# Using All the Shared Memory

**URL:** <https://forums.developer.nvidia.com/t/using-all-the-shared-memory/1183>\
**Category:** CUDA Programming and Performance\
**Created:** [July 23, 2007, 6:51am UTC](https://forums.developer.nvidia.com/t/using-all-the-shared-memory/1183 "2007-07-23T06:51:34Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![rgb1380](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@rgb1380](https://forums.developer.nvidia.com/u/rgb1380)\
**Post date:** [July 23, 2007, 6:51am UTC](https://forums.developer.nvidia.com/t/using-all-the-shared-memory/1183/1 "2007-07-23T06:51:34Z")

</div>

Is there a way to make use of all the ‘16384’ bytes of shared memory in the kernel?

Even when the kernel function has no arguments, the compiler still allocates 16 bytes of shared memory, making it impossible to use the exact ‘16384’ bytes.

---

<div class="post-metadata">

**Author:** ![Simon\_Green](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@Simon\_Green](https://forums.developer.nvidia.com/u/Simon_Green)\
**Post date:** [July 23, 2007, 9:42am UTC](https://forums.developer.nvidia.com/t/using-all-the-shared-memory/1183/2 "2007-07-23T09:42:25Z")

</div>

No. Unfortunately a small amount of shared memory is used to pass arguments (thread id etc.) to the kernel.

---

<div class="post-metadata">

**Author:** ![wumpus](https://sea2.discourse-cdn.com/nvidia/user_avatar/forums.developer.nvidia.com/wumpus/32/12040_2.png) [@wumpus](https://forums.developer.nvidia.com/u/wumpus)\
**Post date:** [July 23, 2007, 10:04am UTC](https://forums.developer.nvidia.com/t/using-all-the-shared-memory/1183/3 "2007-07-23T10:04:37Z")

</div>

> [@](#):
>
> No. Unfortunately a small amount of shared memory is used to pass arguments (thread id etc.) to the kernel.
> 
> [snapback]226517[/snapback]

Block id, probably :)