Hanabal

Is it worth it to get a 64 bit chip yet for my games pc, as the next time i can upgrade will not be for 2 years at least, or should i wait till later this year,

On another completely unrelated note: Is the OEM half life 2 coming with radeon 9800's the full game or single player only.

__________________
I am Hanabal, Phear my elephants

Lasereth

I'd wait for the 64-bit chips. Check out my Athlon 64 thread if you want more info. The prices will be considerably lower by the end of this year.

-Lasereth

__________________
"A Darwinian attacks his theory, seeking to find flaws. An ID believer defends his theory, seeking to conceal flaws." -Roger Ebert

meff

Try out an Opteron if you have money, they can run both code and they're pretty fast too, but the clock speeds have yet to reach those of the top of the line AMD XP/MP and P4 procs.

__________________
Toy-like people make me boy-like.

Lasereth

They don't have to if they're twice as powerful. The 2.0 GHz ClawHammer is proving to be as powerful as a overclocked P4 at about 4 GHz.

-Lasereth

__________________
"A Darwinian attacks his theory, seeking to find flaws. An ID believer defends his theory, seeking to conceal flaws." -Roger Ebert

juanvaldes

If you have to ask no.

oberon

/RANT:
Repeat after me: Clock speed does not necessarily reflect performance.

Ok, that said, to clarify... clock speed is only part of the equation. The *real* metric is execution time. And execution time = instructions per program * clock cycles per instruction * seconds per clock cycle.

You see, the clock rate is only part of the story. If your CPU needs fewer clock cycles per instruction, it will be faster than a CPU with a higher clock rate, assuming they need the same number of instructions per program.
/END RANT

Ok, that said, 64-bit CPUs only make sense in a few areas: bandwidth (I/O, CPU, and memory), and memory size. So unless you have software that will take advantage of one of these, you won't get much out of a 64-bit CPU, especially for your money.

Personally, I'm building a cheap Athlon XP box for now and I'll turn it into my fileserver next year when I get an Opteron.

__________________
"There are finer fish in the sea than have ever been caught." -- Irish proverb

charliex

oberon, not quite, modern processors have multiple pipelines and hardly anyone uses clocks per instructions timing anymore since you can't easily predict it (not a PC , since P3) since you generally don't know the answer till it runs and by sampling it, and even then its not always the same.

the second part doesn't quite work either as its very unlikely you'd have a machine with the same style of CPU that had less cycles per instruction compared against a CPU with the same instruction set with more cycles at a higher clock speed.

say a RISC Vs CISC a lower cycle count for a RISC (generally 1 ) requires more instructions, but it can be beat out by a CISC with a higher number of cycles per instruction and a higher clock speed, which is typical these days.

Your formula for execution time is way off, at the very least its 6 years out of date. , i'm not even sure i understand it whats "seconds per clock cycle"?

[edit] i think i might see what you mean by seconds per clock cycle, how long it takes to execute one cycle ? , which would be a . followed by a lof of zeros and a number, but either way its still way off.

Mad_Gecko

Can you catch a bus? Thats how fast we can send info.......

I think faster than you, but you send information faster than me, Who wins?

__________________
[color=bright blue]W[/color]e Stick To Glass

"If three of us travel together, I shall find two teachers."
Confucious

Hanabal

umm yes, Walks slowly away avoiding eye contact

__________________
I am Hanabal, Phear my elephants

oberon

It's from my book, "Computer Organization & Design", published in 1998, which discusses pipelining in chapter 5 (pipelining first appeared in 1985 in the 80486 & MIPS R2000 CPUs). It's a little old, but I think the principle still applies. Sure "seconds per clock cycle" is a pretty small number and would be better represented inversely. But then it wouldn't work in the formula.

You're right about pipelining though, it does complicate the matter a little bit. My studies of this subject are incomplete, but I think my original point remains intact: clock cycles tell only part of the story.

__________________
"There are finer fish in the sea than have ever been caught." -- Irish proverb

Lasereth

You don't really need any other explanation than running a benchmark test on a 1.8 GHz Pentium 4 and a 1.8 GHz Athlon XP. A PC with the same specs except for the processor would differ drastically on benchmark scores. I'd say the Athlon XP would get about 3,000-5,000 points higher on 3d Mark 2001 easily. Hell, run SiSoft Sandra with an Athlon XP if you want more proof. A 1.6 GHz Athlon XP Palomino core is the equivalent of a 2.2 GHz 400 MHz FSB Pentium 4. Clock speed means what it says: clock speed. Unfortunately, programs/games/applications today use more than simply clock cycles. That's apparent after benchmarking identical PCs sans the same processors.

-Lasereth

__________________
"A Darwinian attacks his theory, seeking to find flaws. An ID believer defends his theory, seeking to conceal flaws." -Roger Ebert

merkerguitars

In my opinion no....definately not worth the cost...sure if i could rip dvd's at 1/4 of the speed and every game would run smooth as silk i would say yes.

__________________
Donate now! Ask me How!

Please use the search function it is your friend.

Look at my mustang please feel free to comment!

http://www.tfproject.org/tfp/showthread.php?t=26985

charliex

oberon, i was pretty close then with the 6 years and pipelined cpus have been around since at least 1962.

while its true that clock speed isn't the global benchmark in PC's it isn't as simple as it appears, and its well beyond the laymans abilities to understand it, which is why it makes such a good marketing ploy. With multiple pipelines, out of order execution, prediction and caching its a phenomenally complex model.

You can sort of compare clock speeds if you normalize them into the same space first, which is approximately what AMD did with their ratings, the advertised clock speed is the normalized rating against intels, they did this because they knew the consumer and marketing depts. would look at clock speeds as a comparison, so they did what they could to show the benefits of their more bang for the buck with a normalized ghz/mhz rating

A 5 GHZ p4 will outrun a 3 GZ P4 , even if the only thing that changes is the clock speed, so ergo the clock speed is important, but as with any statistic it is important to make sure that the scale used is correct.

Anyway with a 64 bit CPU the biggest gain is generally the width of the data transferred, as well as domain size, however alot of systems already transfer at 64/128 bits.

Theres no guarantees a 64 bit system will be faster than a 32 bit one, it may even be slower, it can just be setup to give you the larger domain for a speed penalty.

Its all very complex, and even generic benchmarking really only tells you about that generic benchmark, its easy to make your chip look better since it is such a complex system, and theres a lot of snake oil.

Simplest thing is just to get real world benchmark information using the applications and data you would actually use the machine for as opposed to some tabloid newspaper style benchmark such as 3dmark etc, only that will tell you if its better or not.

oberon

charliex:
You're right, it's very complicated. And of course, the bottom line is still execution time of a given real world application, as you said. Screw the details: clock rate, pipelining, branch/jump prediction models, blah blah blah.

Can we agree on that, then? And stop boring the non computer engineers.

__________________
"There are finer fish in the sea than have ever been caught." -- Irish proverb

Mephisto2

charliex never gives up.

Isn't that right mate?..



Mr Mephisto

Lasereth

A bit of news on the subject:

The new 64-bit AMD chip was codenamed ClawHammer and SledgeHammer. The SledgeHammer eventually became the Opteron, which is out right now. The ClawHammer has evolved into the Athlon 64, which is the desktop chip which will debut on September 23rd of this year. So gamers will most likely be interested in the Athlon 64 chip, which is designed nearly identical to the Opteron except on a non-server (or non multi-processing) level.

-Lasereth

__________________
"A Darwinian attacks his theory, seeking to find flaws. An ID believer defends his theory, seeking to conceal flaws." -Roger Ebert

charliex

and there was me thinking this was a discussion board