aloha1357 opened a new pull request, #1388:
URL: https://github.com/apache/mahout/pull/1388
### Related Issues
<!-- Closes #123 -->
related #1385
### Changes
- [ ] Bug fix
- [x] New feature
- [x] Refactoring
- [ ] Documentation
- [ ] Test
- [ ] CI/CD pipeline
- [ ] Other
### Why
For smaller qubit counts ($N \le 12$), calculating the Fast Walsh-Hadamard
Transform (FWT) using multiple global memory kernel launches becomes heavily
DRAM bandwidth-bound (launch overhead and global memory roundtrips dominate the
execution time). By keeping the entire state vector within the GPU's Shared
Memory (which is much faster and has lower latency than DRAM), we can fuse the
operations into a single kernel launch.
### How
- **Operator Fusion Kernel (`iqp_phase_fwt_normalize_tc_kernel`):** Created
a new fused kernel that handles three steps entirely within Shared Memory:
1. Computes the IQP phase and writes it directly to `extern __shared__
cuDoubleComplex shared_state[]`.
2. Performs the in-place Hadamard FWT over the shared memory buffer.
3. Normalizes the final amplitudes and writes them out to Global Memory.
- **Dynamic Dispatch:** Updated `launch_iqp_encode_tc` to dynamically
allocate Shared Memory and dispatch to this fused kernel when `num_qubits <=
FWT_SHARED_MEM_THRESHOLD`.
## Checklist
- [x] Added or updated unit tests for all changes (Verified passing against
existing CI test suite)
- [x] Added or updated documentation for all changes (Added explanatory
inline comments for PR)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
