[Inductor] Constant folding support

[jgong5]

Motivating Example

Below is a case in MobileBertForMaskedLM which has a concatenation on two model parameters (hidden_states = hidden_states.matmul(torch.cat([self.decoder.weight.t(), self.dense.weight], dim=0))). This concat takes >20% of the single-threaded inference time on CPU but this cost can be saved with constant folding.

class MobileBertLMPredictionHead(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.transform = MobileBertPredictionHeadTransform(config)
        # The output weights are the same as the input embeddings, but there is
        # an output-only bias for each token.
        self.dense = nn.Linear(config.vocab_size, config.hidden_size - config.embedding_size, bias=False)
        self.decoder = nn.Linear(config.embedding_size, config.vocab_size, bias=False)
        self.bias = nn.Parameter(torch.zeros(config.vocab_size))
        # Need a link between the two variables so that the bias is correctly resized with `resize_token_embeddings`
        self.decoder.bias = self.bias

    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
        hidden_states = self.transform(hidden_states)
        hidden_states = hidden_states.matmul(torch.cat([self.decoder.weight.t(), self.dense.weight], dim=0))
        hidden_states += self.decoder.bias
        return hidden_states

cc @mlazos @soumith @voznesenskym @yanboliang @penguinwu @anijain2305 @EikanWang @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @Xia-Weiwen @wenzhe-nrv @jiayisunx @peterbell10 @desertfire @ezyang @msaroufim @wconstab @ngimel @bdhirsh

[ydwu4]

Hi, is there any update on this? Do we still want to keep it open?

[jgong5]

@Le-Zheng can you check if this issue has been addressed with freezing on?

[leslie-fang-intel]

I think constant folding should already be supported in fw_compiler_freezing. @Le-Zheng Could you help to double confirm this issue fixing by adding TORCHINDUCTOR_FREEZING=1?