Dual Core Vs 64 bit

Which is better

  • Dual core

    Votes: 20 83.3%
  • 64 bit

    Votes: 4 16.7%

  • Total voters
    24
Status
Not open for further replies.

darklord

Cyborg Agent
saurav29 said:
lol :) We should have a poll about how Many people are using 64 Bit applications and some one please explain me or not what is the use of 64 bit processing CPU in presence of 32 bit binaries and (Mind you these 64 bit binaries if some can be ported to 32 bits of course with some overheads) Intel sites says that it is future complaint as with more and more 64 bit binaries it will use the resources effectively... BUT i have serious reservations for putting my money in future proofing when they (Read Hardwares Manufacturers) change the specs so rapidly that by the end of the year there are newer plaforms to work with and affordable too..... Most of the people i knew upgrade computer atleast once in an year...... so what the HECK we are fighting about...........
PLZ GUYS post some constructive thoughts rather than who is right and who is wrong and who is dumb and who is smart.......

PS GUYS Get over yourself NO body copies architecture from anybody....... for amd and intel what the bloggers may write use your head... AMD executes 10 instructions per cycle and Intel 8 thats why the true speed of AMD is always less than that of intel what they say is that a 2000+ AMD procy running on ~1800 will excute the same no of instructions executed by Intel running at 2000.... Infact Intel is considered Defacto of Computing standard and AMD rates its cpu on basis of that only....

Please do not think i support or oppose intel i use AMD 3000+ and what i am saying is only on basis of what i know.... No prefrences of AMD or INtel...
Peace
Raj
PS GUYS Get over yourself NO body copies architecture from anybody.------> Nobody claimed of copying an entire architecture,just that the AMD64 Instruction set was copied by Intel from AMD,not the entire architecture.;)
Infact Intel is considered Defacto of Computing standard and AMD rates its cpu on basis of that only.... -------> Heh ? Dude you on crack or something ? It was always AMD who used Product Rating and it was Intel who used to flaunt their clock speeds.Infact Intel clearly mentioned that the GHz race is useless and that they would be 'Rating' the cpus henceforth and hence all current chips have a rating.Get your facts right. Lol

BTW people didnt jump to A64 because it was a 64bit chip but because it comprehensively beat Intel's then architecture,Netburst based CPUs. ;)

AMD executes 10 instructions per cycle and Intel 8 thats why the true speed of AMD is always less than that of intel what they say is that a 2000+ AMD procy running on ~1800 will excute the same no of instructions executed by Intel running at 2000.... -----> Care to explain what that means ? My retarded brain just cannot understand anything:D
 

saurav29

Waiting for 3 Miracles...
:D lol Hey darky if you have such a bright knowledge that you tend to refute every post made by other people try posting some light on the architectural aspect rather that quoting others text and saying it is wrong and at the same time not telling what the Heck is right in thy eyes if i may ask???

Nothing personal :D of course........
Some CPU funda's 101 ........
The cpu core consits of cpu register pairs ( Accummulator (AX), Base Index (BX), Counter (CX) Data (DX) Extra segment (EX) (Stack Segment SS), (Instruction register (IX) Program counter (PC) buses and logic circuits mostly NAND based....... (Duh! NAND being the Universal Gate) The importance of cache is that it is on the same speed as that of cpu registers...
You want cpu to perform a function it goes like this ...
Fetch, Decode and Execute... with a fetch execute being overlapped ... to increase performance... The cpu operates in cycles ... The data is loaded into the registers pairs and operated upon and transfered... the PC gives the address of next instruction which is loaded into the cpu registers and then executed...... The bits of these registers signify the bits of the cpu and hence the instruction and the data processed by them .... suppose a cpu is 8 bit then max no of instrction that can be given to it 2^8 and for 64 bit this becomes 2^68 that are a hell lot of instruction the instruction set does not have so many instrcutions heck they do not have even 2^32 no of instruction... what might be they can operate with 64 bit data now which they previously did by placing the LSB and MSB in two diffrent registers... Then there are micro program controlled and RISC based systems we use microprogram controlled that needed decoding so that cpu performs them in terms it understands.....
you might have heard of 32 bit binaries meaning they can be executed by a 32 bit cpu as both the data and instrcution are in 32 bit only... now how does a 64 bit procy does the same simple the registers are 64 bits so MSB is filled with leading zeros and it works by using LSB's only now if a 64 bit binary is there then it needs to place MSB aand LSB in diffrent registers and works with 2 register pairs now for single piece on data (Hence the over head but 32 bit cpu will execute 64 bit binaries it is upto OS to make it understand HOW?? ....Here comes my doubt .. Somebody plz tell me why the Heck Itanium Sucks in performing 32 bit Instructions... Under no circumstances it will make sense)

