GPU Float64 memset support

Experimenting with Mojo’s GPU package:

from gpu.host import DeviceContext
from gpu.host import DeviceBuffer

fn main() raises:
    with DeviceContext() as ctx:
        var d_buf = ctx.create_buffer_sync[DType.float64](1024)
        ctx.memset_sync[DType.float64](d_buf, 0)

I get this error:

ctx.memset_sync[DType.float64](d_buf, 0)
                                      ^
Kernels/mojo/gpu/host/device_context.mojo:1622:8: note: function instantiation failed
Kernels/mojo/gpu/host/device_context.mojo:1639:10: note: call expansion failed
open-source/mojo/stdlib/stdlib/builtin/constrained.mojo:56:6: note: constraint failed: bitwidth of memset type must be one of [8,16,32]
mojo: error: failed to run the pass manager

Is this a matter of Float64s being unsupported, or is there another way of achieving this?