 Marek Lemańczyk
|
8adeb59acc
|
Fix kernel integration compile errors
|
2025-10-18 18:01:52 +02:00 |
|
|
Marek Lemańczyk
|
6217ea7888
|
Document GPU kernel audit progress
|
2025-10-18 17:42:33 +02:00 |
|
|
mlemanczyk
|
e5d57364b5
|
GPU limit restored. GpuKernelLease will now return to the pool during Dispose
|
2025-10-17 08:49:22 +02:00 |
|
|
Marek Lemańczyk
|
964e748498
|
Pool GPU kernel leases for reuse
|
2025-10-17 07:04:11 +02:00 |
|
|
mlemanczyk
|
fddec17b3f
|
Various try-out changes
|
2025-10-17 06:53:54 +02:00 |
|
|
Marek Lemańczyk
|
67ad36420a
|
Add GPU kernel for small prime factoring
|
2025-10-17 05:31:07 +02:00 |
|
|
mlemanczyk
|
56e413064a
|
Using cleanup
|
2025-10-17 02:40:51 +02:00 |
|
|
mlemanczyk
|
d3695a3dc3
|
Pools made thread static. Various small optimizations. "in" parameter added to read-only method parameters
|
2025-10-15 02:17:13 +02:00 |
|
|
Marek Lemańczyk
|
c9ae56a5ec
|
Remove default prime order overloads
|
2025-10-10 21:51:41 +02:00 |
|
|
Marek Lemańczyk
|
85a98eb009
|
Reduce GPU remainder transfers for divisor filtering
|
2025-10-08 22:28:36 +02:00 |
|
|
Marek Lemańczyk
|
6918bd28ff
|
Make by-divisor GPU kernels reuse read-only operands
|
2025-10-08 18:47:16 +02:00 |
|
|
Marek Lemańczyk
|
cd1debe9d6
|
Various micro- and medium- optimizations
|
2025-10-07 22:14:34 +02:00 |
|
|
Marek Lemańczyk
|
48127845b5
|
GPU errors won't go silently, so that we can analyze and fix them.
|
2025-10-07 21:57:27 +02:00 |
|
|
Marek Lemańczyk
|
31e95e5a8b
|
Merge pull request #120 from mlemanczyk/codex/refactor-cycle-calculation-methods-for-mersenne
Streamline by-divisor cycle equality checks
|
2025-10-05 21:33:10 +02:00 |
|
|
Marek Lemańczyk
|
16181903a1
|
Streamline by-divisor cycle equality checks
|
2025-10-05 21:32:40 +02:00 |
|
|
Marek Lemańczyk
|
bd8686b86b
|
Unused parameter removed
|
2025-10-05 19:14:42 +02:00 |
|
|
Marek Lemańczyk
|
581da24b2f
|
Restore GPU incremental scan logic behind exception
|
2025-10-05 17:39:31 +02:00 |
|
|
Marek Lemańczyk
|
20120f5bc0
|
Fix odd power table updates in GPU heuristics
|
2025-10-05 16:55:27 +02:00 |
|
|
Marek Lemańczyk
|
26cc574020
|
Treat Pollard Rho success as the GPU primality check
|
2025-10-05 06:26:09 +02:00 |
|
|
Marek Lemańczyk
|
0bc97937aa
|
Restore GPU-backed UInt128 pow2 Montgomery helper
|
2025-10-05 02:13:31 +02:00 |
|
|
Marek Lemańczyk
|
72dd69bd53
|
Align GPU pow2 heuristics with Montgomery helper
|
2025-10-04 22:40:27 +02:00 |
|
|
Marek Lemańczyk
|
b2a7190c14
|
Inline GPU capability checks for pow2 heuristics
|
2025-10-04 22:40:23 +02:00 |
|
|
Marek Lemańczyk
|
38df505f4f
|
Implement GPU pow2 heuristics with safe fallbacks
|
2025-10-04 22:40:12 +02:00 |
|
|
Marek Lemańczyk
|
ea08d3f71f
|
Inline GPU overflow tracking helpers
|
2025-10-04 22:40:07 +02:00 |
|
|
Marek Lemańczyk
|
85c6eaece7
|
Persist GPU capability snapshot without volatile
|
2025-10-04 22:40:02 +02:00 |
|
|
Marek Lemańczyk
|
6adac6b7a7
|
Consolidate multiply helpers and guard GPU fallbacks
|
2025-10-04 21:20:29 +02:00 |
|
|
Marek Lemańczyk
|
84ebbed98b
|
Add shift-first multiply helper and document benchmarks
|
2025-10-04 20:01:10 +02:00 |
|
|
Marek Lemańczyk
|
d3cd408d3d
|
Document patch header requirements
|
2025-10-03 22:32:27 +02:00 |
|
|
Marek Lemańczyk
|
941b51ebcf
|
Windowed Pow2Mod implemented and integrated across repository.
