The slab allocation algorithm uses a separate cache for each different object type. Assuming there is one cache per object type, explain why this doesn't scale well with multiple CPUS. What could be done to address this scalability issue?

The slab allocation algorithm uses a separate cache for each different object type. Assuming there is one cache per object type, explain why this doesn't scale well with multiple CPUS. What could be done to address this scalability issue?

Database System Concepts

7th Edition

ISBN:9780078022159

Author:Abraham Silberschatz Professor, Henry F. Korth, S. Sudarshan

Publisher:Abraham Silberschatz Professor, Henry F. Korth, S. Sudarshan

Chapter1: Introduction

Section: Chapter Questions

Problem 1PE

See similar textbooks

Similar questions

Consider a 32-bit microprocessor, with a 16-bit external data bus, driven by an 8-MHz input clock. Assume that this microprocessor has a bus cycle whose minimum duration equals four input clock cycles. What is the maximum data transfer rate across the bus that this microprocessor can sustain in bytes/s? To increase its performance, would it be better to make its external data bus 32 bits or to double the external clock fre- quency supplied to the microprocessor? State any other assumptions you make and explain. Hint: Determine the number of bytes that can be transferred per bus cycle.
For the cache design of the preceding problem, suppose that increasing the line size from one word to four words results in a decrease of the read miss rate from 3.2% to 1.1%. For both the nonburst transfer and the burst transfer case, what is the average miss penalty, averaged over all reads, for the two different line sizes?
The hardware used to replicate Pep/9 has 64K memory. But what is its purpose? So far, just 0x0030 of memory has been written. What gives with all that additional RAM, then? How does each memory region in Pep/9 know what to do with it? Another viewpoint There have only been "global" variables utilised. What impact would using "local" variables have? What distinguishes them? (Think of "stack").
Consider a memory of size 8KB (8192 bytes) that allows dynamic, variable sized partitioning among processes and uses a linked list to keep track of free spaces (hereafter referred to as the free list) in the memory at any given time. Assume that there are 6 processes and assume that their memory size requirements (in bytes) are as given below: Р1: 500, Р2: 600, Р3: 1300, Р4: 2000, Р5: 100, Р6: 200 Assume that the initial state of the free list is as shown below (BA is the base address and Sz is the size of each free space): ВА: 0; Sz: 1100 — ВА: 1200; Sz: 600 > BА: 2000; Sz: 1800> ВА: 6000%; Sz 400
We consider a virtual memory design for a computer architecture. Suppose that the byte- addressed computer has 256 bytes pages, 16-bit virtual addresses, and 16-bit page table entries (PTE) (NUM 1 bits of one PTE are used for reference, valid checking, protection mechanism, etc.). The computer uses two-level hierarchical page tables. 1. How many virtual pages can be addressed by the architecture? 2. What is the maximal size of the physical memory that can be supported by the architec- ture? 3. We are now running a program which is using 300 bytes of memory currently. What is the smallest possible number of PTES to store the running program?
A system with simple paged memory management has addresses of B bits and pages of size P KiB. If all memory is pageable, what is the maximum number of items contained in a process's page table? What if K KiB are reserved for the kernel and are not pageable? Also suppose that the page table is organized in 2 levels, so that the page directory (first level) is addressed by D bits. How many bits will be used to address the second level? How many entries will there be, at most, among all the tables of the pages (first and second level) of a process?
The hardware used to replicate Pep/9 has 64K memory. But what is its purpose? So far, just 0x0030 of memory has been written. What gives with all that additional RAM, then? How does each memory region in Pep/9 know what to do with it? Another viewpoint There have only been "global" variables utilised. What impact would using "local" variables have? What distinguishes them? (Think of "stack").
Imagine that you have implemented an algorithm in a program which has 40% parallelizable code. Now consider the following two cases assuming that scheduling overhead is ignored: -The program is executed on a single processor. -The program is executed on a parallel system with 4 processors. Compute the speed-up gained in the case where multiple processors are used.
We will explore the impact of cache capacity on performance, focusing exclusively on the data cache and excluding instruction storage in the caches. Cache access time is directly linked to its capacity. For the sake of simplicity, let's assume that accessing the main memory takes 100ns, and in a specific program, 50% of instructions involve data access. Two distinct processors, denoted as P1 and P2, are engaged in executing this program. Each processor is equipped with its own L1 cache. L1 size L1 Miss Rate L1 Hit Time P1 64 KB 3.6% 1.26 ns P2 128 KB 3.1% 2.17ns (a) What is the AMAT for P1 and P2 assuming no other levels of cache?
The slab allocation algorithm uses a separate cache for each different object type. Assuming there is one cache per object type, explain why this doesn't scale well with multiple CPUS. What could be done to address this scalability issue?
The simulated Pep/9 hardware has 64K of memory. But how is it put to use? So far, just 0x0030 memory has been written. So, what's the deal with all of that additional RAM? Pep/9 knows what to do with each memory location. Another viewpoint Only "global" variables were used. What effect would using "local" variables have? What distinguishes them? (Think "stack").