Topics - , Addressing modes, GPU, .pdf

By Sonu, Sarthak,Sameer Ali,
Shashank and Shailendra

• CACHE MEMORY
• ADDRESSING MODES
• COMPARING DIFFERENT PROCESSORS IN THE MARKET
• GPUs(MULTI - CORE ARCHITECTURE)
Processing Data 02

Cache is basically temporary memory
(also the fastest) in your CPU which lets
you access some information more
quickly compared to how you would
normally access it from your computer’s
secondary storage device. The data stored
in this memory is from your most used
program/files.
NOTE: Cache memory is also sometimes placed on a separate
chip on the motherboard which has a separate bus interconnect with
the CPU. It is also found in GPUs.
03
Processing Data

L1 : This level is the fastest memory present on the CPU. L1
gets the highest priority in terms of accessing the data
although the size of cache of L1 is the lowest out of all the
levels. (Usually measures in KBs, below 1MB except server
chips.)
L2: L2 cache is slower than L1’s but bigger in size. When CPU
can’t find the data in L1, it goes to L2. L2 cache might lag
behind L1 in terms of speed but it’s still faster than your
system’s RAM.
L3: L3 cache is the largest cache unit out of all levels but also
the slowest one. This is the last level of cache unit CPU
comes looking for when it cant find the required data in L2.
While L1 and L2 cache exist for each core on the chip, L3
cache is more similar to a general memory pool that the
entire chip can make use of.
NOTE: L1,L2 and L3’s size varies depending on the CPU.
04
Processing Data

The different way of specifying the location of an operand in an instruction are called as
addressing modes.
The operation (opcode) field of an instruction specifies the operation to be performed. This
operation will be executed on some data (operand).
Address field contains the information needed to determine the location of the operands
and the result of an operation.
The Purpose of using Addressing modes is:
 To give the programming versality to the user.
 To reduce the number of bits in addressing field of instruction.
Processing Data 05

1. Implied Addressing Mode
2. Stack Addressing Mode
3. Immediate Addressing Mode
4. Direct Addressing Mode
5. Indirect Addressing Mode
6. Register Direct Addressing Mode
7. Register Indirect Addressing Mode
8. Relative Addressing Mode
9. Indexed Addressing Mode
10. Base Register Addressing Mode
11. Auto-Increment Addressing Mode
12. Auto-Decrement Addressing Mode
Processing Data 06

 The definition of the instruction itself specify the operands implicitly.
 It is also called as implicit addressing mode.
Example: The instruction “Complement Accumulator” is an implied mode instruction.
 The operand is contained at the top of the stack.
Example: ADD
1. This instruction simply pops out two symbols contained at the top of the stack.
2. The addition of those two operands is performed.
3. The result so obtained after addition is pushed again at the top of the stack.
Processing Data 07

 The operand is specified in the instruction explicitly.
 Instead of address field, an operand field is present that contains the operand.
Examples:
1. ADD 10 will increment the value stored in the accumulator by 10.
2. MOV R #20 initializes register R to a constant value 20.
 The address field of the instruction contains the effective address of the operand.
 Only one reference to memory is required to fetch the operand.
Example:
1. ADD X will increment the value stored in the accumulator by the value stored at
memory location X.
AC ← AC + [X]
Processing Data 08

 The address field of the instruction specifies the address of memory location that contains the effective address of the
operand.
 Two references to memory are required to fetch the operand.
Example:
1. ADD X will increment the value stored in the accumulator by the value stored at memory location specified by X.
AC ← AC + [[X]]
Processing Data 09

CPU clock speed is measured in hertz, usually gigahertz (GHz). A CPU’s clock rate is basically a measure of how
many clock cycles it can perform per second. For example, a CPU with 2.4 GHz clock rate can perform 2.4 billion
clock cycles per second. Now one must think that the more cycles a CPU can perform, the better it is? Not really.
Back in early days of CPU, higher clock speed and more cores meant faster processor and better speeds but
that’s not the case anymore. Processors today aren’t dependent as much on clock speeds and cores as they
were back then because modern CPUs have several other features that determine how fast they can perform
like dynamically boosting their clock speed under favourable conditions.
Basically, it all comes down to how much work can be done when all parts of a CPU come together. Let’s take 2
CPUs, CPU_1 performs a task in two clock cycles while CPU_2 performs the same task in 1 cycle. Here, CPU_2
might be the better processor even if CPU_1 has a higher clock speed.
NOTE :This definitely does not mean clock speed doesn’t matter at all. For e.g. It matters when you are trying to
decide between two CPUs from same family/category having same no of cores.
Processing Data 10

A little unknown fact in today’s tech world is that even though you are buying a processor with 6 cores, all
6 of those cores might not be used when you are running certain applications.
Why is that so? A lot of the applications and games today are still single-threaded as in the application is
running as one single process and a single process can only run on one core. So even if you have multiple
cores, you won’t be getting the performance of all those cores in that particular application.
That’s why you should always check for single core performance of the processor you are going to buy.
Now, the problem is that companies usually do not reveal any information about single or multi core
performance so you have to rely on third party data sources.
Some of the reliable resources are Tom’sHardware, PassMark's CPU benchmarks and HardwareUnboxed .
Cache is one of the most important parts of a CPU. So much important that if your CPU has small cache
memory, it could be slowing down your PC. So always check how much cache a CPU has when comparing.
Note:This definitely does not mean you should forget about multi core performance,both of them matters equally but a lot
of people to this day still don’t pay much attention to single core performance.
Processing Data 11

Benchmark tools are a great way to compare CPUs in a real world scenarios.
Some of the popular benchmark tools are :
1. Geekbench
2. Cinebench R23
3. Blender
4. 7-ZIP
5. AIDA64
Additionally you can go to AnandTech's Benchmark Tool which is also one of the best independent
hardware review sites to compare different CPUs in different benchmarks tools like the ones stated
above.
Processing Data 12

GPU or graphics card is a vital piece of hardware
just like CPU in your PC. Like the motherboard,
GPU(discrete ones) is also a printed circuit board
which comes with it’s own fans, RAM, own memory
controller, BIOS and other features.
A GPU is primarily responsible for rendering
images on your display device, be that photos,
video games, applications, a file/folder, your
desktop or literally anything else you can think of.
All of these tasks, like video games which require
an enormous computing power to something so
simple like opening a text file requires the use of a
graphics card.
Processing Data 13

GPUs are mainly divided into two types:
: An integrated GPU(iGPU) is basically built in the same
housing as CPU. Majority of CPUs do come with integrated graphics
although they aren’t that powerful and are only useful for basic and
light tasks like rendering your desktop, browsing, watching a video etc.
Although a CPU lineup called APUs(Fig. I) from AMD are usually more
powerful than standard iGPUs found in other CPUs. Albeit being a little
bit powerful, they are still only good enough for light gaming, simple
video editing etc.
: A discrete GPU is as the name suggests, an external GPU
which is added to an expansion slot on the motherboard usually PCI
slot. It delivers more power than the iGPU and can be used for
extensive tasks like video editing, gaming, 3D rendering etc.
Bottom Line:For best performance,opt for a discrete GPU.
Fig I. AMD RYZEN APUs
Processing Data 14

Specific hardware usually varies between GPU models because of different manufacturers but most
modern discrete GPUs have:
GPU: The main PCB which is the actual hardware component similar to a CPU.
Memory: Also known as VRAM(Video RAM), a graphics card comes with it’s own dedicated VRAM to
do tasks.
Interface: Majority of the GPUs today use PCI-E (PCI Express).
Outputs: A GPU can have various outputs, often comprising of HDMI, DP(DisplayPort),VGA, DVI.
Fans/Heat Sink: All discrete GPUs come with fans and a heat sink to help dissipate heat that build-
up during usage.
Power Connectors: Most of the modern GPUs today require a 6 or 8-pin power connector,
sometimes even requiring 2 or 3 of 6/8-pin connectors.
BIOS: The GPU BIOS holds initial setup and program information, retaining data on voltages, memory,
and more when you boot up your machine.
Processing Data 15

INTEGRATED GRAPHICS
• Built into CPU and shared system memory.
• You cannot place both a CPU and GPU on the same chip
without limiting one (or both) of their potential processing
power.
• Sufficient for basic tasks, such as, web browsing and video
playback.
• Integrated graphics cannot handle complex graphics because
of lower processing power and memory.
DISCRETE GRAPHICS
• Have their own dedicated memory and are typically faster
and more powerful.
• A dedicated graphics card has its own PCB, memory (VRAM),
and cooling solution, and, as a result, is not as limited in its
potential power.
• Necessary for more demanding tasks such as gaming and
video editing.
• They have more memory and processing powers, which can
used in handling higher resolution and more complex
graphics.

Topics - , Addressing modes, GPU, .pdf

Topics - , Addressing modes, GPU, .pdf

More Related Content

Similar to Topics - , Addressing modes, GPU, .pdf (20)

More from ShubhamSinghRajput46 (6)

Recently uploaded (20)

Topics - , Addressing modes, GPU, .pdf