AMD Opteron at Newegg

[cyberia]

Originally posted by: kof
Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

Had I known that such a bright person as kof belongs to this discussion board, I might have never joined. He is way totally out of my league.

 

[deeznuts]

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

hahahahahahahahahahahaha. *takes a deep breath*
hahahahahahahahahahahhah.

funniest thing i've read in a while. what most of you don't realize is that 64 bit has almost no use for the desktop market. the main advantage of 64 bit processing is the ability to address more than 4 GB of memory. that's a few years away still, with 512 being the de facto standard right now.

and do you realize that running a process on 64 bit makes it slower than running that process on a 32 bit platform? due to larger addresses.

 

[merlocka]

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

bwAHAHhahah

gosh, I'm hoping this guy was being sarcastic.

I like the term "pushing bits" though. Makes it sound all macho.

 

[damonpip]

Yeah, I wish it was that easy, just double the bits to double the speed

 

[Swanny]

Newegg cracks me up sometimes. Look at the product name.

AMD Athlon Opteron Model 240 1.4GHz 64-Bit PROCESSOR

Notice: Athlon Opteron. It's just the Opteron.

Oh well, I sent them an email so it should get fixed. This is about the 4th email I've sent them on product naming. I keep hoping they'll give me something free for correcting their mistakes

 

[dew042]

Originally posted by: merlocka

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

bwAHAHhahah

gosh, I'm hoping this guy was being sarcastic.

I like the term "pushing bits" though. Makes it sound all macho.

sounds like a drug dealer - i mean "pushing" 64 bits and all

dew.

 

[frankqfrank]

This thread is getting really funny.

 

[oddjob303]

Originally posted by: lobadobadingdong
That's all fine and dandy, but where do you get motherboards, and heatsinks for these? SuSE Linux has an opteron based kernal for it already for 64-bit mode, and it won't be until Octobre or so before windowsXP-64 comes out.

I dont know about the motherboards but the zalman cnps 7000 is supposed to support these puppies, wish i could hear those zalmans in person.

 

[pspada]

It's really not about the clock speed, per se. The big deal about this 64-bit chip is that it is fully 32-bit compatible. Anotherwords today you run your current version of Winblows on it and get better performance than a Athlon processor, but tomorrow you run the new 64-bit version of Winblows on the same system. This is not the case for the Intel 64-bit processors, which require a new 64-bit OS as well and new 64-bit apps. The lack of an installed base to leapfrog Intel's small 64-bit following to critical market mass is a severe determent to the market growth of these Intel chips. This should provide an excellent chance for AMD to lead the market. In fact, I suspect that Intel will be forced to come out with it's own 32-bit compatible 64-bit processor, which will eventually completely replace the current 64-bit line. An interesting effect of AMD being first to have this capability will be that Intel's chips will have to be compatible with the original AMD chips, and I love the irony of Intel being forced to come out with a clone of an AMD chip.

Oh btw, I have a regular "thingie", not a "math thingie".

 

[dexvx]

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

Funny math makes me laugh. You multiplied Mhz by bits, your units are Mhz-bits, which makes absolutely no sense.

 

[frankqfrank]

Originally posted by: dexvx

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

Funny math makes me laugh. You multiplied Mhz by bits, your units are Mhz-bits, which makes absolutely no sense.

well i agree his calculations are nonsensical, but Mhz * bits is just bits/second (without constant), not really a nonsensical unit, not right, but not nonsensical (I never could've imagined using nonsensical three times in a sentence.)

I don't feel like trying to explain stuff, alot of people have some incorrect assumptions about 64 bit ISA in general

 

[ViperV990]

MHz * bit = Mbit/s

...

 

[ivan2]

i dont think he's totally wrong, for Opteron i think it uses 32bit ALUs internally (emulate or whatever u call it) to carry out the 64bit instruction's work, which means that its pipeline is configured such that it will run 32bit application in a lower CPI than 64bit which may take longer pipeline or more cycle to complete (its more complex and with 32bit ALUs doing the job).
it should be taken like this, 64bit=more complex instructions=more CPI(for the 64bit instructions) but also it means more stage pipelines and superscaler are putting into the cpu to achieve a reasonable performance for its low clock rate of 1.4Ghz, all of these pipelining and superscaling will result in a much faster CPI for 32bit instructions which it runs internally.
To complete his statement he just needs to argue that internally around a average of 2 32bit instruction's time is needed to process a 64bit instruction, which im afraid if someone will do the search for me in AMD they will find the spec(its 4oclock and im sleepy now). if not then we have to wait and see the 64bit benchmarks to tell.

and btw this is EE and CS thingy, nothing related to complex math. i think his a noob too but the statement might hold someday if the someone can find out the CPI and other stuff. and im thinking 64bit processor should match with 64bit processor. Just like you match a Crusoe with a Pentium M, but you never match it with a SPARC that's running the similar RISC code. What they are made for is what it means.