Now coming to the cpu rating the cpu is never rated on its clock speed it is rated on how many calculations (Excuse my previous word instructions as these instructions are broken into calculation that is more atomic unit)it can perform in that clock speed ....
AMD cpu exectes 10 calculations so AMD 2000+ will perform (1.67Ghz its true speed 1.67*10*1024=17100, and intel 2.0Ghz will perform... 2.0*8*1024=16384 hence the name AMD 2000+ meaning of + symbol it executes more calculations than intel at 2Ghz ( 1GHZ = 1024 HZ)
Now tell me why the Heck AMD not advertise its cpu as 1.67 G and why as 2000+ if intel is not the standard... It might not be official but marketing it does uses the intels cpu speed as comparision ......... I guess this time you will do LOL to perhaps a ROFL HUH???

After i shift into my new flat i will answer all the queries (As i will have more time then) @DL AND all your LOL's too nothing personal but i think that people should understand a cpu and not go on the marketing gimmicks of some HIGH SOUNDING WORDS LIKE SMARTCACHE.. PIPED SIMD AND NETBURST... these are all simple technologies ... yeah it improves performance but not necessarily you use that.........

Peace
Raj
 

darklord

Cyborg Agent
Some CPU funda's 101 ........
The cpu core consits of cpu register pairs ( Accummulator (AX), Base Index (BX), Counter (CX) Data (DX) Extra segment (EX) (Stack Segment SS), (Instruction register (IX) Program counter (PC) buses and logic circuits mostly NAND based....... (Duh! NAND being the Universal Gate) The importance of cache is that it is on the same speed as that of cpu registers...
You want cpu to perform a function it goes like this ...
Fetch, Decode and Execute... with a fetch execute being overlapped ... to increase performance... The cpu operates in cycles ... The data is loaded into the registers pairs and operated upon and transfered... the PC gives the address of next instruction which is loaded into the cpu registers and then executed...... The bits of these registers signify the bits of the cpu and hence the instruction and the data processed by them .... suppose a cpu is 8 bit then max no of instrction that can be given to it 2^8 and for 64 bit this becomes 2^68 that are a hell lot of instruction the instruction set does not have so many instrcutions heck they do not have even 2^32 no of instruction... what might be they can operate with 64 bit data now which they previously did by placing the LSB and MSB in two diffrent registers... Then there are micro program controlled and RISC based systems we use microprogram controlled that needed decoding so that cpu performs them in terms it understands.....
you might have heard of 32 bit binaries meaning they can be executed by a 32 bit cpu as both the data and instrcution are in 32 bit only... now how does a 64 bit procy does the same simple the registers are 64 bits so MSB is filled with leading zeros and it works by using LSB's only now if a 64 bit binary is there then it needs to place MSB aand LSB in diffrent registers and works with 2 register pairs now for single piece on data (Hence the over head but 32 bit cpu will execute 64 bit binaries it is upto OS to make it understand HOW??
Thanks for your informative post,appreciated :)

....Here comes my doubt .. Somebody plz tell me why the Heck Itanium Sucks in performing 32 bit Instructions... Under no circumstances it will make sense)
Well i am not a architectural guru as you but i know some things for sure.Itanium itself was never bad,it was MEANT to be a 64 Bit CPU and apps needed to be recoded to make full utilisation of its potential.Intel had embedded support for 32Bit apps but it was emulated and hence the performace took a major hit.The hit was so bad that even the Itaniums younger siblings, Xeon and Pentium surpassed it.Mind you the Itanium was insanely expensive.during that time, most of the apps were written in 32bit and not 64bit cos there were no CPUs for it.So although large corporations adopted Itanium,smaller ones just couldnt because of the cost factor and the 32 bit performance issue.It is a well known fact that the Itanium's debacle is attributed to this inability to run 32bit apps efficiently.Then came Opteron and it offered Hybrid 32-64Bit support and it did it efficiently.It made sense to make a massive Opteron based clusters and get insane performance boosts. Since the K8 was a 64bit architecture,its memory controller could address insane amount of Memory and that too at incredibly low latencies.This hit the Itanium the worse.Even Intel started concentrating on Xeons ,the same way as AMD did,to go the cluster way.Itanium was Intel's most cherished and ambitious project but it failed to pick up.:)

