I have been using CUDAv2.4 and everything has worked perfect. But after upgrading, I have found that many kernels fail to compile. Even a simple kernel: function krnl_foo!(ac,bc) b = threadIdx().x r = blockIdx().x ac[b,r] = 0.0 return nothing end when ex…

Yes, that’s unfortunate, but using Int32 indices is probably worth it since it can result in significant register usage reductions. Besides, for code to work on 32-bit platforms you better make those methods generically-typed anyway, so can’t you update those methods to accept an ::Integer? For st…

Problem with CUDAv3

Specific Domains GPU

maleadt November 8, 2021, 11:38am 4

Those intrinsics now return an Int32.

Topic		Replies	Views
Bug with Julia 1.7.1 and CUDA 3.3 GPU bug , cuda	26	2515	June 2, 2022
Cant compile CUDA v3.3.1 on PowerPC : TypeError in Atomic General Usage cuda	2	330	June 25, 2021
Atomic operations stop working when upgrading to CUDA.jl version 3.3 GPU question	4	1096	June 17, 2021
CUDAnative: simple and nice way to define type for all integer/float literals?; size() returning UInt32? GPU	5	842	July 12, 2019
I32 indexing GPU	8	533	March 24, 2025

Problem with CUDAv3

Related topics