{"thread":{"id":"15551","subject":"multi-threading or processing of converts","startedAt":"2008-09-16T14:23:00Z","lastAt":"2008-09-16T14:23:00Z","messageCount":1,"participants":["dhruva"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"90836","messageId":"655551.96611.qm@web95006.mail.in2.yahoo.com","threadId":"15551","inReplyTo":null,"subject":"multi-threading or processing of converts","fromName":"dhruva","fromEmail":"dhruva@ymail.com","sentAt":"2008-09-16T14:23:00Z","receivedAt":"2008-09-16T14:23:00Z","isPatch":false,"sender":{"key":"dhruva@ymail.com","avatar":null},"body":"Hi,\n I am running git-p4 to import around 11147 changes from perforce, it has done 38% in 4h39m. Wonder if it would have been faster to get a bunch of changes in parallel and import them sequentially. I do understand that importing of full history does not happen too often by still it might be a good idea.\n Have a multi-threaded fetcher from source SCM and dump it in some format locally. git fast-import can pull it from the local at its own speed (which will be real fast). Basically, I am suggesting better using the available bandwidth and CPU resources for doing network IO.\n\n-dhruva\n\n\n\n      Cricket on your mind? Visit the ultimate cricket website. Enter http://in.sports.yahoo.com/cricket/\n"}]}