Link - Itanium 32bit deficiency
^^ Not a proof or anything,just to give you idea that even Intel had agreed to Itanium's poor 32bit performance. ;)

AMD cpu exectes 10 calculations
Are you talking about IPC [Instructions per clock cycle] ? If that is the case then, K7 had 9 IPC compared to 6 IPC in Pentiums.

After i shift into my new flat i will answer all the queries (As i will have more time then) @DL AND all your LOL's too nothing personal but i think that people should understand a cpu and not go on the marketing gimmicks of some HIGH SOUNDING WORDS LIKE SMARTCACHE.. PIPED SIMD AND NETBURST... these are all simple technologies ... yeah it improves performance but not necessarily you use that.........
You have noble thoughts and intentions my friend but as people like you or me might be 'interested' in going a bit deep,others are just not bothered or maybe they dont have time or they just dont understand.So we should come up with a simple conclusion for such people regarding a particular product.Thats my outlook, hope you understand ;)

Regarding your logic of Product naming,well i dont think it works that way,here's why i think so,

You say,
AMD Athlon XP 2000+ [1.67 x 10 x 1024 = 17,100.8] Vs. Intel Pentium 4 2.0 GHz [2.0 x 8 x 1024 = 16,384]
Clearly 2000+ has a larger value than P4 2.0GHz,right ?

Ok now please explain these scenarios -

AMD Athlon XP 3200+ [2.2 x 10 x 1024 = 22,528] Vs. Intel Pentium 4 3.2 GHz [ 3.2 x 8 x 1024 = 32,768]
You see P4 3.2 gets a bigger value than the AMD value.

Next,
AMD Athlon XP 2800+ , 3000+ , 3200+ all had multiple versions and they had different clock speeds,
1] AMD Athlon XP 2800+ --> 2250MHz(Thoroughbred Core,266FSB) ,2083MHz (Barton Core,333 FSB) , 2133MHz (Barton Core,266 FSB)

2] AMD Athlon XP 3000+ ----> 2100MHz(Barton Core,400FSB) ,2167MHz (Barton Core,333 FSB)

3] AMD Athlon XP 3200+ -----> 2200MHz (Barton Core,400FSB) , 2333MHz (Barton Core, 333 FSB)

Same model nos. but different clock speeds,how ?? Doesnt make sense.

Well no offense but i know i am not architectural or computing guru but whatever facts i post are atleast true. If i am unsure of something,i do not post it.Its better to not post rather than posting wrong info ;)

As far as product rating goes, well even Intel chucked the clock speed thing long ago and rates their products.How did that happen ?

Seriously dude,nothing personal here just sharing info and correcting some wrong info people post here.Take it in that spirit.The 'LOL' and 'ROFL' are for those who post wrong info,concepts with such authority,it amazes me.

I hope you understand my intentions and take them in the right spirit. :)

Hope to share knowledge with ya ;)

Darky
 

saurav29

Waiting for 3 Miracles...
darklord said:
Thanks for your informative post,appreciated :)
AMD Athlon XP 2800+ , 3000+ , 3200+ all had multiple versions and they had different clock speeds,

1] AMD Athlon XP 2800+ --> 2250MHz(Thoroughbred Core,266FSB) ,2083MHz (Barton Core,333 FSB) , 2133MHz (Barton Core,266 FSB)

2] AMD Athlon XP 3000+ ----> 2100MHz(Barton Core,400FSB) ,2167MHz (Barton Core,333 FSB)

3] AMD Athlon XP 3200+ -----> 2200MHz (Barton Core,400FSB) , 2333MHz (Barton Core, 333 FSB)

Same model nos. but different clock speeds,how ?? Doesnt make sense.
Ok saw this post today only thanks man got something to think about perhaps the "FOOD FOR THOUGHT" :)

Actually shifted into new flat now but into an empty one and will be hell busy for some days getting the basic requirements... :-(

well this particluar number are diffrent because of the clock speed they use ... i will confirm it later as previusly AMD used 133 Mhz they shifted to 166 Mhz and Then to 200 Mhz so the cpu speed is diffrent at diffrent megahertz... This is the precise multiplier we use to OC the CPU too
I think you will be aware of this anyhow..:D AMD uses double bus speed contrast to intels quad pumped bus speeds to (200*4=800) and AMD will be using ( 200*2) = 400Mhz FSB .....

