Dual core processors: The main logic

The dual core and quad core processors have become very popular nowadays but the main reason behind their success is the power management.

There are two basic concepts that one must know, they are :-

Overclocking the processor :-

Let's say that ideally a normal single core processor takes x units of power to give x units of performance. But users generally tend to increase the performance by overclocking the processor ie essentially increasing the clock frequency of the processor. If a user follows the process of overclocking the processor, then the performance increase is about 12-13 percent but the power increase required for that would be about 70 percent.

Under-clocking the processor :-

Now the interesting concept is what if we under-clock the processor that is decrease the clock frequency of the processor?

In this case we could get about 80 percent of the performance at roughly half the power. For example we would get .8x units of performance at .5x units of power.

Now this is the concept that is utilized by the dual core processors and instead of using a single processor we combine the two processors and operate each one of them at half of it's power and so for the same power taken by the normal single core processor we get about 70 percent increase in performance. The aforementioned reason is the core reason behind the success of dual core processors.

Another major factor in dual core processors is as they operate at lower clock rates it helps in reducing the speed mismatch between the processor and the memory and this is a relevant factor because with processor clock rates increasing at very faster speeds the corresponding increase in clock rates of cache memory is not their, therefore a processor's clock cycles are not utilized properly and are wasted in memory read/write cycles.

Hence the concept of under-clocking helps also in reducing the speed mismatch b/w the memory and the processor.

Read More

Intel's Nehalem architecture: What's new ?

Intel's Nehalem architecture is utilized in their i7 Core processors for desktops and Nehalem based Xeon processors for servers and the following features tell you about what's new in them and why you should purchase one of these.

• High performance: The processors perform better than the amd's phenom and intel's core 2 extreme processors available in the market. The extra performance boost is always welcomed by gamers. When performance was tested on the game World of Conflict the phenom processor gave 136 fps, Core 2 extreme processors gave 220 fps whereas the nehalem architecture based processors gave out 250 fps.


• Enhanced Hyper Threading: Hyper threading technology (HTT) is a term used by intel which simply means simultaneous multi-threading which causes the operating system to interact with a single processor as if you had dual processors. So it doubles the number of threads that can be handled by the processor at one time. Thus if you have a Nehalem based quad core processor it can handle 8 threads going by the same logic.


• Turbo boost : Dynamic performance Control :-
  This is similar to Overclocking in the standard processors but unlike the standard processors where you had to manually overclock or under-clock the processor to suit your performance requirements, in this case it is handled automatically by the processor depending upon the system load. So if the load is higher it can automatically increase the clock rate i.e basically increasing the frequency and when the load decreases it automatically lowers it's clock rate.


• QPI(Quick path interconnect):- This technology is similar to amd's hyper transport technology but it is much faster than it. What the QPI does is it gets rid of the FSB(Front Side Bus) in the standard processors. The Xeon and i7 core processors provide 25.6 GB/s per link that is twice that of 1600MHZ FSB (which according to me is very fast :-) ) .
  
  The QPI architecture provides separate integrated memory controllers for each core so that instead of the memory and I/O requests sharing a bus and waiting for many cycles, the QPI and the memory bus are separate. QPI also provides separate channels for writing and reading, so these tasks can be performed in parallel.
  


• 3 Level Cache :- The dual core processors came with a 2 level cache whereas in the nehalem architecture it is expanded to 3 levels of cache. Each processor has it's own L1 and L2 levels of cache and all the processors share the L3 Level cache which is about 8MB in size. The L3 level cache contains copies of the contents of L1 and L2 caches which leads to better performance.


• DDR-3: Higher speed memory :-
  The Nehalem architecture also supports faster (so logically more expensive) RAM: The DDR3 RAM is 2wice as fast compared to the DDR-2 RAM and also maintains the same ratio when it comes to the prefetch buffer. The prefetch buffer is used to buffer the instructions for the processor when it is performing some operation and logically larger prefetch buffer size leads to better performance.


• Lower power consumption:- The Intelligent Power Node Manager for Xeon servers will automatically adjust the power for the optimum power to performance ratio (this was already mentioned when the turbo boost feature was explained). This leads to better power management and hence lower power consumption.
  

Hope this was informational enough and now that you know of the benefits of the nehalem architecture you would be in a better position to know whether or not you want to buy one of these.

Read More