COMP Chap 3-5 (Memories -3) PDF

Definition 1.1. HDD 1.1.1. Definition Hard Disk Drive is a magnetic mass storage device used to store data permanently. The hard disk stores all the memory of your computer, including (see figure 3.41): - The operating system (Windows, Linux, etc.). - Software applications. - Persona...

Definition 1.1. HDD 1.1.1. Definition Hard Disk Drive is a magnetic mass storage device used to store data permanently. The hard disk stores all the memory of your computer, including (see figure 3.41): - The operating system (Windows, Linux, etc.). - Software applications. - Personal data. Figure 3.41. Contents of the HDD. Your hard drive is listed in the "Computer" icon in Windows. The icon representing your disk is labeled as "C:". If your computer has a second disk, you will see an icon labeled "Local disk (D:)". You can also partition a single HDD into 2 partitions (C:) and (D:), or even more (3 partitions, 4...). An example is presented in figure 3.42. Figure 3.42. HDD partitions. There are two types of hard disk (see figure 3.43): - Internal hard disk: Located inside the computer's system unit, connected to the motherboard via PATA, SATA or SCSI cables. - External hard drive: Located outside the system unit, connected to the computer's system unit via a USB connection. 5 HDD Architecture Figure 3.43. The two types of HDD. 1.1.2. HDD Architecture A hard drive (whether internal or external) is a magnetic disk composed of one or more vertically stacked platters. These platters can be made of metal, aluminum, plastic, glass or ceramic, covered with a layer of magnetizable material. Early platters used to be as large as 50 cm in diameter, while today they are typically 3 to 12 cm, or even smaller for pocket computers. The disks rotate very rapidly around an axis in the counterclockwise direction (see figure 3.44) [6, 17]. Figure 3.44. The hard drive (internal). Each platter has two sides: an upper face and a lower face. Each face is associated with a read/write head fixed on a movable arm. Each face is divided into several circles called tracks, numbered from 0 starting from the outer edge. Each track is divided into a certain number of sectors (10 to 100 sectors per track) of fixed lengths, each containing a minimum of 512 bytes of data (see figure 3.45) [6, 17]. 6  Structure of a sector Figure 3.45. Disk with four platters [6, 11]. Each disc has a movable arm capable of radial movement to position the head at a certain distance from the rotation axis of the platter. Each radial distance corresponds to a different track that can be read from or written to. The tracks thus form a series of concentric circles surrounding the axis of the platter. The width of a track depends on the width of the head and the precision of its positioning along the radius. With current technologies, it is possible to construct disks with five thousand to ten thousand tracks per centimeter, resulting in a track width of about 1 to 2 microns (1 micron = 0.001 millimeters). A track on a hard disk is not an actual groove in the material but a simple magnetizable ring, separated from the immediately neighboring tracks by a protective zone called "inter-track space". Sectors are also separated by a space called “inter-sector space”. The collection of data located on the same track on all platters is called “a cylinder”. There are as many cylinders as there are tracks on one face. 1.1.3. Structure of a sector In a sector, there are three distinct zones (see figure 3.46), each serving specific purposes: - Data zone: Used to store the actual data to be saved on the sector. This is the space occupied by the information. - Identification zone: Allows the head to synchronize and identify the data zone. It contains the sector number. - Control zone: Contains a group of bits to control the data contained in the data zone. 7 Disk Drive Performance Figure 3.46. Organization of sectors on a track. 1.1.4. Disk Drive Performance To be able to write or read a sector : - The arm must first reach the correct radial distance; this process is called “ positioning”. Today, the average positioning times between any two tracks ranges from 5 to 10 ms, and it is less than 1 ms between two consecutive tracks. - A second delay, called "rotational latency" elapses while desired sector aligns under the head. Most disks rotate at 5400, 7200, or 10800 revolutions per minute. Therefore, the average wait time (a half rotation) is about 3 to 6 ms. - Data transfer time (measured in bits per second). Each disk is associated with a disk controller. It is a circuit designed to manage the disk's operations. The controller's tasks include receiving commands (read, write, etc.), controlling arm movement, head positioning, and more (see figure 3.47). Figure 3.47. The hard disk controller. 