A100 Accelwattch XML Configuration

[linuohai]

Hi,

I am trying to use AccelWattch to model the power consumption of an NVIDIA A100 GPU. However, I noticed that the provided XML configuration files appear to be identical, and based on the AccelWattch paper, they seem to be derived from measurements on the V100.

I would like to ask whether there is an existing XML configuration file specifically calibrated for the A100. If not, is it acceptable to use the current V100-based XML configuration for A100 power modeling? Additionally, is there any recommended scaling methodology or adjustment procedure for adapting the XML parameters from V100 to A100 (for example, based on frequency differences, SM count, or process technology)?

Any guidance or references would be greatly appreciated. Thank you!