mlemanczyk/even-perfect-numbers-scanner
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Inline legacy GPU sieve and update heuristic fallbacks

Browse Source
This commit is contained in:
Marek Lemańczyk committed 2025-10-26 17:06:13 +01:00
1 parent a7fc34e37a
commit 55554d9565
5 files changed
+10 -14
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

docs/HeuristicIsPrimePlan.md
+1 -1
View File
@@ -140,7 +140,7 @@ The plan incorporates the updated divisor-class heuristics (Groups A/B, wheels,
* [done] Streamed GPU batches from the same enumerator and Montgomery preparation data, removing the filtered-divisor scratch arrays while continuing to feed the existing kernel and hit bookkeeping.
9. [done] **Route legacy prime checks through heuristics**
* [done] Introduce `HeuristicPrimeTester` so heuristic CPU/GPU checks live alongside the legacy `PrimeTester` implementations while callers opt in explicitly.
* [done] Maintain legacy paths for benchmarking via `PrimeTester.IsPrimeInternal`, `PrimeTester.IsPrimeGpuFallback`, and the global `HeuristicPrimeTester.EnableHeuristicPrimeTesting` switch so benchmark harnesses can compare behaviours.
* [done] Maintain legacy paths for benchmarking via `PrimeTester.IsPrimeInternal`, `PrimeTester.Exclusive.IsPrimeGpu`, and the global `HeuristicPrimeTester.EnableHeuristicPrimeTesting` switch so benchmark harnesses can compare behaviours.
10. **Validation and benchmarking**
* Create targeted unit tests covering boundary cases for Group A/B transitions, wheel enumerations, residue deltas, and the stop-on-hit behaviour.
* Expand GPU test harnesses to assert equivalence between CPU/GPU heuristics on sampled composite and prime inputs and to verify the residue-stream contract with by-divisor sessions.
docs/HeuristicIsPrimeSpecification.md
+1 -1
View File
@@ -62,7 +62,7 @@ No other endings are allowed or reordered.
## Default Prime Tester Entry Points
* `PrimeTester.IsPrime` and `PrimeTester.IsPrimeGpu` continue to run the legacy trial-division and GPU sieve paths. Use `HeuristicPrimeTester` when the heuristic CPU/GPU flow is required.
* Toggle `HeuristicPrimeTester.EnableHeuristicPrimeTesting` or fall back to `PrimeTester.IsPrimeInternal` / `PrimeTester.IsPrimeGpuFallback` to control heuristic rollouts in test harnesses.
* Toggle `HeuristicPrimeTester.EnableHeuristicPrimeTesting` or fall back to `PrimeTester.IsPrimeInternal` / `PrimeTester.Exclusive.IsPrimeGpu` to control heuristic rollouts in test harnesses.
* Benchmark harnesses can compare behaviours by calling the legacy helpers on `PrimeTester` alongside the heuristic equivalents on `HeuristicPrimeTester`.
## Integration with the By-Divisor Pipeline