Opened 16 years ago
Closed 14 years ago
#1455 closed feature-request (Invalid)
deluge/libtorrent should utilize multiple processors when checkin files
| Reported by: | balaton | Owned by: | |
|---|---|---|---|
| Priority: | trivial | Milestone: | |
| Component: | libtorrent | Version: | other (please specify) |
| Keywords: | Cc: |
Description
Currently, deluge/libtorrent seems to use only one processor for checking torrents. Using multiple proccessors on a multicore/multicpu machine could speed up this process rapidly.
Change History (7)
comment:1 by , 16 years ago
| Status: | new → pending |
|---|
comment:2 by , 16 years ago
| Status: | pending → new |
|---|
I ran a deluged on a dedicated machine and had it run a check of a torrent. I measured the disk i/o with iotop and it way about 30mb/s read and basically no writes. From that I concluded that there would be plenty of room to check another torrent simultaniously.
comment:3 by , 16 years ago
Right now a moderately fast dual core processor can handle checking of a dozen torrents even with one core. Implementing this for the few systems that have a slow dual core processor and a fast disk is pointless. Besides, I think it is bad for HDD durability if the head has to move back and forth between two torrents for checking. Especially if the CPU is actually capable to max out the disk I/O bottleneck.
comment:4 by , 15 years ago
| Milestone: | 1.3.x → Future |
|---|
comment:5 by , 15 years ago
| Milestone: | Future |
|---|
comment:6 by , 15 years ago
| Priority: | major → trivial |
|---|
comment:7 by , 14 years ago
| Resolution: | → invalid |
|---|---|
| Status: | new → closed |



I suspect the bottleneck happens more from disk I/O than cpu. Have you gathered any statistics to prove otherwise? What data have you gathered that you're willing to share?