64 bit amd?

[Fixion]

Originally posted by Buff



lololololol

please dont spread such mis-information.

64bit athlons are hella fast because of the design. When 64bit WinXP is released, they will get ~20% faster. The fastest runs at 2.4ghz, and is faster then both the 3.4EE from intel, and the fastest mac. If anybody wants to tell me that their G5 dual is faster, the Opteron 248 is faster. The fastest mac is 2.0ghz(I think). The A64 and Apple G are very similar in performance clock for clock, so thats why a 2.2A64>2.0G5

and thats the truth!


o, one last thing, Apple is a freaking lying company. if u believe their publications...well, theres not much hope for u:P

This man speaks the truth.

[Fixion]

Originally posted by Rumble
well give it about 6 months, prosessor speeds are doubled every half year or so. In 6 months it should be around 7GHz, based on trends.

Thats Moore's (sp?) law. It became inaccurate a few years ago.

[Fixion]

Originally posted by Miscue
What's kinda funny is, 64-bit code can be slower than 32-bit code... if everything else is the same. The biggest advantage to 64-bit machines is 64-bit memory addressing, so you can have over 4GBs of memory.

Not quite. You could have more than 4gigs of ram with a 32 bit setup. BUT you had to have an OS and hardware that supported.

The Intel PAE (Physical Address Extension) allows up to 64gigs of physical memory to be used. Although only 4gigs of it can be addressed at one time.

Some Linux kernel notes:

http://www.uwsg.iu.edu/hypermail/linux/kernel/0105.1/1169.html

[Bluestrike_2]

Note: I am in NO way defending Apple's marketing schemes, as I find the "Switch" campaign rather repulsive.

Any company will twist whatever they can in an effort to make their product look better. I am sure that the same can be said for the majority of companies. Yet, I find that the Apple products are ingeniously engineered, IMHO. The Cinema 23" is gorgeous(yes, it is properly calibrated every 2 weeks) and everything I own from them is superb. Now, it is not for everyone, but I have no need to get into this kind of argument, as it serves absolutly NO purpose whatsoever. Whatever works best....

Now, I agree with the don't trust Apple's publications, but I take it one step closer. I don't trust any of the manufacturer publications....

When a company says they are better(CONCAST is a great example)I just shake my head and laugh.

Like with cable and satellite.

Comcast : Satellite sucks! When we sneezed, we lost the picture. Blah, Blah Blah.

Nevermind the fact that rain fade and such is extremely rare. I have DirecTV. Needless to say, the most it went out was for one minute during a crazy storm. That was it, except for a couple lost frames during a heavy rain storm. Nothing Major. Nothing New. All companies have an agenda.

[Miscue]

Originally posted by Buff
\

its possible, but generaly its faster. It also depends on who codes it:P A64's get around 20% faster on WinXP-64, so even though the Fx-53is the fastest in 32bit processing, its even faster in 64bit.

Explain the technical reasons for why it is faster.

Pure 64-bit CPUs fill up cache twice as fast as 32-bit CPUs. Can we say... way more cache misses?

Mind you, I mentioned "pure" 64-bit and not modern hybrids that can have 32-bit data types and such (They use them because it is faster to use both!). Everyone is saying it is faster "because it is 64-bit." This is inaccurate, and there's more to it than that.

[Miscue]

Originally posted by Fixion


Not quite. You could have more than 4gigs of ram with a 32 bit setup. BUT you had to have an OS and hardware that supported.

The Intel PAE (Physical Address Extension) allows up to 64gigs of physical memory to be used. Although only 4gigs of it can be addressed at one time.

Some Linux kernel notes:
http://www.uwsg.iu.edu/hypermail/lin...05.1/1169.html

Of course, you can create a scheme that allows you to use 20 terabytes of memory with an 8-bit CPU... but overextending the addressing ability of a CPU creates a lot of overhead. PAE and such is useful in special cases when it is more desirable to have a lot more program memory, and is more important than the overhead created.

This "exception" you're talking about is kind of cheating. It's like saying you have a car that can carry 10 people like a van, but the disclaimer is that you have to make two trips. It's not the same thing.

[Fixion]

Originally posted by Miscue


Of course, you can create a scheme that allows you to use 20 terabytes of memory with an 8-bit CPU... but overextending the addressing ability of a CPU creates a lot of overhead. PAE and such is useful in special cases when it is more desirable to have a lot more program memory, and is more important than the overhead created.

