Age | Commit message (Collapse) | Author |
|
|
|
|
|
capture other parts of the kernel code
|
|
approach for convgemm
|
|
|
|
local memory support now
|
|
edge cases now
|
|
gemm kernel
|
|
a new kernel
|
|
|
|
|
|
|
|
|
|
|
|
|
|
compiler
|
|
|
|
|
|
|
|
|
|
|
|
where the AMD compiler crashes"
This reverts commit 407ed52cec41445f02e85cb45d08f590960216bb.
|
|
|
|
|
|
AMD compiler crashes
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
names in kernels
|
|
|
|
|
|
loops in kernel accordingly
|
|
|
|
|
|
pragma for several kernels
|
|
based on assumptions
|
|
|
|
|
|
|
|
|
|
changed transpose kernel code
|
|
|
|
|
|
|
|
|
|
optional temporary buffers
|