"vscode:/vscode.git/clone" did not exist on "892a819da4e67be4b94aad4e27921db71c365b85"
Commit 84d0f5c9 authored by Paul's avatar Paul
Browse files

Divide by vec.size

parent 047162bb
...@@ -61,7 +61,7 @@ struct layernorm_compiler : compiler<layernorm_compiler> ...@@ -61,7 +61,7 @@ struct layernorm_compiler : compiler<layernorm_compiler>
} }
auto preloads = preload::broadcasts(axis, inputs); auto preloads = preload::broadcasts(axis, inputs);
auto relements = inputs[0].lens()[axis] / vec.size; auto relements = inputs[0].lens()[axis] / vec.size;
auto nelements = inputs.back().elements() / inputs[0].lens()[axis]; auto nelements = (inputs.back().elements() / inputs[0].lens()[axis]) / vec.size;
auto block_size = compute_block_size(relements, 256); auto block_size = compute_block_size(relements, 256);
hip_compile_options options; hip_compile_options options;
options.set_launch_params( options.set_launch_params(
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment