[PATCH 4/5] [amdgcn] Update lower limits requested by non-leaf kernels

2019-11-14 Thread Kwok Cheung Yeung
The kernel attributes are changed to request at least 64 SGPRs and 24 VGPRs (i.e. the non-kernel maximum, otherwise the callees may not have enough registers to run in) for non-leaf kernels to take advantage of the reduced number of registers used in non-kernel functions. Okay for trunk? Kwok

Re: [PATCH 4/5] [amdgcn] Update lower limits requested by non-leaf kernels

2019-11-15 Thread Andrew Stubbs
On 14/11/2019 15:33, Kwok Cheung Yeung wrote: The kernel attributes are changed to request at least 64 SGPRs and 24 VGPRs (i.e. the non-kernel maximum, otherwise the callees may not have enough registers to run in) for non-leaf kernels to take advantage of the reduced number of registers used i

Re: [PATCH 4/5] [amdgcn] Update lower limits requested by non-leaf kernels

2019-11-15 Thread Kwok Cheung Yeung
On 15/11/2019 11:32 am, Andrew Stubbs wrote: On 14/11/2019 15:33, Kwok Cheung Yeung wrote: The kernel attributes are changed to request at least 64 SGPRs and 24 VGPRs (i.e. the non-kernel maximum, otherwise the callees may not have enough registers to run in) for non-leaf kernels to take advant

Re: [PATCH 4/5] [amdgcn] Update lower limits requested by non-leaf kernels

2019-11-15 Thread Andrew Stubbs
On 15/11/2019 15:51, Kwok Cheung Yeung wrote: On 15/11/2019 11:32 am, Andrew Stubbs wrote: On 14/11/2019 15:33, Kwok Cheung Yeung wrote: The kernel attributes are changed to request at least 64 SGPRs and 24 VGPRs (i.e. the non-kernel maximum, otherwise the callees may not have enough registers