Questions and Answers of Computer Architecture

The cache system can be located between the CPU and the MMU (i.e., a logical cache) or between the MMU and the system random access memory (i.e., a physical cache). What factors determine the optimum
Why is it harder to design a data cache than an instruction cache?
When a CPU writes to the cache, both the item in the cache and the corresponding item in the memory must be updated. If data is not in the cache, it must be fetched from memory and loaded in the
A cache memory may be operated in either a serial or a parallel mode with respect to the main memory. In the serial access mode, the cache is examined for data, and if a miss occurs, the main storage
If a serial mode cache is used and a 5% penalty in speedup ratio over the corresponding parallel access cache can be tolerated, what value of the hit ratio is necessary to achieve this? Assume that
A system has a level 1 cache and a level 2 cache. The hit rate of the level 1 cache is 90%, and the hit rate of the level 2 cache is 80%. An access to level 1 cache requires one cycle, an access to
A computer has a cache with an access time of 1 cycle and an average hit rate of 95 percent. The miss penalty is 100 cycles. a. What is the average cycle time for this computer? b. A level 2 cache
In the context of multilevel caches, what is the difference between a local miss rate and a global miss rate?
Why is the miss rate often quoted (and used) in preference to the hit rate?
What is a victim cache? How is it used?
What type of misses does a victim cache reduce?
What are the essential differences between victim and annex caches?
A processor with memory management has a 4K page size. It has a 32K cache memory with 16-byte cache lines. In order to speed up memory access, you decide to arrange the cache so that the cache is
Suppose a unified cache has the following characteristics: Read/write penalty Miss rate Load instructions (read data) Store instructions (write data) Miss penalty What is the average access time? 1
A 64-bit processor has a 8-MB, four-way setassociative cache with 32-byte lines. How is the address arranged in terms of set, line, and offset bits?
A computer with a separate data cache has a writeback cache memory. Cache line size is 64 bytes. Read access accounts for 80% of memory traffic. The processor, memory, and data buses are all 64-bits
There are three causes of cache miss: compulsory, capacity, and conflict. Define the meaning of these terms. Briefly explain what can be done to minimize their effect.
Why is memory management necessary in a system that uses hard disks?
What forms of protection can memory management provide?
Memory management has a protection function. Does such a facility exist with cache memory?
What are the fundamental differences between cache memory (as found in a CPU) and cache memory found in a hard disk drive?
What are the differences between write-back and write-through caches, and what are the implications for system performance?
A computer with a 32-bit address architecture has a memory management system with single-level 4 KB page tables. How much memory space must be devoted to the page tables?
Consider a fully associative 16-byte cache with four lines of four words. The cache uses a LRU (least recently used) algorithm to deal with line replacement. When the cache is initially empty, lines
A computer runs an instruction set with the characteristics in the following table. What is the average number of cycles per instruction? Class Arithmetic operations Conditional operations Load
Consider the following code that accesses three values in memory scalar integers x and s, and an integer vector y[i]. What is the memory latency in clock cycles for a trip round the loop ( after the
A 64-bit computer has a 128 KB 8-way associatively mapped cache. The cache has 128 sets and a line is 16 words. How many tag bits does each address require?
What type of cache is particularly useful in reducing thrashing in the cache due to repetitive swapping in and swapping out?
A 16-bit CPU has a cache with 32 lines, each of 16 bytes. The CPU accesses a byte at the decimal address 3210. This results in a miss and a line is loaded. Where is the line loaded in the cache?
Given the following data and assuming a clock rate of 1,000 MHz. Calculate the average memory access time. Assume that L2 and DRAM are accessed in parallel with LL. Memory L1 cache L2 cache DRAM
Consider a computer with a 256-byte address space and a two-way 32-byte set associative cache. The computer word size is a byte, each cache line contains four bytes, and each cache four lines. If the
People are always looking for more effective cache mechanisms, particularly for ways of reducing the miss penalty (for example, using means of annex caches or victim caches). A student makes the
A computer has a cache with a hit ratio of 95% and a line size of four 32-bit words. The average processor cache access rate is 100 million/s. Twenty percent of CPU operations are loads/stores with
Consider Problem 9.54 except that the cache uses a write-back mode. On average 25% of cache lines are dirty (have been modified).Data in Problem 9.54,A computer has a cache with a hit ratio of 95%
A computer has a 256 word memory and a 16 word cache. The cache line size is one word. The following sequence of addresses is read in series: Show how the cache memory would be accessed assuming
A computer with a 24-bit address bus has a main memory of size 16 MB and a cache size of 64 KB. The word length is two bytes.a. What is the address format for a direct-mapped cache with a line size
A system has a memory access time of 50 ns and a cache access time of 2 ns. The instruction time is 4 ns (not counting memory access) and the average instruction requires 0.25 memory accesses. If the
Why is the hit rate of an L2 cache usually lower than that of an Ll cache?
A computer has a memory access time of 38 ns and does not use cache. Cache with an access time of 10 ns is added. The computer then runs 90% faster. Estimate the hit ratio.
A system has an L1 cache with a hit rate of 87% (hits takes one cycle). The L2 cache has a hit rate of 90% and a penalty of 10 cycles. The main store has an access time of 200 cycles. What is the
What is the meaning of the following terms (when applied to memory systems technology) ? a. Random access b. Serial access c. Dynamic RAM d. Static RAM e. Read-mostly memory f. Access time g.
A computer has a memory space of 1 MB. a. How many address lines are required to span this address space, assuming it is byte-addressed? b. If this computer has a 16-bit data bus and can access
What is the meaning of memory hierarchy, and why is this concept of importance to the designer of PCs and similar workstations?
A memory component spans the address range 0x00400000 to 0x007FFFFF. What is its capacity?
At the beginning of this chapter, we stated that volatile read-only memory was an oxymoron. However, there is one application where a volatile read-only memory might be very beneficial under certain
A designer can use two different static memory devices to construct a computer with a 16-bit CPU. Both memory chips can hold 222 bits. Suppose it is necessary to construct a computer with a 16-bit
A company designs a computer in 2012. At its launch, the memory is twice as fast as the CPU. It is estimated that each year the CPU will get faster by 20%. Equally, it is estimated that each year the
Why is all ROM RAM but not all RAM ROM?
Why is static RAM more suited to cache memory than DRAM?
What is the typical amount of main store, cache, and hard disk provided by current high-performance, state-of-the-art, cutting-edge personal computers?
We state that a static memory cell requires at least four transistors and a DRAM cell can use one transistor per bit. Could a memory ever use less than one transistor per bit to store data? Can you
Given the timing diagram in Figure P10.12 that is derived from the data sheet of a commercial SRAM, explain its action in words. FIGURE P10.12 CLK ADDR Chip Select CS or SS Write Enable (WE) Output
Why does a DRAM have fewer pins (1/0 connections) than a static RAM of the same size (i.e., capacity in bits)?
In a conventional DRAM, what is the function of the column address and row address strobes?
In the context of DRAM timing, what is a pseudo maximum? What DRAM parameter is a pseudomaximim? What are the implications of a pseudo maximum for design engineers?
What fundamental properties of matter are exploited to implement memory systems? Can you think of any properties that have not yet been exploited and suggest ways in which they may be used to
What are the differences between Ovonic and ferroelectric memories?
A DDR SDRAM module in a PC is clocked at 133 MHz and is interfaced to a 64-bit data bus. What is the designation of the memory system in terms of the PCxx00 standard?
The speed of DRAM is increasing by about 7% a year, while the performance of processors is increasing by about 60%. What is the consequence of this state of affairs in the short term? What is the
You have a microprocessor chip with a 16-bit data bus. The computer accesses words; that is, it performs only 16-bit reads and writes. You are going to build a minimal computer which requires the
A computer has a 64-bit data word and a 32-bit address. Interleave addressing is used with four banks. The total amount of memory is 1,024 MB. Show how the processor's 32-address bits are partitioned
A computer with a 64-bit data bus uses the following memory chips. In each case, the chip is specified by locations X data width. For each of these chips, state the minimum number of chips required
A DDR3 DRAM is specified as 9-9-9-24. What does this mean?
Suppose you could include cache memory in DRAM chips. How would you organize it, and what would the advantages be? What changes might have to be made to the computer system architecture?
