jinz2014

Results 57 comments of jinz2014

Extension to cg scan

Thanks.

no viable conversion from 'typename AccessorCommonT::AccessorSubscript<2 - 1, local_accessor_base<unsigned int, 2, (mode)1026, (placeholder)0>>' to 'volatile uint *'

Sorry, the header file (EGS.h) was missing in the repository.

no viable conversion from 'typename AccessorCommonT::AccessorSubscript<2 - 1, local_accessor_base<unsigned int, 2, (mode)1026, (placeholder)0>>' to 'volatile uint *'

Please let me know the building issue so I will update the file. No problem

CURAND_RNG_PSEUDO_MRG32K3A in curand_mtgp32_uniform_example.cpp

The example name is [curand_mtgp32_uniform_example.cpp] Sorry, I didn't describe it clearly.

The SYCL sub-group does not support mask options for sycl::reduce_over_group.

https://github.com/readwrite112/AGAThA/tree/main/AGAThA/src Attached paper [3627535.3638474.pdf](https://github.com/oneapi-src/SYCLomatic/files/14574750/3627535.3638474.pdf) @readwrite112

Implement `local_accessor::get_multi_ptr`

Could you please explain the reason of change from get_pointer() to get_multi_ptr in SYCL 2020?

Codes not migrated

I added a CUDA example (https://github.com/zjin-lcf/HeCBench/tree/master/warpselect-cuda) for your reference. Then, I changed the generated DPCT codes manually for the shuffle functions and the copy constructor of the Tensor class. However,...

Codes not migrated

@lwqq00 I reopened the issue.

Codes not migrated

@lwqq00 I would try to change the cuda code

Codes not migrated

@tomflinda Users may or may not change CUDA codes. On the other hand, may you please suggest code changes needed in the SYCL program to work around the trivially copyable...