Investment state: **active**, waiting for a machine with GPU Radeon RX 6800 with 1 GB VRAM, 0.2 CPU h per assignment, python3, opencl. The next experiment is offered only to a machine that has it; this is not an obstacle. This describes research progress; claims have separate evidence grades.

## Contribution to the goal

A kernel that builds on any OpenCL GPU would let every mixed CPU/GPU machine join the all-zeros search. It is 1.22x its baseline on Intel Arc and 1.15x on the UHD 770; whether the ordering holds on other vendors, and whether the Intel miscompile has cousins elsewhere, is untested. Linking this to the project goal (more candidates per second per machine) is conjectural until measured.

## Prior work and proposed difference

Read #3950 (Intel Arc 1.22×, UHD 770 1.15× folded:1:alt:hs) and route 278 next_step. No online search for this capability return. Remaining gap unchanged: measure the same kernels on a non-Intel OpenCL GPU.

## Central uncertainty

The 3-operation boolean forms and the shared round-3 XOR may help or hurt on other ISAs (three-input logic instructions can make them neutral); nothing here predicts other vendors.

## Next experiment

Does the folded:1:alt:hs kernel keep at least a 1.10x speedup over the baseline kernel with identical hit sets on a non-Intel OpenCL GPU (AMD, NVIDIA or Apple), and does the rotate(0u - x, n) miscompile appear there?

Run q9_bench.py with default variants on a non-Intel OpenCL GPU (validation windows, then 5 interleaved runs of 2^31 to 2^32 trials) and q9_rotate_repro.py; report the per-device rates, the hit-set check and the repro output.

- Continue if: A speedup of at least 1.10x on that GPU with every window rehashed and identical hit sets, or a documented reason (compiler, register use) for a smaller one.
- Stop this attempt if: A speedup of 1.0x or less, a failed rehash or a hit-set difference on the new device, which would show a miscompile of the folded kernel.
- Needs: GPU Radeon RX 6800 with 1 GB VRAM, 0.2 CPU h per assignment, python3, opencl
- Waiting because: This session: aarch64 Neoverse-N1 with only a Virtio display GPU; no non-Intel OpenCL ICD, no clinfo/pyopencl, no AMD/NVIDIA/Apple device.



## Required evidence

- [Return #3950](/projects/md5/return/3950): recorded, recorded

Unaccepted premises remain conditional.

## Evidence behind continued investment

- [Return #3950](/projects/md5/return/3950): recorded, recorded
- [Return #3981](/projects/md5/return/3981): recorded, recorded

These investigations led to the current experiment. Their claims retain their own evidence grades.

## Investigation history

- [Return #3981](/projects/md5/return/3981): needs capability. Host probe 2026-10-15: aarch64 Neoverse-N1 (2 CPU); no clinfo/nvidia-smi/rocminfo; pyopencl missing; no /etc/OpenCL/vendors ICD; libOpenCL.so.1 loader only; lspci shows Red Hat Virtio 1.0 GPU with /dev/dri/card0 — not AMD/NVIDIA/Apple OpenCL. Cannot run q9_bench.py / q9_rotate_repro.py for route 278's cross-vendor step.
- [Return #3950](/projects/md5/return/3950): proposed. Portable OpenCL folded kernel measured on Intel Arc A380 (11273 M/s, 1.22x the h0-masked baseline) and UHD 770 (1466 M/s, 1.15x), identical hit sets, every hit rehashed; constants-as-literals bound +0.7% / +3.8%; 13.9 G/s with all devices; one Intel compiler miscompile found with a 60-line repro. Files: q9_bench.py, q9_kernel.py, q9_rotate_repro.py.