1.1.5. Disk Drive Capacity It refers to the amount of information an HDD can hold and is measured in bytes. The capacity of a disk depends on its geometry and the sector capacity. Disk geometry = number of heads, number of cylinders and number of sectors per track.  Reminder Disk capacity= number of sectors per track × number of cylinders × number of heads × sector capacity. 8  Power supply  Note Some manufacturers express the capacity of their disks by including even the space for identification and control zones. However, it is more honest to measure the capacity by considering only the actual data zone, which is generally less than 15% of the total disk capacity. 1.1.6. Power supply To function, a hard drive requires an electrical power supply. This is provided by the power supply unit. For IDE disks, a MOLEX connector is used, connected next to the IDE port (see figure 3.48). Figure 3.48. MOLEX connector. For SATA disks, the ports used for data and power have the same name “SATA”. To differentiate them, the power port is larger than the data port (see figure 3.49). Figure 3.49. IDE (left) and SATA (right) drive connectors. Some SATA drives also have a MOLEX connector, as seen above, so they can be used with a power supply unit that only offers MOLEX connectors. However, it is crucial not to use the both MOLEX connector and the SATA power connector!  Note:[6,11, 15] - As seen in the figures, sectors on different tracks do not have the same size, but all sectors contain the same amount of information. - The heads are electromagnets that lower and lift to read or write information. - The heads do not touch the surface, they are a few microns (< 0.25µm) above the disk's surface. - The disks rotate, creating a wind of approximately 250km/h! with thousands of rotations per minute. - The heads are movable laterally to scan the entire disk's surface. - All the heads move forward or backward simultaneously (not only the head involved in reading/writing). 9 SSD - Only one head can read or write at a given time. - A file is considered "sequential" if all its sectors are placed in the same cylinder, in order or on several neighboring cylinders. When one cylinder is full, the reading process moves to the next cylinder. A file is considered “direct/random access” if its sectors are scattered in different locations on the disk. Reading a sequential file is much faster than reading a direct/ random access file. - On hard disks, addressing is done physically by defining the data's position with the coordinates: cylinder / head / sector. 10  Definition 1.2. SSD 1.2.1. Definition Isn't incredible that in our modern age, there are still mechanical parts in our computers, such as the read/write heads of hard disks? These inherently “slow” parts are real bottlenecks in the system's operation. Well you are not alone in thinking that! This is why the SSD (Solid State Drive) was invented. SSDs are the worthy successors of our beloved hard disks, so we quickly tend to call them "SSD disks". Yet, they have absolutely nothing in common with a disk: The technology used is different, they don't rotate, and they don't have platters or a motor... They are not even rounding! So, it is a linguistic misnomer due to the history of computing. But then, if an SSD has neither platters nor mechanical read/write heads, how does it work? The answer lies in two words: flash memory. Indeed, the memory of an SSD is of the “flash” type, meaning that the data is written in small memory cells within a chip. There are, therefore, no mechanical elements. Each memory cell can be read or written with the same delay, regardless of its positioning on the chip. With a traditional hard drive, you had to wait for the read head to move across the platter's surface to fetch data. It was long, too long, but that's no longer the case with SSDs. 1.2.2. Controller Alongside its flash memory chip, the SSD is equipped with a controller, which is responsible for selecting the cells to read and store data. This controller is the real strength of SSDs, as it can access several cells simultaneously, increasing the data rate throughput! Instead of reading one cell, then another, and so on, the SSD reads multiple cells at once and consolidate the data. The time savings are significant. 1.2.3. Format and connectors Now let's look at the different formats of SSDs and their connectors. The majority of SSDs are in the 2.5-inch form factor, the same size as laptop hard disks. You can also find them in 3.5-inch, usually slightly cheaper, but less common. In both cases, the most common interface is SATA (see figure 3.50). Figure 3.50. 2.5-inch SSD “SATA”. That covers the “standard” SSD formats, which closely resemble classic hard disks. However, there are also SSDs in the form of "daughter cards" which connect to a PCIe port (see figure 3.51). These SSDs offer the highest data transfer rates, but are also the most expensive. 