Commit Graph
132 Commits
Author SHA1 Message Date
illuzen 45e0d7d384 sort of almost 2025-11-17 21:35:25 +08:00
illuzen 36cb0efb8a closer 2025-11-17 18:38:04 +08:00
illuzen 93696a5f4b a bit of debug 2025-11-17 17:40:38 +08:00
illuzen e7156dd890 closer 2025-11-17 17:05:35 +08:00
illuzen 9635482098 test vectors 2025-11-17 16:54:20 +08:00
illuzen c214aa9d47 kinda poseidon2 2025-11-17 16:51:32 +08:00
illuzen 6fc2e9a209 separate wgsl file 2025-11-17 16:49:51 +08:00
illuzen 4578e076ce toy hasher 2025-11-17 16:47:23 +08:00
illuzen 50958ef4da ok wgpu runs hello world 2025-11-17 16:37:38 +08:00
Nikolaus Heger e92c4f0769 update comment 2025-11-07 14:40:13 +08:00
illuzen 63d7e114c2 Support new pow algo on cpu, not gpu (#22)
* support new pow algo, remove montgomery

* format

* remove redundant check

* use local poseidon until we release

* back to release

* remove unnecessary fxn

* comment out cuda for now

* comment out open-cl

* fix more tests

* remove cuda jobs
2025-11-03 16:23:18 +08:00
Nikolaus Heger f35660c9e9 clippy 2025-10-16 16:03:25 +08:00
Nikolaus Heger 3466eff8b0 fmt 2025-10-16 16:01:10 +08:00
illuzen c0bfca592d added benchmarks 2025-10-13 23:10:41 +08:00
illuzen 534b3ac2f8 qpow-math from git 2025-10-13 21:53:11 +08:00
illuzen ff3e65361a remove qpow reimplementation 2025-10-13 21:36:44 +08:00
illuzen aa1204bab9 switch to poseidon
pow-core updated, checked that it is compatible with chain/main
2025-09-25 22:27:17 +08:00
Cezary Olborskiandczareko abc3fcfcd3 chore: RM - refreshed (#15)
* chore: Proper version distribution

* chore: version based on tags

---------

Co-authored-by: czareko <czareko@users.noreply.github.com>
2025-09-23 17:22:49 +08:00
Cezary Olborski c820e0bd84 chore: Lock update + logs warn fix (#9) 2025-09-23 12:48:51 +08:00
Cezary Olborski 4a4d3a612a chore: quantus-miner-api dependency (#8) 2025-09-23 12:25:38 +08:00
Cezary Olborski cc536cef9e fix: aarch64 compilation (#6)
* fix: aarch64 compilation

* chore: Clippy fixes
2025-09-23 10:08:17 +08:00
rob thijssen 1103eb3efe Add miner-telemetry crate and integrate telemetry support
- Introduce miner-telemetry crate for WebSocket-based telemetry
reporting - Add telemetry configuration options to miner-cli CLI - Pass
telemetry CLI args as environment variables for miner-service - Enable
miner-service to bootstrap telemetry from env vars and emit events -
Emit periodic system.interval telemetry with mining stats and metadata -
Add miner-telemetry as dependency to miner-service and workspace member
2025-09-17 14:26:51 +08:00
rob thijssen 05a51fe5df Simplify short SHA extraction in build script 2025-09-17 00:12:26 +08:00
rob thijssen 058a287dcd Use if let Ok instead of if let Some for GITHUB_SHA check 2025-09-17 00:11:54 +08:00
rob thijssen 74858c3280 Sort features in Cargo.toml for miner-service metrics flag 2025-09-17 00:10:32 +08:00
rob thijssen b4e9bf0c42 Add FoundOrigin enum to track candidate source in engine results 2025-09-16 23:09:25 +08:00
rob thijssen a0a7712493 Add found-origin metrics and logging for CUDA engine candidates
- Track candidate origin (cpu, gpu-g1, gpu-g2) in metrics with counters
and gauges - Log detailed info on candidate discovery per origin in CUDA
engine - Emit per-job found-origin gauge in miner-service with actual
job_id - Cleanup origin metrics on
2025-09-16 22:49:47 +08:00
rob thijssen a824148b68 Lower CUDA log level from debug to trace for finer-grained output 2025-09-16 21:47:44 +08:00
rob thijssen cdf197cd15 Add extensive debug and info logging to CUDA engine initialization and
kernel launches

