feat: GF-T backprop microcode generator for the microsequencer - #1864
Merged
Conversation
tools/gft_backprop_microcode.py: generates the full-backprop microcode (op, a, a_mod, b, b_mod, dst steps over a register file) for an arbitrary 2-layer net (n_in, n_hid, n_out) to run on the on-FPGA microsequencer (one shared GftSmul + one shared GftSadd). Self-contained GF-T interpreter; self-test: the generated (2,2,1) XOR microcode trains XOR to 4/4. Any topology runs on the SAME ~3K-LUT datapath; only microcode length (=time) and register file grow. Emitting the Verilog case-ROM from steps is mechanical (board/bpseq.v is the hand-written (2,2,1), silicon-built at 2.93M fasm). Makes the microsequencer a programmable ternary NN trainer under the ~17M ceiling. Refs #1764 Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Contributor
Contributor
|
📓 NotebookLM Notebook linked to this PR
This notebook contains session context, decisions, and artifacts for this work. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
tools/gft_backprop_microcode.py: generates full-backprop microcode for an arbitrary 2-layer net (n_in, n_hid, n_out) to run on the on-FPGA microsequencer (one shared GftSmul + GftSadd). Self-contained GF-T interpreter; self-test: the generated (2,2,1) XOR microcode trains XOR 4/4. Any topology runs on the SAME ~3K-LUT datapath; only microcode length (time) + register file grow. Makes the microsequencer a programmable ternary NN trainer under the ~17M ceiling. board/bpseq.v is the hand-written (2,2,1), silicon-built at 2.93M fasm. docs/NOW.md updated. Refs #1764