[PERFORMANCE] Grouped convolutions appear dominated by overhead

Assessment

This issue has not been assessed yet.

Description

I'm profiling the models here and I'm getting poor results when grouped convolutions are used.

For example, 1x1_groups looks at scaling upt he number of groups in an 8-channel 1x1:

This plot makes it look like the overhead is what's dominating the calculation since the compute time scales linearly with the number of groups.

This is bad because it means that grouped convolutions are basically useless--better to just use the full matrix with a ton of zeroes.

I suspect that compile-time optimizations may solve this, but I'm surprised that it's this bad.

Contributor guide

No contributing guide indexed for this repository

First steps

1. Read the whole issue, then the project's contributing guide.
2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
3. Fork the repository and make your change on a branch.
4. Open a pull request that references the issue number.

More from sdatkinson/NeuralAmpModelerCore

• get_dsp() aborts the process on well-formed JSON that is not a NAM model Open

  Difficulty 2/5 1-3 hours Newbie friendliness 86/100

  sdatkinson/NeuralAmpModelerCore#328 · Opened Sep 5, 2026

• Documentation: C++20 required on Windows Open
  documentation

  Difficulty 2/5 1-3 hours Newbie friendliness 68/100

  sdatkinson/NeuralAmpModelerCore#305 · 1 comment · Opened Jul 12, 2026

• Build on FreeBSD: "minor" and "major" in get_dsp.h cause build error Open

  Difficulty 2/5 1-3 hours Newbie friendliness 65/100

  sdatkinson/NeuralAmpModelerCore#304 · 2 comments · Opened Jul 4, 2026

• Architecture parsers are dropped when Core is linked as a static library Open

  Difficulty 4/5 3-5 days Newbie friendliness 55/100

  sdatkinson/NeuralAmpModelerCore#327 · 1 reaction · Opened Sep 5, 2026

• [BUG] Loading certain A1 models after A2 models can cause audio crackling / dropouts Open

  Difficulty 4/5 3-5 days Newbie friendliness 45/100

  sdatkinson/NeuralAmpModelerCore#303 · 2 comments · 1 reaction · Opened Jun 29, 2026

All issues in sdatkinson/NeuralAmpModelerCore

Similar issues

• Eval bug: SIGSEGV in token-counting routes when the request arrives while the server is sleeping (stale vocab/mctx captured before the wake barrier) Open

  Difficulty 2/5 1-3 hours Newbie friendliness 86/100

  ggml-org/llama.cpp#29188 · Opened Sep 20, 2026

• Sensor initialization takes very long when `--initial-sim-time` is set to current UNIX timestamp Open

  Difficulty 2/5 1-3 hours Newbie friendliness 78/100

  gazebosim/gz-sensors#662 · 1 comment · Opened Sep 20, 2026

• Aligned columns on the serial monitor while keeping CSV on the SD card Open
  enhancement

  Difficulty 2/5 1-3 hours Newbie friendliness 76/100

  NorthernWidget/Margay_Library#31 · Opened Sep 20, 2026

• v1 manifest list: existing_files_count is the table-wide total-data-files Open
  comp-datalake

  Difficulty 2/5 1-3 hours Newbie friendliness 88/100

  ClickHouse/ClickHouse#121222 · Opened Sep 20, 2026

• Case collision between `Canvas` and `canvas` test directories causes platform-dependent source hashes Open

  Difficulty 2/5 1-3 hours Newbie friendliness 68/100

  LadybirdBrowser/ladybird#12123 · Opened Sep 20, 2026

More C++ issues