Purge dim name and distutils

[calad0i]

Description

Get rid of dim names in the whole repository, based on the DA PR.

Type of Change

• Cleaning up

Tests

N/A

Checklist

• all

[vloncar]

Now you're adding unrelated commits to this PR 😄

[calad0i]

Now you're adding unrelated commits to this PR 😄

Let me move the distutil one to separate PR? It is deprecated and breaks on python 312 and 313.

[JanFSchulte]

I think that was about the change to the bit-exact algorithm which probably needs to go the DA PR :-D

[jmitrevs]

I think test_rnn.py fails due to changes in this PR. Otherwise, I think getting this in would be very good.

[JanFSchulte]

The pytest error is fixed now, but there seems to be an error in the synthesis test now:

../../../../myproject_test.cpp:15:20: error: 'N_OUT000' was not declared in this scope

I suspect the name gets malformed somehow, but from reading the code changes it's not obvious to me what's going wrong.

Otherwise I think this is ready for merge.

[calad0i]

The pytest error is fixed now, but there seems to be an error in the synthesis test now:

../../../../myproject_test.cpp:15:20: error: 'N_OUT000' was not declared in this scope

I suspect the name gets malformed somehow, but from reading the code changes it's not obvious to me what's going wrong.

Otherwise I think this is ready for merge.

Caused by global search & replace at https://github.com/fastmachinelearning/hls4ml/blob/main/hls4ml/writer/vivado_accelerator_writer.py#L314. I have no idea why it is written like this.

[jmitrevs]

I noticed that vivado_accelerator_writer has some lines like line.replace(out.size_cpp(), 'N_OUT').replace(out.name, 'outputs').replace(out.type.name, 'output_axi_t') and newline = line.replace(out.size_cpp(), 'N_OUT').replace(out.type.name, 'output_axi_t'). Could that be the reason? Thought that should only affect the accelerator backend.

[calad0i]

I noticed that vivado_accelerator_writer has some lines like line.replace(out.size_cpp(), 'N_OUT').replace(out.name, 'outputs').replace(out.type.name, 'output_axi_t') and newline = line.replace(out.size_cpp(), 'N_OUT').replace(out.type.name, 'output_axi_t'). Could that be the reason? Thought that should only affect the accelerator backend.

INFO: [HLS 200-10] Creating and opening solution '/var/jenkins_home/workspace/hls4ml_pr_1321/test/hls_prj/KERAS_3layer-backendVivadoAccelerator-boardpynq_z2-xc7z020clg400_1-c5-io_parallel-rf1-ap_fixed_16_6_-Resource-Vivado-2020.1/myproject_prj/solution1'. The failure point is at vivado accelerator, so it is.