Btw got the answer for 3200+ :D what can i say AMD screwed up with this procy and there lies their false claims with 3200+ whereas it was about 31% slower .... might have matched the performance but it was slow in regards to their own comparisonal rating that they adopted...
take a look here you will see what i am talking about (*www.tomshardware.com/2003/05/13/high/page8.html)

I took the example of the lower speed because that was the very basis of where these rating started and AMD started taking intel Head on.... but later they started loosing to intel and C2D they were beaten flat back (and the crown stripped) but i think they are coming up with a true multi core (both 2 and 4 cores) Pocy unlike intel C2d where they fuse two chips on one single procy but then it is still in testing phases and with AMD in troubled water with ATI hostile takeover and market reviews and their loyals are shifting to C2D .... I sinclerely hate to see intel monopolizing the procy market and to begin it was AMD who bought the quality and competition to Desktop segment.. but AMD future at present does not seemed too bright and if they cannot beat C2D and perhaps launch their procy early with troubles what happens next is no mans guess!!:mad:
 

darklord

Cyborg Agent
saurav29 said:
Ok saw this post today only thanks man got something to think about perhaps the "FOOD FOR THOUGHT" :)

Actually shifted into new flat now but into an empty one and will be hell busy for some days getting the basic requirements... :-(

well this particluar number are diffrent because of the clock speed they use ... i will confirm it later as previusly AMD used 133 Mhz they shifted to 166 Mhz and Then to 200 Mhz so the cpu speed is diffrent at diffrent megahertz... This is the precise multiplier we use to OC the CPU too
I think you will be aware of this anyhow..:D AMD uses double bus speed contrast to intels quad pumped bus speeds to (200*4=800) and AMD will be using ( 200*2) = 400Mhz FSB .....

Btw got the answer for 3200+ :D what can i say AMD screwed up with this procy and there lies their false claims with 3200+ whereas it was about 31% slower .... might have matched the performance but it was slow in regards to their own comparisonal rating that they adopted...
take a look here you will see what i am talking about (*www.tomshardware.com/2003/05/13/high/page8.html)

I took the example of the lower speed because that was the very basis of where these rating started and AMD started taking intel Head on.... but later they started loosing to intel and C2D they were beaten flat back (and the crown stripped) but i think they are coming up with a true multi core (both 2 and 4 cores) Pocy unlike intel C2d where they fuse two chips on one single procy but then it is still in testing phases and with AMD in troubled water with ATI hostile takeover and market reviews and their loyals are shifting to C2D .... I sinclerely hate to see intel monopolizing the procy market and to begin it was AMD who bought the quality and competition to Desktop segment.. but AMD future at present does not seemed too bright and if they cannot beat C2D and perhaps launch their procy early with troubles what happens next is no mans guess!!:mad:

AMD used 133 Mhz they shifted to 166 Mhz and Then to 200 Mhz so the cpu speed is diffrent at diffrent megahertz... This is the precise multiplier we use to OC the CPU too ----> That is not the multiplier.That is the base frequency of the FSB.

AMD screwed up with this procy and there lies their false claims with 3200+ whereas it was about 31% slower .... might have matched the performance but it was slow in regards to their own comparisonal rating that they adopted... ---> False claim ? I wouldnt say that.Product rating is indicative and not confirmative that it will beat the competitors CPU IMHO. ;)

AMD started taking intel Head on.... but later they started loosing to intel and C2D they were beaten flat back (and the crown stripped)----> Huh ?? K8 never happened ?:D AMD Dominated for good 3 years. Prior to K9 launch AMD Market share was what ? 8-10% well after K8 its nearly 25% of total global CPU market,which is not a small feat to achieve in a tiny time frame of 3 years.

