Smartphone players have done much on the increasing screen
size, loading advanced versions of operating system, reducing thickness of the device
and so on. From last year beginning, the quad-core mania started. Instead of
blindly counting on the theoretical benchmarks and expecting the device to
perform more swiftly than the dual-core device, it is always better to
understand other factors that actually affect the device performance in
real-life tasks.
Theoretically multi-core can be compared to room with
multiple doors. When too many people gathered in the room are asked to leave
the room simultaneously, a 4 door room can be emptied easily compared to a 2
door room. Besides the fast processing, quad-core chips also promise better
battery life. This is because, a single core has to work hard to complete a
task on its own, while multiple cores does the same work with more ease, depleting
the battery slower.
Theory is just theory. There are many factors that actually
affect the processing speed.
Requirement for operating system support
No matter how many cores are there in the processor, there
is nothing much it can do without the help of software. There should be software
to divide a particular task and assign to each core a portion of the task. Technically
this is called “Multithreading” and this process can be done only with the help
of software. So it is important to have a software that supports
multithreading.
Lack of multithread execution in apps and games
Suppose you have a multi-cored device loaded with
multithreading supported software. But what if the apps or games are not coded
to support multithreaded execution? It is highly complicated for developers to write
codes or debug apps and games to support such execution.
Limitations in video streaming
The most demanding applications that should ideally get
benefited from multiple cores are gaming and videos. When threaded games are
not being developed, streaming videos from various online sources also is not
getting benefitted from multi-cores. The reason is that the videos steams
cannot be easily broken up. So they end up being processed using a single core;
other cores may be either idle or being used for less demanding tasks such as
email.
Memory bottlenecks
Apps processing can be divided into multiple cores. But the
multiple cores are still using shared memory space; that means there is no
dedicated memory assigned to each core. In some cases, the memory
architecture can also become a bottleneck. A single channel memory architecture reduces
the data transfer rate between the memory and chipset, compared to a dual channel
memory.
Battery life myth
“More cores means less power” theory is not always right. It
all depends on how intelligently the software is able to do the optimization to
minimize the energy consumption. The energy efficiency is improved only when more demanding tasks are being carried out. But most of the time, multiple
cores are being used for less demanding tasks, ultimately increasing the energy
utilization.
To conclude: Performance is not just connected to number of
cores, but tied to an optimized ecosystem of hardware and software.
No comments:
Post a Comment