Please input ur comment.

 

[ivan2]

edit: double posted, damn the network

 

[Mrburns2007]

So your saying that a Athlon 64 with Windows XP 64 running a 64 bit app won't be any faster then all 32 bit........I was under the impression there was more to it then just addressing more memory. I had heard that it could give a 15% boost.

 

[kidsafe]

Wow...the stink of FUD....

 

[kof]

Originally posted by: ivan2
i dont think he's totally wrong, for Opteron i think it uses 32bit ALUs internally (emulate or whatever u call it) to carry out the 64bit instruction's work, which means that its pipeline is configured such that it will run 32bit application in a lower CPI than 64bit which may take longer pipeline or more cycle to complete (its more complex and with 32bit ALUs doing the job).
it should be taken like this, 64bit=more complex instructions=more CPI(for the 64bit instructions) but also it means more stage pipelines and superscaler are putting into the cpu to achieve a reasonable performance for its low clock rate of 1.4Ghz, all of these pipelining and superscaling will result in a much faster CPI for 32bit instructions which it runs internally.
To complete his statement he just needs to argue that internally around a average of 2 32bit instruction's time is needed to process a 64bit instruction, which im afraid if someone will do the search for me in AMD they will find the spec(its 4oclock and im sleepy now). if not then we have to wait and see the 64bit benchmarks to tell.

and btw this is EE and CS thingy, nothing related to complex math. i think his a noob too but the statement might hold someday if the someone can find out the CPI and other stuff. and im thinking 64bit processor should match with 64bit processor. Just like you match a Crusoe with a Pentium M, but you never match it with a SPARC that's running the similar RISC code. What they are made for is what it means.

Please input ur comment.

To complete my statement look at

graphs such as this
and quotes such as this

Itanium 2 POVRay Benchmark Results (CPU)
By Brian Neal
Friday, April 18, 2003 10:50 PM EDT
According to these benchmark results, the Itanium 2 is turning in some impressive times for the freely available POVRay raytracer. At 00:11:07, the 1 GHz Itanium 2 is better than twice as fast as its closest listed competitor, an overclocked Athlon XP running at 2.29 GHz (00:26:07). The highest-ranked Pentium 4, a Northwood overclocked to 3.24 GHz (154 MHz FSB), achieves a result of 00:28:31.

from here .
That's a 1Ghz 64 bit Itanium 2 versus 2,29 Ghz, 32 bit XP.

Of course it's also a $3000 processor versus a $150 processor.

Sytem design is critical. If you run your systems under DOS (if it's possible) with 16Mb of memory, one 5.25" floppy and only use edit you probably won't see a lot of improvement. If you've the right 64 bit apps to take advantage of a fully tricked out system the 64 bit performace does show up at ballpark twice the processing power.

 

[El Norm]

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

you for got to factor in other stuff, look its really easy ((1.8*64)*400bus*8xAGP*2DDR Memory*USB2.0*52x CDR drive*100Herts Refresh rate+wireless rechargable mouse)/(Windows xp*amount of programs installed))= MAN THAT IS CRAZY FAST!!

looks complicated but if u failed as many math classes as i did it should make sense:beer:

 

[cameronj]

Originally posted by: kof

Originally posted by: Mrburns2007
1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

That would be because you made it up, and that calculation has nothing to do with how performance actually works...

 

[cyberia]

Originally posted by: kof

Originally posted by: Mrburns2007 1.8 Ghz keeping up with a 3 Ghz cpu is impressive!

Actually, with a properly designed system, 1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ). But I realize that's a "math thingy" many of you might not understand.

What I am concerned more with is the tone in which kof made his statement - regardless of how correct or incorrect he was.

Kof, can you provide a reference to support your statement (not the benchmark links that show an outcome not necessarily substantiating your theories)?

