64-bit client

[Llevar]

Re: Re: Re: Allows me to educate you boobs

Originally posted by Janglur
Correct, on most parts.
I cited /an/ Opteron. There's several typed out there. 220, 420, 820, etc.

And, these ultra-fast Opterons are alost some $2000+ dollars, for the CPU alone.
Who here wants a few extra FPS /that/ badly?


As for the 64-bit data instruction sets, it's true this will become more efficient in time. But seeing as how Half-Life 2 was programmed intiially in 32-bit instructions, to utilize 64-bit instructions would mean entirely redesigning it. I doubt they will go through that trouble for a 5-10% speed increase, especially since they'll not come near using 100% of a high-end 32-bit processor /anyway/.

The 64-bit data isntruction won't become very useful until we've had it for another ten years, same as 32-bit.

Remember: We have 128, and even 256 bit console systems now. But they're mostly still using 32-bit instruction sets!
These extended instruction sets have little place in computers, yet. A few might make a novel save in typing here and there, but one set isn't going to make a game suddenly run twice as fast.

Bitrate evolution and programming evolution are seperate worlds. And the latter moves much, much slower.


Remember: NASA uses mostly 16-bit processors for it's most important controls. Because the instructions, while limited, are reliable and more easily predicted.

I agree entirely that Valve will not have a 64-bit client. It would be a pretty unsound business decision to dedicate thousands of man-hours to recoding a game for a yet unreleased operating system(a Windows one at that, which really means at least two service packs until its semi-reliable) with a slight performance gain for the 0.0000000001% of the user base who will actually have the 64-bit chips in the next couple of years. Coupled with the absence of expertise with the new instruction set results in a big no-no for Valve porting the client to 64-bit based processors.

In my eyes, if anything will make a real breakthrough in the speed with which memory intensive applications run in the next few years it will be a new mass storage technology that will bring these devices to the same order(or even anywhere close) of speed as RAM.

 

[Llevar]

Originally posted by Janglur
64-bit OSs will effectively make 'bubbles' to emulate a 32-bit environment.

Based on the fact that all non-legacy software is coded in 32-bit and is not about to be ported over any time soon I would say you have an accurate forecast of at least the first five years of the 64-bit processors being on the market.

Unless of course the compilers that will come out for the new instruction set will be able to take advantage of the new features without the need to restructure the high level code greatly.

 

[Janglur]

Which is unlikely, at best.

 

[Frost]

Originally posted by Gorgon
increase performance only in business appli.....
for games would increase per only 10% and that is nothing

Actually you are wrong on all accounts. Please do some research or have some hard facts before you go telling someone that there would be no to little increase in performance. From what your saying it sounds like we could run our PC's without a CPU and the GPU is all you need for games LOL! 64-Bit processors combined with 64 Bit coded software will allow around a 30% boost in speed over a 32-bit processor running at the same frequency in Half-Life 2. They have already announced a 64 Bit server for Half-Life and have said somewhere that they will take advantage of 64-Bit processors in Half-Life 2 in some interview, However Gabe has made it seem to me like they won't support it right away because he's unsure of a Windows XP update or new version coming out before the game is released. When asked which brand processor I should buy (Either an AMD 64 Bit or a Intel 3GHz+ w/800 FSB & HT) Gabe suggested I go with an Intel because he was unsure about the Windows release, which would you would need to have in combination with Half-Life 2 coded in 64 bit to get the advantage. 64 Bit processing is the future and it will certainly be okay to buy a 64 Bit processor when they are released for desktops in late August/early September because you will be able to run programs and Windows in 32 Bit mode until they upgrade them to 64 Bit then you will have the technology already to use it right away. I was planning on building a new PC right before Half-Life 2 ships but if they ship on September 30th, then I'm going to have to wait, its just not enough time to figure out which processor and motherboard to go with, not to mention the releases of the new graphics cards from ATI and Nvidia have been pushed back a bit from the 6 month cycle and won't be out till after Half-Life 2 hits shelves and Steam. Maybe if they do delay Half-Life 2 till November 25th then I'll be able to have the system by then, lots of new stuff is coming guys, do a lot of research when spending your hard earned money to play Half-Life 2.

Recommended Reading:

http://www.tomshardware.com
http://www.anandtech.com
http://www.hardocp.com

 

[Janglur]

Now would be a good time to shamelessly plug myself.


