This project is read-only.
This page is dedicated to documenting any errata that have been discovered in the Practical Rendering and Computation with Direct3D 11 book. If you find something that should be added here, please post a comment or edit this wiki page!

Resource Format Requirement for UAVs

On pages 63 and 65, it is mentioned that the format parameter of a UAV description must be set to DXGI_FORMAT_R32_UNKNOWN. However, this format does not exist, and should instead be listed as DXGI_FORMAT_UNKNOWN. Thanks to fpmore for pointing this out.

For further information, please see

Last edited Dec 7, 2011 at 9:14 PM by jzink, version 7


Anonymous777 Aug 26, 2013 at 12:22 PM 
Errata: Corrected form P = Paragraph ; [ ] = where mistake occured

159 under Constant Buffers : The same [] recommendation….
396 1st P: the current bone transformation for each bone is [then] calculated….
395 3rd P: ……how much each vertex is influenced by each [of] the objects….
417 1st sentence: ….but it definitely [is] one that had to be handled.
493 next to last P: …. to show [ ] their age.
509 last P: ….. and will [] resut to the render target.
547 last P: Since there [are] eight different…..
559 1st P: … it may lead to invalid data being written [to the] structured buffer.

Nessphoro Nov 26, 2012 at 5:05 AM 
At page 270, section 4.2.3, in the sentence " can be up to 32 float4sor 128 scalars..." there is missing space between "float4s" and "or."

Nessphoro Nov 25, 2012 at 11:32 PM 
At page 228, under Unordered Access View, there seems to be a typo, because the sentence doesn't make any logical sense "...stage can use utilize...", either "use" or "utilize" is extraneous.

kore3d Aug 2, 2012 at 6:10 AM 
At page 111 number of listing is incorrect. The number in the text at the third paragraph is 2.51.

kore3d Jul 30, 2012 at 7:01 PM 
At page 98, "a single trilinear filtered sample will perform 16 memory fetches" is not true. The trilinear filtering uses only 8 fetches per sample (4 fetches per slice).

gtachiki May 15, 2012 at 10:57 AM 
In listing 8.3, it seems like output.light needs to be calculated before output.position is transformed from world space to clip space.