This "exception" you're talking about is kind of cheating. It's like saying you have a car that can carry 10 people like a van, but the disclaimer is that you have to make two trips. It's not the same thing.

This isn't a scheme. Its an extention to the addressing instruction set of processors. If you want to make it be able to access more, lets say 20 terrabytes, you would have to redesign the processor and create an instruction set (or extentions to current ones) that can do that. So it doesn't really count as cheating. Ram paging is something that is built into processors.

The "special hardware" I was talking about are memory riser boards (to motherboards that support them). But thats if you can't come by 4 1 gig sticks of ram and you don't have a mobo that has 4 ram slots.

This selection turns Intel PAE (Physical Address Extension)
+ mode on. PAE implements 3-level paging on IA32 processors. PAE is
+ fully supported by Linux and PAE mode is implemented on all recent
+ Intel processors (Pentium Pro and better).

EDIT: Not like this maters, not many people have more than 4 gigs of ram.

[Buff]

Explain the technical reasons for why it is faster.

Pure 64-bit CPUs fill up cache twice as fast as 32-bit CPUs. Can we say... way more cache misses?

Mind you, I mentioned "pure" 64-bit and not modern hybrids that can have 32-bit data types and such (They use them because it is faster to use both!). Everyone is saying it is faster "because it is 64-bit." This is inaccurate, and there's more to it than that.

why is it faster? uhm....they took the Athlon core, updated it, put the memory controller on-chip(very LL), gave it dual-channel ram, gave it a 1.6ghz HT bus, and a few other things, and then 64bit extensions. All I know is that under WinXP-64 Beta, it encodes media stuff, etc 20% faster then in WinXP-32. I dont own one yet though....So thats a rough estimate. I guess i'll learn more later this year when I start my education as a computer engineer

o ya, talking about ram, I saw an Opteron 2xx board that had room for 16x2gig ram slots....
wow. might make a nice rig for HL2.....

[Buff]

Now, I agree with the don't trust Apple's publications, but I take it one step closer. I don't trust any of the manufacturer publications....

I think Nvidia lies even more though.....
I normally trust benchmarks I get from xtremesystems.org/forums members, and oc info from them, before I buy.... saved me alot of cash over the years.

[Automaggin2]

Originally posted by Buff


I think Nvidia lies even more though.....
I normally trust benchmarks I get from xtremesystems.org/forums members, and oc info from them, before I buy.... saved me alot of cash over the years.

I trust only tomshardware.com

[Bluestrike_2]

As sad as it may be, everyone has some sort of bias in the computer world. Even someone who never used one. Because then they migh think a computer is taboo

[Buff]

I trust only tomshardware.com

uhm, thats almost as bad as apple or nvidia. go to xtremesystems.org/forums (home of the world records for 3dmark2k1 and 2k3), and ask if tomshardware is trustworthy.

[Automaggin2]

Originally posted by Buff


uhm, thats almost as bad as apple or nvidia. go to xtremesystems.org/forums (home of the world records for 3dmark2k1 and 2k3), and ask if tomshardware is trustworthy.

You obvioulsy have no idea what you are talking about then. It is completley unbiased tests.




Apple's most famous tests were run using dual G5's against p4's with hyper threading turned off.

[Buff]

You obvioulsy have no idea what you are talking about then. It is completley unbiased tests.

hehe, I aint talking about biasim, I am talking about noobism I have been building comps since I was 15......
I even set a record for 3DMark2k3(for 9500np) that lasted about 2.5 months

please dont say I have no idea......

[Bluestrike_2]

Only the paranoid truly survive.

Never trust 'em!!!

Ah, hell! While I'm at it, I wanna add my favorite quotes.

"Some call me crazy because I run. I call them crazy because they don't"

"How Do You Run so Fast?" "Well, what do you do when you hear a gun?"

Ok, that was kind of pointless, so onto my real topic.

With technology today, everyone is <B>*POOF*</B> BIASED. Mmmk.... Some do their best to create equal tests, some don't. I agree w/ Buff.

other than a few select PC units, dual G5s are faster. Can't wait for the Crucial Ram comes in, for a grand total of 4gigs!

Course, I'm broke right now. Anyone wanna send some Crucial Ram over this way?

<B>*Warning: No Cussing*</B> - Miscue

Edit: Sorry Miscue

Edit: Typo. Instead of four gigs I meant to type 1.5. I wasn't thinking then. Sorry