Here's where top hangs out on Libero. Seems it would be a better situation if Tor would actually use the second core.
top - 12:05:07 up 5 days, 21:35, 1 user, load average: 0.33, 0.43, 0.34 Tasks: 130 total, 2 running, 128 sleeping, 0 stopped, 0 zombie Cpu0 : 47.2%us, 21.0%sy, 0.0%ni, 17.6%id, 0.0%wa, 0.0%hi, 13.8%si, 0.3%st Cpu1 : 0.0%us, 0.3%sy, 0.0%ni, 99.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 1914136k total, 718704k used, 1195432k free, 139588k buffers Swap: 3981304k total, 0k used, 3981304k free, 116320k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1340 _tor 20 0 566m 224m 36m R 71.7 12.0 6168:37 tor 9887 nobody 20 0 21644 13m 892 S 7.3 0.7 0:31.85 scamper 9889 root 20 0 102m 916 772 S 0.7 0.0 0:02.80 sed 9921 root 20 0 15028 1312 992 R 0.3 0.1 0:00.19 top 1 root 20 0 19232 1480 1212 S 0.0 0.1 0:00.27 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
Hi,
Did you consider the numCPUs option ? -- Mit freundlichen Grüssen / Sincerely yours
Sebastian Urbach
---------------------------------------------- Those who would give up essential Liberty, to purchase a little temporary Safety, deserve neither Liberty nor Safety. ---------------------------------------------- Benjamin Franklin (1706 - 1790), Inventor, journalist, printer, diplomat, and statesman
Here's where top hangs out on Libero. Seems it would be a better situation if Tor would actually use the second core.
top - 12:05:07 up 5 days, 21:35, 1 user, load average: 0.33, 0.43, 0.34 Tasks: 130 total, 2 running, 128 sleeping, 0 stopped, 0 zombie Cpu0 : 47.2%us, 21.0%sy, 0.0%ni, 17.6%id, 0.0%wa, 0.0%hi, 13.8%si, 0.3%st Cpu1 : 0.0%us, 0.3%sy, 0.0%ni, 99.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st Mem: 1914136k total, 718704k used, 1195432k free, 139588k buffers Swap: 3981304k total, 0k used, 3981304k free, 116320k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 1340 _tor 20 0 566m 224m 36m R 71.7 12.0 6168:37 tor 9887 nobody 20 0 21644 13m 892 S 7.3 0.7 0:31.85 scamper 9889 root 20 0 102m 916 772 S 0.7 0.0 0:02.80 sed 9921 root 20 0 15028 1312 992 R 0.3 0.1 0:00.19 top 1 root 20 0 19232 1480 1212 S 0.0 0.1 0:00.27 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd
_______________________________________________ tor-relays mailing list tor-relays@lists.torproject.org https://lists.torproject.org/cgi-bin/mailman/listinfo/tor-relays
The NumCPUs option only causes OnionSkins (circuit creation requests) to be processed on additional threads. This is only a portion of the CPU cost of a tor relay (doesn't include TLS, circuit-level AES, packet handling/memory operations, etc), and now that the network is switching over to the considerably faster ntor circuit handshake, it will become an even smaller one.
Making Tor fully multithreaded for all of its CPU-intensive operations is a large task.
For now, the best option for high speed relays is to run multiple Tor daemons with separate OR ports. If you have AES-NI, you can do about 300Mbit per Tor daemon. If not, you can do about 100Mbit per daemon. You can run at most two daemons per external IP, so if you have a higher capacity uplink than that, you have to get more IPs from your provider.
Sebastian Urbach:
tor-relays@lists.torproject.org
-
Mike Perry -
Sebastian Urbach -
tor＠t-3.net