On Wed, Sep 8, 2021 at 10:54 AM Hongtao Liu wrote:
>
> On Wed, Jul 21, 2021 at 3:43 PM liuhongt wrote:
> >
> > Hi:
> > As discussed in [1], this patch support _Float16 under target sse2
> > and above, w/o avx512fp16, _Float16 type is storage only, all operations
> > are emulated by soft-fp and
On Wed, Jul 21, 2021 at 3:43 PM liuhongt wrote:
>
> Hi:
> As discussed in [1], this patch support _Float16 under target sse2
> and above, w/o avx512fp16, _Float16 type is storage only, all operations
> are emulated by soft-fp and float instructions. Soft-fp keeps the intermediate
> result of the
Hi:
As discussed in [1], this patch support _Float16 under target sse2
and above, w/o avx512fp16, _Float16 type is storage only, all operations
are emulated by soft-fp and float instructions. Soft-fp keeps the intermediate
result of the operation at 32-bit precision by defaults, which may lead to
