-
Notifications
You must be signed in to change notification settings - Fork 10
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
Add graph capture validation pass #1195
base: main
Are you sure you want to change the base?
Conversation
@@ -450,6 +457,7 @@ def _execute(binaries): | |||
program_index, | |||
total_inputs[loop], | |||
total_outputs[loop], | |||
self["--use-graph-capture"], |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can we add this in trt.runtime.DebugEnv? It seems like this feature is a debug mode feature and the infra is already there to add specific bools like this one
debug_env = ttrt.runtime.DebugEnv.get(
self["--load-kernels-from-disk"], self["--enable-async-ttnn"]
)
const std::unordered_set<uint32_t> &programInputs, | ||
const std::unordered_set<uint32_t> &programOutputs, | ||
::ttnn::MeshDevice *meshDevice, bool useGraphCapture) { | ||
if (useGraphCapture) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
this would be something like debug::Env::get().useGraphCapture
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
changes inlined
Implement a mlir pass which performs graph capture validation of output TTNN graph. By default, this pass is disabled. #1183
fyi @mbezuljTT @s-jovic