Tune CUDA gaph sizes on B200 and H100 (#3077)
This commit is contained in:
@@ -214,13 +214,7 @@ std::pair<int, int> get_graph_limits(Device& d) {
|
||||
mb = 400;
|
||||
break;
|
||||
case 900: // H100
|
||||
ops = 30;
|
||||
mb = 400;
|
||||
break;
|
||||
case 1000: // B200
|
||||
ops = 50;
|
||||
mb = 500;
|
||||
break;
|
||||
case 1200: // Consumer Blackwell
|
||||
ops = 100;
|
||||
mb = 1000;
|
||||
|
||||
Reference in New Issue
Block a user