From what I understand the latest server software is multi-threaded. How many threads does it use? Even before the update people were boasting about their hexacore processors on their threads!
So how many threads does it use and how much load is put on them?
Rollback Post to RevisionRollBack
"Terminator like robots may one day rule the world, as long as they don't run Windows Vista"
“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” — Albert Einstein
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Not the Spout plugin, the Spout project is writing their own server from scratch, without using any Notchcode in the process. If the rumors are true, it'll end up being the unofficial replacement for Bukkit, and it should support multithreading.
“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” — Albert Einstein
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Once Spout comes out, every server owner with a shred of common sense will quickly move over. Let's see, open source and not based on Notchcode, with multithreading and a open community, or closed source and regulated API with worse performance.
“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” — Albert Einstein
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Once Spout comes out, every server owner with a shred of common sense will quickly move over. Let's see, open source and not based on Notchcode, with multithreading and a open community, or closed source and regulated API with worse performance.
Thinking as a big admin here:
Apart from Multi-threading what is in it for me? Open source isn't a big benefit. And it definitely does not make up for having to wait for updates when Mojang's version will come out in sync with the client.
I love the idea of Spout's client server relationship but it would be far better suited as part of Minecraft itself- having things download to the client = cool. Having to download something to get it to work = sadface.
NB: Has Spout had any plans to work with Forge? If they worked together then it would be awesome (as Forge mods don't modify the actual Jar file they can be added on in runtime I guess).
Rollback Post to RevisionRollBack
"Terminator like robots may one day rule the world, as long as they don't run Windows Vista"
Thinking as a big admin here:
Apart from Multi-threading what is in it for me? Open source isn't a big benefit. And it definitely does not make up for having to wait for updates when Mojang's version will come out in sync with the client.
I love the idea of Spout's client server relationship but it would be far better suited as part of Minecraft itself- having things download to the client = cool. Having to download something to get it to work = sadface.
NB: Has Spout had any plans to work with Forge? If they worked together then it would be awesome (as Forge mods don't modify the actual Jar file they can be added on in runtime I guess).
Since the server is being written from scratch, they claim updates will be fast.
Multithreading is a huge advantage.
Open source and open community allows for them to do things and make mods that the Mod API would never allow. In addition, they claim that it'll support legacy bukkit mods, unlike the API, which isn't going to provide this type of transition path.
“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” — Albert Einstein
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Since the server is being written from scratch, they claim updates will be fast.
Multithreading is a huge advantage.
Open source and open community allows for them to do things and make mods that the Mod API would never allow. In addition, they claim that it'll support legacy bukkit mods, unlike the API, which isn't going to provide this type of transition path.
Sometimes you just have to drop legacy support to move on. On and by the way Bukkit was written from scratch and its updates were behind in some cases by weeks other times they were pretty fast.
Open Source is an interesting point but it isn't a huge benefit considering people decompile MC anyway and sometimes there stuff gets added in (unless it is released under an annoying license like GNU/GPL that won't let the coders give it to Mojang). Multi Threading is an advantage, but how many threads will it use and how well will it use them? Also most servers (as in nearly every one of the 24/7 servers) are on VPSes where they only get a share of a CPU (which is normally less than the power of a core). In these cases Multi-threading has no advantages.
Rollback Post to RevisionRollBack
"Terminator like robots may one day rule the world, as long as they don't run Windows Vista"
Sometimes you just have to drop legacy support to move on. On and by the way Bukkit was written from scratch and its updates were behind in some cases by weeks other times they were pretty fast.
Open Source is an interesting point but it isn't a huge benefit considering people decompile MC anyway and sometimes there stuff gets added in (unless it is released under an annoying license like GNU/GPL that won't let the coders give it to Mojang). Multi Threading is an advantage, but how many threads will it use and how well will it use them? Also most servers (as in nearly every one of the 24/7 servers) are on VPSes where they only get a share of a CPU (which is normally less than the power of a core). In these cases Multi-threading has no advantages.
All Bukkit plugins will break with the new mod API, while with Spout you'll be able to use most of them as-is until Spout versions are released.
The transition from Bukkit to the Mod API will be chaos, possibly worse than the hMod to Bukkit transition.
The Bukkit project is basically giving the middle finger to their developers, by merging with vanilla and abandoning the Bukkit API.
Multithreading is extremely useful for larger servers. Quite a few VPS and dedis have multiple cores, and being able to spread the load across multiple threads is a great feature.
Spout will allow developers much greater freedom with creating plugins, and they'll be able to propose changes to the API itself, unlike with vanilla.
“Two things are infinite: the universe and human stupidity; and I'm not sure about the universe.” — Albert Einstein
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
All Bukkit plugins will break with the new mod API, while with Spout you'll be able to use most of them as-is until Spout versions are released.
The transition from Bukkit to the Mod API will be chaos, possibly worse than the hMod to Bukkit transition.
The Bukkit project is basically giving the middle finger to their developers, by merging with vanilla and abandoning the Bukkit API.
Multithreading is extremely useful for larger servers. Quite a few VPS and dedis have multiple cores, and being able to spread the load across multiple threads is a great feature.
Spout will allow developers much greater freedom with creating plugins, and they'll be able to propose changes to the API itself, unlike with vanilla.
As a developer and a (private) server admin I can say that I can't wait for Bukkit's move. Developers will have to re code their mods but hopefully it won't be to painful. But after that there will be no more waiting for Bukkit to update (which sucks) which is a huge benefit. Lots of servers with mods like IC2 run vanilla software with just to try and get the "update time" shortened (NB: They have to wait for extra things to have IC2 on Bukkit).