 

[ivan2]

An freshly design 64bit system and one that runs 32bit internally, which one will you think its more efficient? I think we should all wait until there's some 64bit Operton marks to tell if its the-thing for 64bit world. But again, 250 bux xeon replacement is very good for many.

 

[JDLT1]

So where are the motherboards for thise bad boys?

 

[frankqfrank]

Originally posted by: El Norm


you for got to factor in other stuff, look its really easy ((1.8*64)*400bus*8xAGP*2DDR Memory*USB2.0*52x CDR drive*100Herts Refresh rate+wireless rechargable mouse)/(Windows xp*amount of programs installed))= MAN THAT IS CRAZY FAST!!

looks complicated but if u failed as many math classes as i did it should make sense:beer:

You forgot the coefficient of friction.

 

[deeznuts]

Originally posted by: kof

Originally posted by: ivan2
i dont think he's totally wrong, for Opteron i think it uses 32bit ALUs internally (emulate or whatever u call it) to carry out the 64bit instruction's work, which means that its pipeline is configured such that it will run 32bit application in a lower CPI than 64bit which may take longer pipeline or more cycle to complete (its more complex and with 32bit ALUs doing the job).
it should be taken like this, 64bit=more complex instructions=more CPI(for the 64bit instructions) but also it means more stage pipelines and superscaler are putting into the cpu to achieve a reasonable performance for its low clock rate of 1.4Ghz, all of these pipelining and superscaling will result in a much faster CPI for 32bit instructions which it runs internally.
To complete his statement he just needs to argue that internally around a average of 2 32bit instruction's time is needed to process a 64bit instruction, which im afraid if someone will do the search for me in AMD they will find the spec(its 4oclock and im sleepy now). if not then we have to wait and see the 64bit benchmarks to tell.

and btw this is EE and CS thingy, nothing related to complex math. i think his a noob too but the statement might hold someday if the someone can find out the CPI and other stuff. and im thinking 64bit processor should match with 64bit processor. Just like you match a Crusoe with a Pentium M, but you never match it with a SPARC that's running the similar RISC code. What they are made for is what it means.

Please input ur comment.

To complete my statement look at

graphs such as this
and quotes such as this

Itanium 2 POVRay Benchmark Results (CPU)
By Brian Neal
Friday, April 18, 2003 10:50 PM EDT
According to these benchmark results, the Itanium 2 is turning in some impressive times for the freely available POVRay raytracer. At 00:11:07, the 1 GHz Itanium 2 is better than twice as fast as its closest listed competitor, an overclocked Athlon XP running at 2.29 GHz (00:26:07). The highest-ranked Pentium 4, a Northwood overclocked to 3.24 GHz (154 MHz FSB), achieves a result of 00:28:31.

from here .
That's a 1Ghz 64 bit Itanium 2 versus 2,29 Ghz, 32 bit XP.

Of course it's also a $3000 processor versus a $150 processor.

Sytem design is critical. If you run your systems under DOS (if it's possible) with 16Mb of memory, one 5.25" floppy and only use edit you probably won't see a lot of improvement. If you've the right 64 bit apps to take advantage of a fully tricked out system the 64 bit performace does show up at ballpark twice the processing power.

that doesn't mean anything. it just means the itanium architecture is more efficient then amd's or even intel's. this has nothing to do with 64 bit vs. 32 bit as stated in your original argument.

just to show how bad and flawed your argument is, i'll use your original calculation.

comparing P4 vs. XP.

1.8Ghz pushing 64 bits should outdo 3 Ghz pushing 32 bits by 20% ( (1.8*64)/(3*32) ).

so with a P4 at (3.24*32) vs. XP at (2.29*32), the P4 should be performing 30% faster, but it's actually slower.

don't you see, the 64 bit vs. 32 bit has almost no bearing on performance in the context of what is being discussed here. the difference in performance you see is due to architectural designs (like P4 vs. XP) not 64 bit vs. 32 bit.

read this:

http://www.anandtech.com/guides/viewfaq.html?i=112

about 64 bit taking over and intel having to license the technology ... not likely. with intel's market share and force, they could probably design their own, and still take over. there's a lot of people who will not buy amd even if their life depended on it. not smart, but true.

 

[kof]

that doesn't mean anything. it just means the itanium architecture is more efficient then amd's or even intel's. this has nothing to do with 64 bit vs. 32 bit as stated in your original argument.

Duh - itanium is INTEL