Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Already on GitHub? Sign in to your account

array indexing of 8- and 16-bit types #58

Closed
tmcdonell opened this Issue Jul 5, 2012 · 2 comments

Comments

Projects
None yet
1 participant
Member

tmcdonell commented Jul 5, 2012

Indexing an array of 8- or 16-bit type within scalar code yields spurious results:

ghci> let xs = use $ fromList (Z:.10) [0..] :: Acc (Vector Int8)
ghci> CUDA.run $ A.map (\i -> xs ! index1 (A.fromIntegral i)) xs
Array (Z :. 10) [0,4,4,6,4,-123,6,-121,4,69]
Member

tmcdonell commented Dec 20, 2013

I don't think this will manifest when using 2.0 and later cards, because of the way indexing is implemented (texture references vs. array indexing)

@tmcdonell tmcdonell modified the milestone: _|_ Apr 14, 2017

Member

tmcdonell commented Jul 4, 2017

Works with accelerate-llvm-ptx

@tmcdonell tmcdonell closed this Jul 4, 2017

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment