2x2=4? 4 32bit processors from a dual core?

[xiangkc]

Greeting! This is what i have
Intel Core 2 Duo E4300 Allendale
GIGABYTE GA-965P-DS3
Being a first time dual core owner for about a year now... I realize... 64bit x 2 processor running under 32bit windows xp... how come I only see two graphs on my task manager? shouldn't there be 4? Recently I played around and enable my friend's hyperthreading and realize the task manager showed two processor. 64*2 /32 = 4 processors for xp?

SOoo... the question is.. how do i make my core2duo to show 4 graphs in the task manager? is it even possible? what's the downfall or enhance performance?

Thank you all.

ps.. it's hard to google this problem

 

[nukemaster]

Core2's do not use hyperthreadding.

All hyper threading was a way for P4 based systems(only those with HT) to fake a second cpu. That said, i has nothing to do with 64/32 bit. When the cpu is running in 32bit windows, it just runs 32bit code(its capable of 64 but needs the OS to support it...XP-64/Vista/Linux/ect).

The biggest downfall to HT was that there was still only one cpu. so if an app was using SSE or another instruction set it was NOT available to other apps(in some cases this even slowed down apps that thought there was a real 2nd cpu). With you real dual core each cpu has it own full set of instructions....

There was a Pentium D extreme edition with 2 cores and HT. That made it look like 4, but its still 2. There will still be a boost, but clock for clock Core2 kills p4(as in a 1.8ghz Pentium dual core[core2 based value chip] beats the PentiumD[2 x p4] @ 3.0)....

With quad core becoming the new mainstream(look at how cheap they are). its just a matter of time before you get 4 boxes

Check out the cpu charts
http://www23.tomshardware.com/cpu_2007.html

 

[skittle]

There is no problem. Your just confused.

64bit (x86-64) is just a superset of the x86 instruction set, kind of like SSE. Basically it just allows your computer to address more memory space than a 32bit system.

 

[waxdart]

If I used a old 16bit OS you wouldn't see more Cores. You just don't get the full benefit.

2 cores, 2 graphs. If you want 4 graphs buy a Q6600 and join the bigs boys with 4 graphs. Not that it will make much diff in games and things not made to use 4 cores.

 

[ZozZoz]

Are you for real? oh my god..... that's hilarious.

 

[tenshin111]

I think you should switch to Mac....your life will be much easier!

 

[xiangkc]

There is no problem. Your just confused.

64bit (x86-64) is just a superset of the x86 instruction set, kind of like SSE. Basically it just allows your computer to address more memory space than a 32bit system.

okay.. for somereason i got this idea that 64bit p4 can become two "fake" 32bit processors

 

[nasim_30]

I think you should switch to Mac....your life will be much easier!

Ha ha ha! very good ans.
One day my friend told me, he bought a 9.6 Ghz processor.
Do you guys undrstand?
It was Q6600, 2.4x4= 9.6.
Ha ha ha!

 

[xiangkc]

lol i bet a mac user can say something like that too. since they have intel core2duo now. how about amd + mac?

 

[thefumigator]

the bits are the address lenght to do just that address the memory. Just like addressing each home on a street. If you use 3 numbers, lets say, from 000 to 999 then you can address 1000 houses.

If your street is too long and has more than 1000 homes, then you need an extra number, 4 in total, which makes the range from 0000 to 9999 which are 10000 addressable homes in total.

Same with memory, just replace each "home" with "byte"; and "number" with "bit" where it can be zero or one. Amount of bits are 32 or 64.

The other things that sometimes are asociated with the processor bits is the registries (somewhat like small internal CPU memory where you hold the data ready to be processed). Some registries are 32bit some are 64 and so on. I will let anyone who is more into this to explain it...

That's it.

 

[xiangkc]

the bits are the address lenght to do just that address the memory. Just like addressing each home on a street. If you use 3 numbers, lets say, from 000 to 999 then you can address 1000 houses.

If your street is too long and has more than 1000 homes, then you need an extra number, 4 in total, which makes the range from 0000 to 9999 which are 10000 addressable homes in total.

Same with memory, just replace each "home" with "byte"; and "number" with "bit" where it can be zero or one. Amount of bits are 32 or 64.

The other things that sometimes are asociated with the processor bits is the registries (somewhat like small internal CPU memory where you hold the data ready to be processed). Some registries are 32bit some are 64 and so on. I will let anyone who is more into this to explain it...

That's it.

Nice I understand it now. care to explain how hyperthreading works? in a more "house and number" style

 

[Pinhedd]

HT works by providing a CPU with two threads of execution at once, each thread is simultaniously filled with instructions from the cache memory. When the CPU gets a cache miss on an incorrect branch prediction (loads the wrong code) rather than going idle while the pipeline is refilled, it switches to the other thread and begins executing it.

 

[thefumigator]

as pinhedd pointed out, that's hyperthreading. Which makes sence why is not as good as having 2 separate cores for doing 2 jobs.

Hyperthreading helped a lot when multitasking was heavy, I still remember the tom's multitasking testing, Pentium with 2 cores + ht each core could beat in the 4 threads bench against one A64 X2.

 

[MarkG]

HT works by providing a CPU with two threads of execution at once, each thread is simultaniously filled with instructions from the cache memory. When the CPU gets a cache miss on an incorrect branch prediction (loads the wrong code) rather than going idle while the pipeline is refilled, it switches to the other thread and begins executing it.

Not quite. Hyperthreading runs both threads simultaneously, if possible; any execution units not used by the first thread in the current clock cycle are available to the second thread... obviously if the first thread is stalled that means the second thread has full use of the CPU, but if the first thread is running and the instructions it's executing aren't using all the execution units, other instructions from the second thread can run in the same clock cycle if they only require the available units.

To be honest, I suspect that 80-90% of the time the second thread really is only using the CPU because the first has stalled, but it is somewhat smarter than you make out.

 

[yomamafor1]

Ha ha ha! very good ans.
One day my friend told me, he bought a 9.6 Ghz processor.
Do you guys undrstand?
It was Q6600, 2.4x4= 9.6.
Ha ha ha!

I'm really not surprised. Once I found a seller selling an E6400, and claiming it to have 4.2Ghz capability.

I wonder if he had sold that E6400 yet.

 

[Doctor_HV]

Ha ha ha! very good ans.
One day my friend told me, he bought a 9.6 Ghz processor.
Do you guys undrstand?
It was Q6600, 2.4x4= 9.6.
Ha ha ha!

But hey, could we at least say that it is as fast as some old P4 at 9.6GHz would be...
And if we have double the L2 cache, isn't it as 8x some P4 at 2.4GHz with 1MB of L2 cache?
or it is not going that way?
I was wondering about that, and I just joined this forum, so I want to ask that in a proper thread.