CPU scaling issue: 4 vs 8 vs 10 CPUs — quip-miner-cpu 0.3.2-rc3

[subzeroua]

Hi,

I'm testing quip-miner-cpu 0.3.2-rc3 in Docker and noticed that increasing the CPU allocation does not significantly increase the reported mining throughput.

My tests:

4 CPUs:
~0.7 jobs/s

8 CPUs:
~0.7–0.8 jobs/s

10 CPUs:
Docker CPU usage: ~989%
Reported throughput: ~0.9 jobs/s

With 10 CPUs the logs show:

progress: 70 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528
progress: 80 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528
progress: 90 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528
progress: 100 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528
progress: 110 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528
progress: 120 jobs | 0.9 jobs/s | reads=105 sweeps=846 | best=-14452 | requires energy<=-14528

The container is actually consuming almost 10 CPU cores (~989%), so the CPU allocation appears to be effective.

Could you clarify:

1. Does the "jobs/s" value represent aggregate throughput across all CPU SA workers?
2. Is ~0.9 jobs/s expected with 10 CPU cores?
3. Is there a way to see the number of active SA workers and their individual/aggregate throughput?
4. Could there be a bottleneck or configuration issue preventing throughput from scaling with the number of CPU cores?

Version:
quip-miner-cpu 0.3.2-rc3
protocol 1

Thanks!