blackbeard/miner
1
0
Fork 0
Code Issues 2 Pull Requests Actions Packages Projects Releases 2 Wiki Activity

Add debug option to force GPU winner for early exit

Browse Source 
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.
This commit is contained in:
rob thijssen
2025-09-14 15:06:22 +03:00
parent 456742e425
commit 737de2113a
2 changed files with 75 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/engine-gpu-cuda/src/kernels/qpow_kernel.cu
+38
View File
@@ -337,6 +337,7 @@ __device__ uint8_t C_SAMPLER_TARGET_BE[64];
__device__ uint8_t C_SAMPLER_THRESH_BE[64];
__device__ uint32_t C_SAMPLER_INDEX;
__device__ uint32_t C_SAMPLER_DECISION;
__device__ __constant__ int C_DEBUG_FORCE_WIN;
// Load/store helpers (little- and big-endian)
__device__ __forceinline__ uint64_t load64_le(const uint8_t* p) {
@@ -629,6 +630,43 @@ for (uint32_t j = 0; j < iters; ++j) {
if (atomicAdd(found_flag, 0) != 0) {
return;
}
// Debug: force a winner to exercise GPU->CPU signaling
if (C_DEBUG_FORCE_WIN != 0 && tid == 0 && j == 0) {
if (atomicCAS(found_flag, 0, 1) == 0) {
if (out_index) {
*out_index = tid * iters_per_thread + j;
}
if (out_win_tid) {
*out_win_tid = tid;
}
if (out_win_j) {
*out_win_j = j;
}
if (out_distance_be) {
#pragma unroll
for (int i = 0; i < 64; ++i) {
out_distance_be[i] = 0;
}
}
if (out_dbg_y_be) {
#pragma unroll
for (int i = 0; i < 64; ++i) {
out_dbg_y_be[i] = 0;
}
}
if (out_dbg_h_be) {
#pragma unroll
for (int i = 0; i < 64; ++i) {
out_dbg_h_be[i] = 0;
}
}
if (C_SAMPLER_ENABLE && tid == 0) {
C_SAMPLER_INDEX = tid * iters_per_thread + j;
C_SAMPLER_DECISION = 1u;
}
}
return;
}
// y_hat = y_hat * m_hat
uint64_t yhat_next[8];
crates/engine-gpu-cuda/src/lib.rs
+37 -7
View File
@@ -460,6 +460,34 @@ impl CudaEngine {
c_samp_h.copy_from(&zero64)?;
}
}
// Optional debug: force device winner when MINER_CUDA_FORCE_WIN is set
if let Ok(force_env) = std::env::var("MINER_CUDA_FORCE_WIN") {
let force: i32 = if force_env == "0"
|| force_env.eq_ignore_ascii_case("false")
|| force_env.is_empty()
{
0
} else {
1
};
if let Ok(mut c_force) = module
.get_global::<i32>(CString::new("C_DEBUG_FORCE_WIN")?.as_c_str())
{
c_force.copy_from(&force)?;
if force != 0 {
log::warn!(target: "miner", "CUDA G2: MINER_CUDA_FORCE_WIN enabled (device will force early-exit at tid=0,j=0)");
} else {
log::debug!(target: "miner", "CUDA G2: MINER_CUDA_FORCE_WIN disabled");
}
} else {
log::warn!(target: "miner", "CUDA G2: MINER_CUDA_FORCE_WIN set but device symbol C_DEBUG_FORCE_WIN unavailable");
}
} else if let Ok(mut c_force) =
module.get_global::<i32>(CString::new("C_DEBUG_FORCE_WIN")?.as_c_str())
{
let zero: i32 = 0;
let _ = c_force.copy_from(&zero);
}
// Flag ready = 1
let one: i32 = 1;
c_ready.copy_from(&one)?;
@@ -490,7 +518,11 @@ impl CudaEngine {
}
let active_threads_usize = active_threads as usize;
let grid_dim = (((active_threads as u32) + block_dim - 1) / block_dim).max(1);
log::info!(target: "miner", "CUDA G2 launch: grid_dim={grid_dim}, block_dim={block_dim}, threads={}, iters={effective_iters}, batches={batches}", active_threads);
let force_enabled = std::env::var("MINER_CUDA_FORCE_WIN")
.ok()
.map(|v| v != "0" && !v.is_empty() && !v.eq_ignore_ascii_case("false"))
.unwrap_or(false);
log::info!(target: "miner", "CUDA G2 launch: grid_dim={grid_dim}, block_dim={block_dim}, threads={}, iters={effective_iters}, batches={batches}, force_win={}", active_threads, force_enabled);
let current_be = current.to_big_endian();
let end_be = end.to_big_endian();
let cur_prefix = hex::encode(&current_be[..16]);
@@ -686,8 +718,9 @@ impl CudaEngine {
}
}
// Advance by covered window and continue to next batch, honoring cancel between batches
hash_count = hash_count.saturating_add(covered);
current = current.saturating_add(U512::from(1u64 + covered));
let attempts = active_threads * (effective_iters as u64);
hash_count = hash_count.saturating_add(attempts);
current = current.saturating_add(U512::from(1u64 + attempts));
if cancel.load(AtomicOrdering::Relaxed) {
break;
}
@@ -774,10 +807,7 @@ impl CudaEngine {
}
}
// Not found in this window
hash_count = hash_count.saturating_add(covered);
// Advance and continue loop (no G1 copy-back in G2 mode)
current = current.saturating_add(U512::from(1u64 + covered));
// Not found in this window (no additional advance; already advanced per batch)
continue;
} else {
// G1 launch: computes y for (current + t + 1) for each thread t in [0, num_threads)