Latest news
The arrows code is done, both on the client and on the server. While we still need some more tests, and probably various fine tunning, everything works fine right now.
Xaphier is working at a vertex shader rendered for the cal3d animations, and he is making good progress. There are still various issues with his code, but once it is done, there will be a significant frame rate increase, especially for people with good videocards. The idea is to move some of the CPU workload to the GPU, for two reasons:
1. Sometimes videocards can do the animation faster than the CPU can.
2. Sometimes the CPU is busy with other stuff (such as compressing video, compiling a program, rendering images, etc.), so by having some workload switched to the GPU, you can play the game at a decent frame rate even while the CPU is busy with other stuff.
A big German magazine (c't) will include EL on their DVD a the end of the month. They will also review the game, and have an article about it.
Now, this is a big magazine (400K readers), and if even 1% of their readers decide to try the game at the same time (like, say, the day it is published), we will have some problems handling so many users.
I am going to increase the server users limit to 2.5K, and hope that there will be no issues. I estimate that our current game server should be able to handle over 5K people at the same time, but the CPU load does not increase linearly with the number of users, so we'll just have to see what happens.
The web server, on the other hand, might not be so lucky, so it is quite possible that for a few days after that magazine is published the website will be very slow, or even unreachable at all.
But of course, those small problems are worth it, it's not often that you can get your game exposed to almost half a milion people (and probably even more if the article will be published on the website too).
Xaphier is working at a vertex shader rendered for the cal3d animations, and he is making good progress. There are still various issues with his code, but once it is done, there will be a significant frame rate increase, especially for people with good videocards. The idea is to move some of the CPU workload to the GPU, for two reasons:
1. Sometimes videocards can do the animation faster than the CPU can.
2. Sometimes the CPU is busy with other stuff (such as compressing video, compiling a program, rendering images, etc.), so by having some workload switched to the GPU, you can play the game at a decent frame rate even while the CPU is busy with other stuff.
A big German magazine (c't) will include EL on their DVD a the end of the month. They will also review the game, and have an article about it.
Now, this is a big magazine (400K readers), and if even 1% of their readers decide to try the game at the same time (like, say, the day it is published), we will have some problems handling so many users.
I am going to increase the server users limit to 2.5K, and hope that there will be no issues. I estimate that our current game server should be able to handle over 5K people at the same time, but the CPU load does not increase linearly with the number of users, so we'll just have to see what happens.
The web server, on the other hand, might not be so lucky, so it is quite possible that for a few days after that magazine is published the website will be very slow, or even unreachable at all.
But of course, those small problems are worth it, it's not often that you can get your game exposed to almost half a milion people (and probably even more if the article will be published on the website too).
4 Comments:
Gratz on the magazine deal, I might have to learn German now.
Thats cool, congratulations.
But I would propose to expand the webserver. If the news is comming also on heise.de you can expect that the server will go down ;) It would not be the first time.
greetings V
Yes, I heard that before, and I think it is possible that the webserver will go down.
However, the brief nature of this event (a few days at most) can't justify buying another webserver.
Maybe after all the stress testing the player database could be cleaned, so very old not developed character names would be freed. I know you do that from time to time, but it is just a suggestion.
Post a Comment
<< Home