Daniel Lowell

When cmake-get pulls from recipes it pull 1.64 not 1.58

Building with CUDA 9.0

1

Hi I'm building warp-ctc with cuda 9 My machine is 17.04 Ubuntu with pytorch installed. The test_cpu builds ok, but not the test_gpu. Here is a sample error I get:...

Build with CUDA 9.0 failure

2

Hi I'm building warp-ctc with cuda 9 My machine is 17.04 Ubuntu with pytorch installed. The test_cpu builds ok, but not the test_gpu. Here is a sample error I get:...

Currently the manifest for this repo is pointed to branches: ``` ``` Is it possible instead to have these components per ROCm release to have specific commit hashes for the...

v_mac in gfx10 architectures not supported

17

gfx10 does not support the below instructions present in v4r1 and other kernels: ``` In file included from gridwise_convolution_implicit_gemm_v4r1_nchw_kcyx_nkhw_lds_double_buffer.cpp:1: In file included from ./common_header.hpp:22: ./amd_inline_asm.hpp:24:21: error: instruction not supported on...

Dynamic Kernels Assignments

5

### Purpose This project exists to minimize our reliance on compile time parameterization in MIOpen's source kernels. The goal isn't to sacrifice performance, but rather determine a ways of reducing...

Precompiled kernels should contain compiled defaults

10

Need to assess whether this is feasible. Two types of defaults exist: 1. Some solver have hard-coded defaults -- one kernel for everything. 2. Others are configured at runtime using...