|
|
|
@ -157,7 +157,7 @@ static int accepted, rejected;
@@ -157,7 +157,7 @@ static int accepted, rejected;
|
|
|
|
|
int hw_errors; |
|
|
|
|
static int total_queued, total_staged, lp_staged; |
|
|
|
|
static bool localgen = false; |
|
|
|
|
static unsigned int getwork_requested = 0; |
|
|
|
|
static unsigned int getwork_requested; |
|
|
|
|
static char current_block[37]; |
|
|
|
|
static char longpoll_block[37]; |
|
|
|
|
static char blank[37]; |
|
|
|
@ -563,17 +563,17 @@ static bool submit_upstream_work(const struct work *work)
@@ -563,17 +563,17 @@ static bool submit_upstream_work(const struct work *work)
|
|
|
|
|
} |
|
|
|
|
|
|
|
|
|
utility = accepted / ( total_secs ? total_secs : 1 ) * 60; |
|
|
|
|
efficiency = getwork_requested ? cgpu->accepted * 100.0 / getwork_requested : 0.0; |
|
|
|
|
efficiency = cgpu->getworks ? cgpu->accepted * 100.0 / cgpu->getworks : 0.0; |
|
|
|
|
|
|
|
|
|
if (!opt_quiet) { |
|
|
|
|
printf("[%sPU %d] [%.1f Mh/s] [Q:%d A:%d R:%d HW:%d E:%.0f%% U:%.2f/m] \n", |
|
|
|
|
cgpu->is_gpu? "G" : "C", cgpu->cpu_gpu, cgpu->total_mhashes / total_secs, |
|
|
|
|
getwork_requested, cgpu->accepted, cgpu->rejected, cgpu->hw_errors, |
|
|
|
|
cgpu->getworks, cgpu->accepted, cgpu->rejected, cgpu->hw_errors, |
|
|
|
|
efficiency, utility); |
|
|
|
|
print_status(); |
|
|
|
|
} |
|
|
|
|
applog(LOG_INFO, "%sPU %d Requested:%d Accepted:%d Rejected:%d HW errors:%d Efficiency:%.0f%% Utility:%.2f/m", |
|
|
|
|
cgpu->is_gpu? "G" : "C", cgpu->cpu_gpu, getwork_requested, cgpu->accepted, cgpu->rejected, cgpu->hw_errors, efficiency, utility |
|
|
|
|
cgpu->is_gpu? "G" : "C", cgpu->cpu_gpu, cgpu->getworks, cgpu->accepted, cgpu->rejected, cgpu->hw_errors, efficiency, utility |
|
|
|
|
); |
|
|
|
|
|
|
|
|
|
json_decref(val); |
|
|
|
@ -1061,6 +1061,8 @@ static bool get_work(struct work *work, bool queued)
@@ -1061,6 +1061,8 @@ static bool get_work(struct work *work, bool queued)
|
|
|
|
|
bool ret = false; |
|
|
|
|
int failures = 0; |
|
|
|
|
|
|
|
|
|
getwork_requested++; |
|
|
|
|
|
|
|
|
|
retry: |
|
|
|
|
if (unlikely(!queued && !queue_request())) { |
|
|
|
|
applog(LOG_WARNING, "Failed to queue_request in get_work"); |
|
|
|
@ -1194,9 +1196,10 @@ static void *miner_thread(void *userdata)
@@ -1194,9 +1196,10 @@ static void *miner_thread(void *userdata)
|
|
|
|
|
/* obtain new work from internal workio thread */ |
|
|
|
|
if (unlikely(!get_work(&work, requested))) { |
|
|
|
|
applog(LOG_ERR, "work retrieval failed, exiting " |
|
|
|
|
"mining thread %d", mythr->id); |
|
|
|
|
"mining thread %d", thr_id); |
|
|
|
|
goto out; |
|
|
|
|
} |
|
|
|
|
mythr->cgpu->getworks++; |
|
|
|
|
work.thr_id = thr_id; |
|
|
|
|
needs_work = requested = false; |
|
|
|
|
work.blk.nonce = 0; |
|
|
|
@ -1439,11 +1442,10 @@ static void *gpuminer_thread(void *userdata)
@@ -1439,11 +1442,10 @@ static void *gpuminer_thread(void *userdata)
|
|
|
|
|
"gpu mining thread %d", mythr->id); |
|
|
|
|
goto out; |
|
|
|
|
} |
|
|
|
|
mythr->cgpu->getworks++; |
|
|
|
|
work->thr_id = thr_id; |
|
|
|
|
requested = false; |
|
|
|
|
|
|
|
|
|
getwork_requested++; |
|
|
|
|
|
|
|
|
|
precalc_hash(&work->blk, (uint32_t *)(work->midstate), (uint32_t *)(work->data + 64)); |
|
|
|
|
work->blk.nonce = 0; |
|
|
|
|
work_restart[thr_id].restart = 0; |
|
|
|
|