jha-lab / acceltran

[TCAD'23] AccelTran: A Sparsity-Aware Accelerator for Transformers
BSD 3-Clause "New" or "Revised" License
33 stars 8 forks source link

question about top.sv and PE #6

Closed TT-RAY closed 1 year ago

TT-RAY commented 1 year ago

I found that you did delete some CNN module codes, but those modules are still called in synthesis/top/top.sv, and the code of the transformer accelerator PE module is not provided.

shikhartuli commented 1 year ago

I have removed top.sv as well. We do not provide the code for AccelTran's PE module. Instead, we use the reports of the sub-modules (MACs, softmax, and layer-norm) for calculations.