https://avatars.discourse-cdn.com/v4/letter/d/ecc23a/32.png) [@dwellman](https://forums.percona.com/u/dwellman)\
**Post date:** [October 22, 2007, 5:04pm UTC](https://forums.percona.com/t/per-thread-buffers-what-are-they-really/493/3 "2007-10-22T17:04:06Z")

</div>

mysqltuner uses the first method. In full disclosure, that is the very reason why I sought a hard, fast answer. But there doesn’t seem to be one.
