Slashdot Mirror
More on 64-bit Gaming
waytoomuchcoffee writes "Valve has announced "immediate availability" of a linux 64-bit dedicated Counterstrike server, designed to run on AMD's upcoming Opteron. This follows on the heels of Unreal Tournament 2003, previously reported on Slashdot. Gamespy has a related story up on a presentation of the future of 64-bit gaming (sponsored by AMD) at last week's Game Developers Conference. As Intel is in no hurry to make the jump to 64-bit desktops, this leaves AMD to court the gaming market."
I agree; for now. There are currently no games that *need* 64 bits for their next version, but why not plan ahead? In the next 10 years gaming will have taken a major jump into the realistic world. Doom 3 starts this jump, even at 32 bits. 4 GB will be required for this jump to continue forward. I want to look into my 3D monitor and see a monster; not a cartoon.
This is a nice way to run a benchmark on a new chip. In a straight port of code highly optimized for x86-32, Counter-Strike dedicated server tests with both 32- and 64-bit versions revealed a 30% clock-for-clock gain, and is expected to show further performance gains in future upgrades. Its also nice to see Valve putting in the effort to keep it customers happy and running at high speeds. www.valvesoftware.com Operon web page at AMD
Motherboard in my IBM NetVista computer (1.6GHz, 512MB) just died yesterday after playing new IL-2 Sturmovik FB for couple hours. I am sure, with currently fastest available 32-bit computers you will experience some kind of slowness if you, let say, increase resolution or fly aircraft over complex terrain or with many AI objects in the scene. Therefore, biggest limitation of 32-bit computing has come to desktop due to very sophisticated games and simulations. Intel is downplaying desktop 64-bit computing because the only solution they have requires complete rewrite of all applications, which very few can afford. AMD has great chance with Opteron to take the lead and shake Intel Empire.
Does anyone know what the story is with Valve? They worked for years and then came out with Half-Life, a huge monster hit. They re-wrote major portions of the Quake engine to do it! They had huge frickin levels! This wasn't two guys in a garage somewhere.
And then... lots of nothing. Half-Life: Opposing Force and Half-Life: Blue Shift were done by Gearbox. Counter-strike was mostly done by people outside Valve.
Did Valve lay off most of the people who worked on Half-Life? If not, how are they paying all those salaries? Does Valve have any actual projects in the pipeline (little pun there) or is Team Fortress 2 all they are working on? (And does anyone think TF2 will actually ever ship?)
For that matter, who the heck is Gearbox? Was it spun off in some way from Valve, or is it something else? How big is Gearbox?
I think that something really bad must have happened to Valve. But I don't have any idea what it was.
steveha
lf(1): it's like ls(1) but sorts filenames by extension, tersely
How much of the 30% gain was due to the jump to 64 bits and how much was due to improved compilers, unrelated improvements to the CPU architecture, higher clock speed, etc?
you could use 64-bit integers to repleace floating point numbers
Sure, but why?
You wouldn't have the same accuracy that you would get from using floats
Correct...
it's a nice middle-ground between 32-bit ints, floats, and double-precision floats
Absolutely wrong. Anything that is going to be scaled or rotated (think matrix transformations, vector multiplication, and physics calculations) needs to have floating point representation, unless the processor architecture is incapable of it (the Gameboy Advance or other embedded platforms, say). You can have pseudo floating point with ints (last x bits are behind the decimal, say), but the software needs to do extra work: there's special cases for multiplication, etc. Most processors these days are designed to do floating point operations as fast as anything else, so adding unnecessary overhead that ignores basic funtionality would be stupid.
If you have a 64-bit cpu, use 64-bit floats instead of 32. 64-bit floats are superior for games because of increased dynamic range for lighting (less color banding with scaling) and other values, and of course less floating point round-off errors (less matrix drift).
Game servers do no graphical calculations, so there's no advantage there, but some of the physics could be done centrally to prevent the clients from making up their own and cheating.