A CPU with a 24-bit address bus and 16-bit data bus implements the following memory blocks: Design an address decoder to implement this arrangement. 1 M byte of ROM using 256K X 8-bit chips 8 M
It has been reported that flash memory was reaching the limit of its density. Why do you think that this may be so?
What is wear leveling, and why does it have to be undertaken?
You have been asked to design the on-board computer for a deep-space vehicle. What special considerations would you have to take into consideration, and how would these affect your design?
What limits the areal density of a hard disk drive?
This chapter refers to an article on magnetic storage in the December 2000 edition of / EEE Spectrum with the title "Magnetic Storage: The medium that wouldn't die. " The title implies th at magnetic
Explain the following terms as they are applied to magnetic storage technology. a. Domain b. Permeability c. M agnetostriction d. Remnant field e. Ferromagnetic f. Curie temperature g.
What are the major obstacles faced by the designer of hard drives?
Making hard drives for laptops and portable computers imposes a different set of design criteria on the engineer. What are the special considerations that the designer of mobile hard disks has to
Hard disk drives once used combined inductive read/ write heads. Today, high-performance drives have GMR read heads. a. What is a GMR read-head? b. What are the advantages of separate read and
Why is it difficult to fully erase data from a magnetic medium such as tape or disk?
Why does data have to be encoded before it can be written to a disk?
The following data sequence is to be MFM encoded for recording on a diskette. Draw the resulting waveform that would be presented to the write head. 0101001100001111
What are the criteria that determine the optimum size of a sector? Are these criteria permanent or do they change with time? What other developments in computing affect the answer to this question?
A hypothetical disk drive had the following parameters. The seek time for a short seek is given as 1 + 0.2Jn ms, and the seek time for a long seek is 3 + 0.003n ms. Assume that short seeks are less
A hard disk has the following parameters. • 6 surfaces • 20K tracks per surface • 256 sectors per track • 512 bytes per sector • Rotational speed 7,200 rpma. What is the capacity of
How can disk drives be made faster (i.e., what options are available to the manufacturer)?
Describe the following disk head scheduling algorithms: FIFO, SSTF, SCAN, and C-SCAN.
A hypothetical disk drive has the following properties. What is the approximate time required to access the following sectors, assuming that a new seek is not required if successive sectors are on
What is SMART technology, and how can it help the corporate hard disk user?
The most popular (i.e., commercially successful) RAID levels are 1 and 5. Why?
If the probability of failure of a single drive is 0.000001 per 1000 hours, what is the probability of failure of exactly one drive in an array of four? a. One drive in an array of four. b. Two
When calculating the reliability of systems with multiple disks such as RAID arrays, it is conceptually obvious that, if one drive has a probability of failure of one in a thousand, the probability
Why is a CD so relatively insensitive to the effect of contamination by dust particles that are far larger than the bits stored on the disk?
Why is a laser required to read data from an optical disk?
Why is it necessary to use a code such as 8-to-14 to encode data bits before recording them on a CD?
What is the size of a spot on the recording surface of a CD if the wavelength of the laser light is 780 nm and the numerical aperture is 0.45?
What is the difference between constant linear and constant angular velocity? Why is one used in magnetic recording and the other in optical recording?
A disk rotates at 7,200 rpm, and the head is over a track 2 in. from the center. What are the corresponding angular and linear velocities?
How are writable CDs implemented?
If magnetic tape (pre-1980s) was ½ in. wide and stored data in nine tracks at 128 characters/in., what was the areal density in bits per square inch? What is the areal density ratio between this and
In 1996, NEC announced a 64-MB flash memory chip with a 98 mm2 die size. What is the corresponding areal density in bits/in.2?
By 2002, a 125-mm2, 1-Gb flash memory had been announced. What is the corresponding areal density in bits/in.2?
In September 2010, Toshiba launched the first embedded flash memory modules (i.e., for incorporation on circuit boards rather than as plug-in modules). These devices had a capacity of 128 GB in a 17
What is the minimum spot size of a DVD laser if its wavelength is 650 nm and the lens has a numerical aperture of 0.6?
What is the minimum spot size of a Blu-ray laser if its wavelength is 405 nm and the lens has a numerical aperture of 0.85?

Showing 200 - 300 of 1374