Include detailed logs for device info, context creation, buffer
allocations, kernel resolution and launch parameters, and data transfers
to help diagnose CUDA behavior and failures. Also add fallback and error
logging enhancements.
2025-09-16 21:02:32 +08:00
rob thijssen 8f051f543c Add build.rs to embed version with git SHA in miner-cli binary 2025-09-16 12:24:24 +08:00
rob thijssen fbd63b580f revert all --version semver changes 2025-09-15 09:38:50 +08:00
rob thijssen e899709f72 Revert "Add build script to append git SHA to version info"
This reverts commit 7674e3141f.
2025-09-15 09:29:11 +08:00
rob thijssen 7674e3141f Add build script to append git SHA to version info 2025-09-15 09:14:40 +08:00
rob thijssen 5867ddefe9 append short git sha to the semantic version returned by the cli when
invoked with `--version`
2025-09-15 09:01:27 +08:00
rob thijssen b108ccefbe simplify sha3 digest handling in qpow kernel
add optional cuda sampler parity logging for debugging false positives
2025-09-15 08:44:34 +08:00
rob thijssen 2480d7828b Cache CUDA context and module in CudaEngine for reuse
Initialize CUDA context and load module once per engine instance,
storing them in a mutex-protected cache to avoid repeated setups. Use
cached context/module for kernel launches, creating a fresh stream each
call. Adjust grid dimension calculation to consider multiprocessor
count.
2025-09-14 18:49:19 +03:00
rob thijssen 7e6eee8c72 Add watchdog logging and improve result channel handling in miner
service
2025-09-14 18:33:50 +03:00
rob thijssen 2c7709413b Use Rust 1.58+ format strings in miner logs and remove unused binding 2025-09-14 17:35:23 +03:00
rob thijssen 4ba983a4c0 Add job ID to miner thread logs for better traceability 2025-09-14 17:05:14 +03:00
rob thijssen 2596e62003 Limit counted CUDA attempts to covered nonces for accuracy 2025-09-14 16:03:34 +03:00
rob thijssen 0d81f3a346 Autotune CUDA desired_iters and improve logging
Add dynamic tuning of desired_iters based on coverage and thread count.
Enhance info logs with attempts and utilization metrics. Update example
config.
2025-09-14 15:43:36 +03:00
rob thijssen f37e70e1b1 Fix nonce advancement logic and update CUDA miner env settings 2025-09-14 15:31:46 +03:00
rob thijssen 737de2113a Add debug option to force GPU winner for early exit
Introduce C_DEBUG_FORCE_WIN device constant to force a winner on GPU at
tid=0,j=0. Enable control via MINER_CUDA_FORCE_WIN environment variable
in Rust to trigger early exit and aid debugging.
2025-09-14 15:06:22 +03:00
rob thijssen 456742e425 fix kernel indexing and use correct digest bytes in qpow CUDA kernel
enable sampler in miner config and increase verify iterations to 2048
2025-09-14 14:47:44 +03:00
rob thijssen c3fa990402 optional host-side verification for CUDA G2 batches
support for verifying first W candidates on the host when
MINER_CUDA_VERIFY is set, logging any missed winners.

add a metric for missed winners and info logging for found flags.
2025-09-14 14:29:28 +03:00
rob thijssen e9eada661d mark job result as served and track serving metrics in miner service
keep completed jobs with served results for cleanup and log serving
info. add metric increments for result readiness and serving events.
re-verify nonce distance before serving and update job state
accordingly.
2025-09-14 14:05:26 +03:00
rob thijssen 531eb7df99 refine mining metrics to reduce redundancy and improve cleanup
remove redundant estimated rate gauges and ensure all related job and
thread metrics are cleaned up on job cancellation and completion.

update GPU CUDA metrics to use backend-specific estimated rate.
2025-09-14 13:08:42 +03:00
rob thijssen c385f25af8 add set_job_estimated_rate_backend function to metrics crate 2025-09-14 12:54:01 +03:00
rob thijssen 80f3bb3eae Add janitor thread to prune stale job and thread metrics 2025-09-14 12:50:58 +03:00
rob thijssen ca9f97a265 adjust active thread count based on desired iterations in CUDA kernel
launch
2025-09-14 12:18:43 +03:00