I build computers.

Here's some I have so far:

(Dragon Lineup
Drake $287.50 A nice budget PC
Wyvern $517.50 An all-around home PC
Draco $575 A simple home PC for most chores
Wyrm $1069.50 Sadistic, Black, and Sexy.
Trogdor $1,192.55 A low-cost gaming machine
Trogdor2 - The Burninator $1,236.25 Burninating the peasants (CD+DVD burner)
Bahamut $2,541.50 God is NOT going to like that we stole his PC.



Right now, all i've got are AMDs to build. This is because I can buy them at discount, or at bulk. I can't get pentiums for discount or bulk, none of my providers offer them.

I think it's due to them costing so damned much.






Note I build on demand, I do not keep stock of entire PCs. You have to pay, at least, the upfront fee for me to build the PC. (The upfront fee is the total cost of parts). Then pay the comission after you've received the PC (at latest).

 

[FictiousWill]

It would be a pretty unsound business decision to dedicate thousands of man-hours to recoding a game

Wazzat? Wouldn't they just recompile their source code (ascii) with a new 64-bit compiler?

 

[Torsion]

I didnt read this whole thread but only games optimized for a 64 mb cpu would see improvement. Games that arent (todays games) would run the same or even see performance reduction.

 

[qckbeam]

I don't think that you recompile the old 32-bit code in a new 64-bit compiler. You have to go in and write new instructions for the 64-bit CPU to replace the old 32-bit instructions and this would take a long time.

 

[Janglur]

qckbeam is correct.

This is because instruction sets are machine code, not programming code.

 

[Llevar]

Originally posted by FictiousWill
Wazzat? Wouldn't they just recompile their source code (ascii) with a new 64-bit compiler?

As I pointed out before, it is not a fact that major alterations to high level code will not be required to take advantage of the new processor. If a compiler will be made that will allow people to just port stuff from 32-bit to 64-bit it will have to either be really smart(read long development time, probably several years) or not offer much in terms of performance increase. And if it doesn't offer much in terms of performance increase then why not just run the game in a 32-bit emulator on the 64-bit box.

 

[Llevar]

Originally posted by Janglur
qckbeam is correct.

This is because instruction sets are machine code, not programming code.

Why would you say that? Applications are written in high level languages so theoretically a switch from 32-bit to 64-bit could take place entirely on compiler level without changing the high level grammar, afterall it is the compiler that pumps out the instructions not the coders.

 

[Janglur]

But then it would do the same thing. There'd be no performance increase.
In fact, there'd be a decrease, as the commands get larger (64-bits wide), it'd take up more space. Up to twice as much.

Requirements: 1 GB of RAM, 4 GB of free HDD space


We don't want to see that for several years yet.

 

[Llevar]

Originally posted by Janglur
Note I build on demand, I do not keep stock of entire PCs. You have to pay, at least, the upfront fee for me to build the PC. (The upfront fee is the total cost of parts). Then pay the comission after you've received the PC (at latest).

You do that for a living? I did tech support(including new system builds) for about a year and a half a few years back and found it pretty unrewarding, professionally.

 

[Janglur]

Tech Support is, indeed, unrewarding.

But I can make a paltry sum building them.

It's a hobby, i'm not quitting my day job. =3

 

[Llevar]

Originally posted by Janglur
But then it would do the same thing. There'd be no performance increase.
In fact, there'd be a decrease, as the commands get larger (64-bits wide), it'd take up more space. Up to twice as much.

Requirements: 1 GB of RAM, 4 GB of free HDD space


We don't want to see that for several years yet.

Well, most compilers are a lot smarter than just parsing the code and spitting out instructions. They do all kinds of optimizations. Supposing for a second that the current instruction set had no shift instruction but the one coming out in a few months did. You should hope that there would be compilers that would take your current code in c++ or whate have you and generate shifts where possible to replace existing sets of instructions that have the same functionality. The capability to write such compilers is definitely there and they are guaranteed to be made at some point since the existing 32-bit codebase is just too large to recode manually and emulation will surely not work correctly for many applications. I do think however that we won't see them for the next couple of years.

 

[Janglur]

Exactly. We won't see compilers capable of 32-to-64 bit conversions that are worth performing for several years, at least.

 

[Ahnteis]

[laughs at naysayers]

You might want to read what valve has said on the subject. And learn a little bit about computers and programming.