11 An SSD Or a Classic Hard Disk? Figure 3.51. 2.5-inch SSD “PCIe”. Finally, there are SSDs that connect via mini-SATA. They are mainly used in laptops, NetBooks, tablets, smartphones, and more (see figure 3.52). Figure 3.52. SSD "mini-SATA". 1.2.4. An SSD Or a Classic Hard Disk? Should You Choose an SSD Instead of a Classic Hard Disk? To make a decision, we need to consider the advantages and disadvantages of SSDs. Advantages of SSDs : - Larger storage capacity. - Tremendous speed. - Absence of any mechanical parts allows for much faster data access. - Shock resistance and no noise (due to the absence of mechanical parts). - Lower power consumption. At first glance, there is no doubt: SSDs are preferable! But, not everything is so perfect. The major disadvantages of SSDs are: - Much higher price. - Lower lifespan compared to classic hard disks (10,000 to 100,000 write cycles). This is due to the number of read/erase/write cycles that the memory cells can endure. SSDs are more effective for reading than at writing. However, for the operating system, that's precisely what's needed! An OS spends most of its time accessing system files and rarely needs to modify them. The same principle applied to the majority of software. Therefore, SSDs are well-suited for being the system drive. A 60 GB or 128 GB 12  The Graphics Card SSD is usually sufficient for the OS and software. Alongside that, you should have a classic hard disk with good capacity to store all your data (documents, photos, etc.). The SSD/HDD combination is an excellent way to achieve a very responsive system while retaining ample storage capacity. It was mentioned that the SSDs have a shorter lifespan. Is it wise to place the system on a disk that can fail at any moment? It is better to lose the system than to lose data (you can reinstall Windows, but you cannot find your lost photos). Some common recommendations to increase the lifespan of an SSD include: - Limiting unnecessary writes. - Storing all data on a classic hard disk, which complements the SSD reserved for the system. - Disabling unnecessary features given the SSD's speed (disk indexing, extended hibernation, etc.). - In general, avoiding any write processes that have no real purpose. The following table presents the differences between an SSD and an HDD. Table 3.8. Differences between an SSD and an HDD. Characteristics SSD HDD Random access time About 0.1 ms From 2.9 to 12 ms Read/ write speed From 27 MB/s to 3 GB/s From 12 to 260 MB/s Noise Very little to none Variable but tends to increase with time and usage Vulnerabilities Power los scan render the drive Sensitive to shocks, vibrations, and unrecoverable on certain (older) magnetic fields models Size 4, 5.7 , 6.35 cm (1,8”, 2,5”) 4, 5.7, 6.35, 8,89 cm (1,8”, 2,5”, 3,5”) Mass Few tens of grams Up to nearly 700g Cost / capacity Ratio About 0,4 €/ GB (2014) About 0,04 €/ GB (2014) Storage Capacity Up to 16 TB Up to 10 TB Power Consumption 0,1 W to 0,9 W (idle), up to 0,9 W 0,5 W to 1,3 W (idle), up to 2 W to (active) 4W (active) 2. The Graphics Card It is an electronic component responsible for converting the digital data to displayable graphics data that can be used by a display device. There are two types of cards: 13 Graphics Card Interfaces and Connectors - IGP (Integrated Graphics Processor): Integrated directly into the motherboard (Northbridge) or within the processor. It is less expensive but has limited performance. - Dedicated Graphics Card: An additional card with its own memory. It is more powerful (see figure 3.53). Figure 3.53. Nvidia Graphic Card. 2.1. Graphics Card Interfaces and Connectors - Interface: The type of bus used to connect the graphics card to the motherboard is AGP or PCI Express bus. - Connectors (see figure 3.54) : - VGA: or SUB-D15 = 3 series of 5 pins. - DVI (Digital Video Interface): Sends digital data to the screens to avoid the conversions to/from analog. - S-Video: Many cards are equipped with an S-Video port to display on a television (also called “TV-out”). - HDMI (High-Definition Multimedia Interface): Combines video and audio signals in a single connector. An extension of DVI, replacing (S-Video). Figure 3.54. Graphics card connectors. 2.2. Graphics Cards Manufacturers Currently, there are only two major graphics card manufacturers: AMD and NVIDIA, but there are many graphics cards brands. Nvidia and AMD supply their graphics chips to numerous manufacturers such as: Hercules, Sapphire, HIS, Asustek, Gigabyte, Point of view, MSI, XFX, and many others. The performance of the cards will be very similar or even identical when they use the same Nvidia or AMD graphics chip. Therefore, the choice of the brand depends on the price and properties of the card. 14