>8.8-17.6gts
still too slow in era of ai, they need to get it over 80gts so cpu can finally be viable for ai inferencing