Chaos? I think it could be relatively smooth, unless the Spout Server is out (I can only find the Bukkit plugin) all of the developers will just wait. With some sense Mojang will release the API as a server only update meaning people will stay with Bukkit until the API has the mods they wait.
Multi-Threading will be useful as you have stressed in all your posts but once again we have lived without for years. VPSes nearly always have multiple cores and threads. But think of it this way: A four cored Xeon @ 3.6GHz with 32GB of RAM. That's 8GB of RAM per core. Most servers slide between 10-30 players which can be hosted comfortably on 8GB of RAM and a single 3.6GHz core (depending on plugins, mostly this is RAM so the extra cores aren't needed).
Dedicated servers do have multiple cores but remember popularity is about up-time. Right now most of the better servers are on Bukkit so there is no major competition if you upgrade as soon as Bukkit is released. With the new API servers running Spout based servers will loose huge amounts of users (and income from donations and such) each update because the updated API servers have all the new features while Spout is updated.
Also I take back what I said:
Spout seems to be going right from scratch (non of Mojangs code).
This could result in quick updates in some cases and really really long ones in other cases (if Notch adds some awesome new things which involve piles of code Spout will have to work out how he did it and code it before they can update).
So how many threads does it use and how much load is put on them?
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Thanks for the help.
-Donator on over 10 servers and 3 sites (My philosophy: "Don't be cheap, donate to Minecraft communities")
Not the Spout plugin, the Spout project is writing their own server from scratch, without using any Notchcode in the process. If the rumors are true, it'll end up being the unofficial replacement for Bukkit, and it should support multithreading.
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
-Donator on over 10 servers and 3 sites (My philosophy: "Don't be cheap, donate to Minecraft communities")
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Thinking as a big admin here:
Apart from Multi-threading what is in it for me? Open source isn't a big benefit. And it definitely does not make up for having to wait for updates when Mojang's version will come out in sync with the client.
I love the idea of Spout's client server relationship but it would be far better suited as part of Minecraft itself- having things download to the client = cool. Having to download something to get it to work = sadface.
NB: Has Spout had any plans to work with Forge? If they worked together then it would be awesome (as Forge mods don't modify the actual Jar file they can be added on in runtime I guess).
Since the server is being written from scratch, they claim updates will be fast.
Multithreading is a huge advantage.
Open source and open community allows for them to do things and make mods that the Mod API would never allow. In addition, they claim that it'll support legacy bukkit mods, unlike the API, which isn't going to provide this type of transition path.
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
Lol. This ^ every server owner with a shred of common sense.
Sometimes you just have to drop legacy support to move on. On and by the way Bukkit was written from scratch and its updates were behind in some cases by weeks other times they were pretty fast.
Open Source is an interesting point but it isn't a huge benefit considering people decompile MC anyway and sometimes there stuff gets added in (unless it is released under an annoying license like GNU/GPL that won't let the coders give it to Mojang). Multi Threading is an advantage, but how many threads will it use and how well will it use them? Also most servers (as in nearly every one of the 24/7 servers) are on VPSes where they only get a share of a CPU (which is normally less than the power of a core). In these cases Multi-threading has no advantages.
All Bukkit plugins will break with the new mod API, while with Spout you'll be able to use most of them as-is until Spout versions are released.
The transition from Bukkit to the Mod API will be chaos, possibly worse than the hMod to Bukkit transition.
The Bukkit project is basically giving the middle finger to their developers, by merging with vanilla and abandoning the Bukkit API.
Multithreading is extremely useful for larger servers. Quite a few VPS and dedis have multiple cores, and being able to spread the load across multiple threads is a great feature.
Spout will allow developers much greater freedom with creating plugins, and they'll be able to propose changes to the API itself, unlike with vanilla.
"Never try to teach a pig to sing; it wastes your time and it annoys the pig." — Robert Heinlein
As a developer and a (private) server admin I can say that I can't wait for Bukkit's move. Developers will have to re code their mods but hopefully it won't be to painful. But after that there will be no more waiting for Bukkit to update (which sucks) which is a huge benefit. Lots of servers with mods like IC2 run vanilla software with just to try and get the "update time" shortened (NB: They have to wait for extra things to have IC2 on Bukkit).
Chaos? I think it could be relatively smooth, unless the Spout Server is out (I can only find the Bukkit plugin) all of the developers will just wait. With some sense Mojang will release the API as a server only update meaning people will stay with Bukkit until the API has the mods they wait.
Multi-Threading will be useful as you have stressed in all your posts but once again we have lived without for years. VPSes nearly always have multiple cores and threads. But think of it this way: A four cored Xeon @ 3.6GHz with 32GB of RAM. That's 8GB of RAM per core. Most servers slide between 10-30 players which can be hosted comfortably on 8GB of RAM and a single 3.6GHz core (depending on plugins, mostly this is RAM so the extra cores aren't needed).
Dedicated servers do have multiple cores but remember popularity is about up-time. Right now most of the better servers are on Bukkit so there is no major competition if you upgrade as soon as Bukkit is released. With the new API servers running Spout based servers will loose huge amounts of users (and income from donations and such) each update because the updated API servers have all the new features while Spout is updated.
Also I take back what I said:
Spout seems to be going right from scratch (non of Mojangs code).
This could result in quick updates in some cases and really really long ones in other cases (if Notch adds some awesome new things which involve piles of code Spout will have to work out how he did it and code it before they can update).