but i think they are coming up with a true multi core (both 2 and 4 cores)-----> Which age are you in ? AMD has Dual Core for more than a year now :)

unlike intel C2d where they fuse two chips on one single procy ----> Conroe aka C2D is a 'native' Dual Core ;) Its Kentsfield aka C2Q or Quad Core that is just 2 Conroe glued together :)

AMD in troubled water with ATI hostile takeover and market reviews and their loyals are shifting to C2D ------> AMD may not be 'earning' as before but their sales havent plummeted. They are selling like hot cakes cos of the cheaper prices ;) Infact IIRC AMD sells ALL the chips they make.Remember,it cant fulfil the demand and hence Charter also makes CPUs for them and still the demand is so high,nothing is left in the inventory.Basically the no. of chips made by AMD = No. of Chips Sold ;)

AMD future at present does not seemed too bright and if they cannot beat C2D and perhaps launch their procy early with troubles what happens next is no mans guess!!-----> K10 will be good,trust me,its a heavily reworked architecture ;)
 
Last edited:

saurav29

Waiting for 3 Miracles...
darklord said:
AMD used 133 Mhz they shifted to 166 Mhz and Then to 200 Mhz so the cpu speed is diffrent at diffrent megahertz... This is the precise multiplier we use to OC the CPU too ----> That is not the multiplier.That is the base frequency of the FSB.

but i think they are coming up with a true multi core (both 2 and 4 cores)-----> Which age are you in ? AMD has Dual Core for more than a year now

unlike intel C2d where they fuse two chips on one single procy ----> Conroe aka C2D is a 'native' Dual Core ;) Its Kentsfield aka C2Q or Quad Core that is just 2 Conroe glued together

AMD in troubled water with ATI hostile takeover and market reviews and their loyals are shifting to C2D ------> AMD may not be 'earning' as before but their sales havent plummeted. They are selling like hot cakes cos of the cheaper prices ;) Infact IIRC AMD sells ALL the chips they make.Remember,it cant fulfil the demand and hence Charter also makes CPUs for them and still the demand is so high,nothing is left in the inventory.Basically the no. of chips made by AMD = No. of Chips Sold ;)

AMD future at present does not seemed too bright and if they cannot beat C2D and perhaps launch their procy early with troubles what happens next is no mans guess!!-----> K10 will be good,trust me,its a heavily reworked architecture ;)
Man i cannot find anything wrong with the 1st statement Okay word Multiplier was wrongly used and hence misintrepreted....:( Sorry about that i referred in terms of OC in which Multiplier remains same and we change the speed of Bus so get a greater clock speed... I used the word precise precisely for it ..... Tell me why do you catch each word do you do it to irritate me....:confused: I think i know enough architecture to understand what is meant by FSB (base frequency)......... Or for some reason you doubt that:D
i think you get why a core speed is diffrent for same cpu

(Multiplier x base FSB= Frequency):confused:

1] AMD Athlon XP 2800+ --> 2250MHz(Thoroughbred Core,266FSB)(133Mhz)

,2083MHz (Barton Core,333 FSB) (166 Mhz), 2133MHz (Barton Core,266 FSB) (133 Mhz)

2] AMD Athlon XP 3000+ ----> 2100MHz(Barton Core,400FSB) ,2167MHz (Barton Core,333 FSB)

3] AMD Athlon XP 3200+ -----> 2200MHz (Barton Core,400FSB) , 2333MHz (Barton Core, 333 FSB)

Same model nos. but different clock speeds,how ?? Doesnt make sense.

so on then there was design level; (what word to use ??? ) Fabrication level ... Nope word i will think of later.... :mad:

OKay i used the word multicore i swear i wanted to write ( 4 and 8 cores ) you caught the point ....... No arguing that.... :DI know that AMD has dual core ...... I think most of the places the Intels Quad core is referred as C2d only with change in modelno to Q instead of E but then i donot follow hardware too much more intrested in how the things work then who the heck works better.........;)

As far as i know AMD cpu have no takers with intel cutting on the prices and amd cpus selling for peanuts..... I use the word peanuts because if you play the cost cutting game then the profit margins dwindle..... The effort goes in more and output less this bleeds the company dry..... AMD has put hands in bad managment ....... Then i donot know the real scenario. ... People who bought AMD 3000+ at 1000$USD are the bread earners for the company than like me who bought it for 50$ are just patrons while the company would have got only 30$ and rest going into the channel.... i think doesn't help the company much they will need a lot of money to finance and keep going...... aka running cost and at the same rate financing the R&D....... Can be costly affair but i think we better leave economics out of it.......:-x

K10 is a distinct dream dude wake up (its in testing stages and if they go on cheating the people with "INDICATIVE" marketing gimmicks and IMHO amd will really loose it and loose it BIG TIME!!!" and i thought that i was the only person on Crack & Weed ... Their AM2 platform is still plagued by the lack of good hardware to extract the potential of the cpu and its hardware so much that people advocating that 939 is better than AM2 .... These things takes time and with ... AMD being beaten by intel the failure of company is likely if not Inevitable ......:-x

Ask yourself question Dude .... most of the peoples in forums who buy the latest and best hardwares kinda breadearners of the Company which cpu and platform they use now...... and you will understand what i am talking about...................:-x
 

darklord

Cyborg Agent
^^ Yo first of all,nothing personal here :D Just point out some goof ups thats all ;) Take it in that spirit,chill.

Ok,

Man i cannot find anything wrong with the 1st statement Okay word Multiplier was wrongly used and hence misintrepreted.... Sorry about that i referred in terms of OC in which Multiplier remains same and we change the speed of Bus so get a greater clock speed... I used the word precise precisely for it ..... Tell me why do you catch each word do you do it to irritate me.... I think i know enough architecture to understand what is meant by FSB (base frequency)......... Or for some reason you doubt that
i think you get why a core speed is diffrent for same cpu

(Multiplier x base FSB= Frequency)
Point is my friend,maybe your concepts are clear but remember, others also read here and they shouldnt get some wrong ideas,thats all ;)
so on then there was design level; (what word to use ??? ) Fabrication level ... Nope word i will think of later....
Didnt get you .....

OKay i used the word multicore i swear i wanted to write ( 4 and 8 cores ) you caught the point ....... No arguing that.... I know that AMD has dual core ...... I think most of the places the Intels Quad core is referred as C2d only with change in modelno to Q instead of E but then i donot follow hardware too much more intrested in how the things work then who the heck works better.........
oh so you typed that by mistake....lol :D happens....chalta hai ;)

As far as i know AMD cpu have no takers with intel cutting on the prices and amd cpus selling for peanuts..... I use the word peanuts because if you play the cost cutting game then the profit margins dwindle..... The effort goes in more and output less this bleeds the company dry..... AMD has put hands in bad managment ....... Then i donot know the real scenario. ... People who bought AMD 3000+ at 1000$USD are the bread earners for the company than like me who bought it for 50$ are just patrons while the company would have got only 30$ and rest going into the channel.... i think doesn't help the company much they will need a lot of money to finance and keep going...... aka running cost and at the same rate financing the R&D....... Can be costly affair but i think we better leave economics out of it.......
its ok if you dont wanna believe me but fact remains,AMD sales have not declined ;)

K10 is a distinct dream dude wake up (its in testing stages and if they go on cheating the people with "INDICATIVE" marketing gimmicks and IMHO amd will really loose it and loose it BIG TIME!!!" and i thought that i was the only person on Crack & Weed ... Their AM2 platform is still plagued by the lack of good hardware to extract the potential of the cpu and its hardware so much that people advocating that 939 is better than AM2 .... These things takes time and with ... AMD being beaten by intel the failure of company is likely if not Inevitable ......
Its not a distinct dream.But if you insist,then well so be it ;) All i can say is K10 was taped out last year and working samples are there with AMD and they are fine tuning them to the last bit and make them ready for launch.
Wanna bet that K10 based CPUs will be out by June in the market ???
Seriously i am game for a bet,say 1000 bucks ;) what say ?

Ask yourself question Dude .... most of the peoples in forums who buy the latest and best hardwares kinda breadearners of the Company which cpu and platform they use now...... and you will understand what i am talking about...................
Lol chiull dude,you got your point through...relax ;)
Let's have a healthy discussion shall we ? ;)
 

saurav29

Waiting for 3 Miracles...
LOL :D sorry man some problems with office and too much work and accomodation fuse was blown off last night........

Nothing that "last legal drug on earth" cannot take care of...:D

Peace
raj
 
Status
Not open for further replies.
Top Bottom