>We see 1.3% of Qualcomm's NPU 45 Teraops/s claim To me that suggests that the test is wrong. I could see intel massaging results, but that far off seems incr

>We see 1.3% of Qualcomm's NPU 45 Teraops/s claim To me that suggests that the test is wrong. I could see intel massaging results, but that far off seems incredibly improbable