|
2025-10-03 19:37:15 +02:00 |
|
|
Marek Lemańczyk
|
ebe48d5f2f
|
Divisor-Cycle Cache Rework
|
2025-10-02 19:53:09 +02:00 |
|
|
Marek Lemańczyk
|
b8aa845c5c
|
Clarify cycle stepper reset inlining
|
2025-10-02 13:30:56 +02:00 |
|
|
Marek Lemańczyk
|
36d316d19c
|
Add TODOs for divisor metadata caching
|
2025-10-01 22:52:21 +02:00 |
|
|
Marek Lemańczyk
|
513fc7acb0
|
Optimizations and tiny cleanup
|
2025-10-01 13:02:07 +02:00 |
|
|
Marek Lemańczyk
|
604e91b3e3
|
Return scalar MulMod results and benchmark native reducers
|
2025-10-01 00:41:27 +02:00 |
|
|
Marek Lemańczyk
|
4bbdebc874
|
Optimizations and TODOs added
|
2025-09-30 20:22:01 +02:00 |
|
|
Marek Lemańczyk
|
8d44b75ca5
|
Cache native modulo bit masks
|
2025-09-30 17:13:43 +02:00 |
|
|
Marek Lemańczyk
|
ebe3b306b9
|
Refine Mul64 benchmarks with GPU-friendly variant
|
2025-09-30 04:24:40 +02:00 |
|
|
Marek Lemańczyk
|
daab18504b
|
Performance improvements
|
2025-09-30 03:21:51 +02:00 |
|
|
Marek Lemańczyk
|
0bdf85730d
|
Tiny optimization
|
2025-09-29 23:28:35 +02:00 |
|
|
Marek Lemańczyk
|
f63254f0f0
|
Refactor GPU-compatible Mul64 handling
|
2025-09-29 22:35:46 +02:00 |
|
|
Marek Lemańczyk
|
1ba04047c1
|
Extract by-divisor scanner into shared tester
|
2025-09-29 20:49:01 +02:00 |
|
|
Marek Lemańczyk
|
7ef8a304e9
|
Clarify GPU batch capacity guards
|
2025-09-25 22:58:57 +02:00 |
|
|
Marek Lemańczyk
|
9958338820
|
Initialize cycle remainder stepping upfront
|
2025-09-25 21:09:28 +02:00 |
|
|
Marek Lemańczyk
|
9b7611fe62
|
Trim GPU exponent batches using cached divisor cycles
|
2025-09-25 20:23:25 +02:00 |
|
|
Marek Lemańczyk
|
9b9c6959be
|
Reuse cycle remainder stepper in by-divisor testers
|
2025-09-25 18:51:47 +02:00 |
|
|
Marek Lemańczyk
|
b40c61b527
|
Tiny optimizations
|
2025-09-25 12:43:38 +02:00 |
|
|
Marek Lemańczyk
|
05077f7e90
|
Fixed regression with ref taken only once while it may move since..
|
2025-09-25 10:31:34 +02:00 |
|
|
Marek Lemańczyk
|
b792eb90e0
|
Small optimization
|
2025-09-25 09:35:02 +02:00 |
|
|
Marek Lemańczyk
|
74747adc7c
|
Performance improvements
|
2025-09-25 08:29:27 +02:00 |
|
|
Marek Lemańczyk
|
8b95a42f27
|
Divisor cycle cache will now respect --divisor-cycles-batch parameter (renamed from --divisor-cycles-batch-size)
|
2025-09-25 07:02:44 +02:00 |
|