-
Notifications
You must be signed in to change notification settings - Fork 528
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
support of __float2bfloat162_rn #3425
Comments
Can you share some more information, where are you seeing this error. |
I hope this link will eventually let you add all missing functions related to conversions among float, half and bfloat16. https://github.com/ROCm/HIPIFY/blob/amd-staging/docs/tables/CUDA_Device_API_supported_by_HIP.md |
CUDA provided |
// Following math functions included in ROCM6.2 SDK : Here is an simple implementation of
it means accept two floats, and convert each of them to bf16 and pack it into 32bit number. The signature '__float2bfloat162_rn' also implemented in hip, it accepts a single float:
here is the implemenation:
@houseroad @jinz2014 please updated your SDK to rocm 6.2 using this script |
Thanks for the answer. |
error: use of undeclared identifier '__float2bfloat162_rn'; did you mean '__float22bfloat162_rn'?
return __float2bfloat162_rn(a);
Thanks
The text was updated successfully, but these errors were encountered: