Latest activity

Hi @Mountain, replied to you by email.

What's New: Added new NPU power consumption test. Updated TFLite runtime. Updated TFLite GPU, NNAPI, Qualcomm QNN, Hexagon NN and...

Happy to help! And yes of course, let's connect. I'll send a DM. Cheers, Manuel

Happy to help! And yes of course, let's connect. I'll send a DM. Cheers, Manuel

hi， The current AndroidT version and release, there is a new nnapi operator, will aibenchmark update and adapt to this?

This is a very brief answer, but the general idea is as follows: HTP = rebranded compute DSP (since Snapdragon 888 / Hexagon v68)...

This is a very brief answer, but the general idea is as follows: HTP = rebranded compute DSP (since Snapdragon 888 / Hexagon v68)...

This is a very brief answer, but the general idea is as follows: HTP = rebranded compute DSP (since Snapdragon 888 / Hexagon v68)...

Thanks for the quick reply. Are you familiar with the internal components of the HTP? Eg the difference between the HTA, HVX and HMX...

Thanks for the quick reply. Are you familiar with the internal components of the HTP? Eg the difference between the HTA, HVX and HMX...

I recently got an S23 Ultra and ran the benchmark. The score was 2031K which does blow away anything else on the device list so far. You...

One can potentially extract all models directly from the benchmark APK file. Feel free to use this forum for sharing or comparing the...

Yes, partly: the Hexagon 6xx family is denoted as DSPs in QNN, while the Hexagon 7xx family - as HTPs. Here you can find the full list...

INT8 models were running with the TFLite GPU delegate. No, these are plain NPU/GPU runtime results. Because of the bug in the iOS...

In the standard benchmark mode, only INT8 inference is tested. However, one can also check the results of FP16 inference in the PRO...