Lag in Uru
There has been a lot of discussion about lag recently, with many people blaming the servers. Allow me to quote a chat that I found with chip. For those who do not remember/know, chip was an intern at Cyan, and he was ultimately the one who worked on the Untìl Uru servers and released them to the public.
<chip> oh man, can we please start arguing about plasma again
<chip> the problemm isn’t plasma2.0 or the plNetCore
<chip> its the fucking physics crap
<chip> it was sloooow
<chip> you put like 30 people on the same screen
<chip> it just goes to all shit
<chip> * same area
<chip> I learned lots at cyan. mostly that you can optimize the backend servers all you want
<chip> but in the end
<chip> a shitty client (in this case in c++ and python)
<chip> can still make the whole thing suck
So there you have it, the servers are not the problem. There were a number of improvements during MOUL (including removing the cones entirely), but for lag to disappear entirely will be a client-side issue and has nothing to do with the servers.
Marten said,
Friday, January 9, 2009 at 17:20
This is actually good news, from my perspective.
A'lan said,
Tuesday, January 13, 2009 at 5:10
Chip speaks the truth. Wasn’t he working on Apache after he left Cyan? Brilliant fellow.
Grogyan said,
Friday, January 16, 2009 at 20:12
Not entirely true, as I found out with first hand experience.
With a modern CPU and graphics card, even with 200 in the same vicinity as me I got fairly decent frame rates of ~15 – 20FPS
Interesting enough my old 7900GTX out perfromed the 8800 GTS 512 I currently own, which dropped the frame rates even lower.
Now that physics are processed on the GPPU (I intentionally put the extra P in) we should see an improvement on the frame rates.
I am sure you have more knowledge than you want to admit to fix a lot of problems in the client
Hoikas said,
Sunday, January 18, 2009 at 15:32
Physics in MOUL are not processed on the PPU or GPU.
Using the PPU or GPU requires explicit permission in the code… And you must OPTIMIZE CAREFULLY or you will end up burning more time than you did CPU cycles.