Reply to thread

Message

<blockquote data-quote="bert_r" data-source="post: 419903" data-attributes="member: 90426">Regarding the hashing speed: I just spent several hours trying to improve the hashing speed, and it seems it's already quite fast. I tried several strategies in c#, including starting several threads to hash the chunks parallel, and the largest improvement I could manage is 4%. Thinking that C would be the fastest language to calculate such things, I also tried creating a separate C dll to do the hashing, resulting in a 10% speed improvement.The best that I could manage (until now) is a 10% speed improvement, which isn't much. My files take an average of 6 seconds to hash, which would lower to 5.4 ... hardly a difference. So unless somebody has a brilliant idea, I'm afraid we're stuck with the current hashing speed (except for small improvements).[USER=64817]lwerndly[/USER]: I also love the v2 screenshots so far. If you need any help, let me know!One other problem I noticed, which may also be a problem in v2: When scanning new files, it would probably be best to check if the hashed file already exists. A few days ago I renamed a few directories (I have the habit to start a directory name with an underscore for series I'm currently watching, I can find them easier that way because they're sorted on top). When I opened the dir in MP, all the files in the renamed directory were rehashed (which is logical) and I had lost all watched statuses (which is less logical). When I checked the database, all these files occurred twice (same hash, different dir).</blockquote>

[QUOTE="bert_r, post: 419903, member: 90426"] Regarding the hashing speed: I just spent several hours trying to improve the hashing speed, and it seems it's already quite fast. I tried several strategies in c#, including starting several threads to hash the chunks parallel, and the largest improvement I could manage is 4%. Thinking that C would be the fastest language to calculate such things, I also tried creating a separate C dll to do the hashing, resulting in a 10% speed improvement. The best that I could manage (until now) is a 10% speed improvement, which isn't much. My files take an average of 6 seconds to hash, which would lower to 5.4 ... hardly a difference. So unless somebody has a brilliant idea, I'm afraid we're stuck with the current hashing speed (except for small improvements). [USER=64817]lwerndly[/USER]: I also love the v2 screenshots so far. If you need any help, let me know! One other problem I noticed, which may also be a problem in v2: When scanning new files, it would probably be best to check if the hashed file already exists. A few days ago I renamed a few directories (I have the habit to start a directory name with an underscore for series I'm currently watching, I can find them easier that way because they're sorted on top). When I opened the dir in MP, all the files in the renamed directory were rehashed (which is logical) and I had lost all watched statuses (which is less logical). When I checked the database, all these files occurred twice (same hash, different dir). [/QUOTE]

Verification