Transcript Competing For Memory Vivek Pai / Kai Li Princeton University

Competing For Memory
Vivek Pai / Kai Li
Princeton University
Mechanics

Feedback optionally anonymous



Quiz 1 Question 2 answer(s)




No real retribution anyway
Do it to make me happy
#regs != #bits
Registers at top of memory hierarchy
Lots of acceptable answers
Last Quiz, Feedback still being digested
2
The Big Picture
We’ve talked about single evictions
 Most computers are multiprogrammed





Single eviction decision still needed
New concern – allocating resources
How to be “fair enough” and achieve good
overall throughput
This is a competitive world – local and
global resource allocation decisions
3
Lessons From Enhanced FIFO

Observations




it’s easier to evict a clean page than a dirty page
sometimes the disk and CPU are idle
Optimization: when system’s “free”, write
dirty pages back to disk, but don’t evict
Called flushing – often falls to pager daemon
4
x86 Page Table Entry
Page frame number U P Cw Gl L D A Cd Wt O W V
31
12
Reserved
Valid
Writable
Owner (user/kernel)
Write-through
Cache disabled
Accessed (referenced)
Dirty
PDE maps 4MB
Global
5
Program Behaviors
80/20 rule


Locality


> 80% memory references
are made by < 20% of
code
Spatial and temporal
# page faults

Working set
Working set

Keep a set of pages in
memory would avoid a lot
of page faults
# pages in memory
6
Observations re Working Set
Working set isn’t static
 There often isn’t a single “working set”




Multiple plateaus in previous curve
Program coding style affects working set
Working set is hard to gauge

What’s the working set of an interactive
program?
7
Working Set

Main idea


Keep the working set in memory
An algorithm



On a page fault, scan through all pages of the
process
If the reference bit is 1, record the current time
for the page
If the reference bit is 0, check the “last use time”



If the page has not been used within d, replace the page
Otherwise, go to the next
Add the faulting page to the working set
8
WSClock Paging Algorithm



Follow the clock hand
If the reference bit is 1, set reference bit to 0,
set the current time for the page and go to
the next
If the reference bit is 0, check “last use time”


If page has been used within d, go to the next
If page hasn’t been used within d and modify bit is 1


Schedule the page for page out and go to the next
If page hasn’t been used within d and modified bit is 0

Replace this page
9
Simulating Modify Bit with
Access Bits



Set pages read-only if they are read-write
Use a reserved bit to remember if the page is
really read-only
On a read fault


If it is not really read-only, then record a modify in
the data structure and change it to read-write
Restart the instruction
10
Implementing LRU without
Reference Bit

Some machines have no reference bit


VAX, for example
Use the valid bit or access bit to simulate



Invalidate all valid bits (even they are valid)
Use a reserved bit to remember if a page is really
valid
On a page fault



If it is a valid reference, set the valid bit and place the
page in the LRU list
If it is a invalid reference, do the page replacement
Restart the faulting instruction
11
Demand Paging


Pure demand paging relies only on faults to
bring in pages
Problems?



Remedies



Possibly lots of faults at startup
Ignores spatial locality
Loading groups of pages per fault
Prefetching/preloading
So why use it?
12
Speed and Sluggishness



Slow is > .1 seconds (100 ms)
Speedy is << .1 seconds
Monitors tend to be 60+ Hz =
<16.7ms between screen paints

Disks have seek + rotational delay



Seek is somewhere between 7-16 ms
At 7200rpm, one rotation = 1/120 sec = 8ms.
Half-rotation is 4ms
Conclusion? One disk access OK, six are bad
13
Memory Pressure
“Swap” space
 Region of disk used to hold “overflow”
 Contains only data pages
(stack/heap/globals). Why?
 Swap may exist as “regular file,” but
dedicated region of disk more common
14
Disk Address


Use physical memory
as a cache for disk
Where to find a page
on a page fault?


Virtual
address
space
invalid
Physical
memory
PPage# field is a disk
address
Observation: OS
knows that pages are
real but not in
memory
15
Imagine a Global LRU
Global – across all processes
 Idea – when a page is needed, pick the
oldest page in the system
 Problems? Process mixes?




Interactive processes
Active large-memory sweep processes
Mitigating damage?
16
Source of Disk Access

VM System


Filesystem


Main memory caches - full image on disk
Even here, caching very useful
New competitive pressure/decisions


How do we allocate memory to these two?
How do we know we’re right?
17
Partitioning Memory

Originally, specified by administrator




20% used as filesystem cache by default
On fileservers, admin would set to 80%
Each subsystem owned pages, replaced
them
Observation: they’re all basically pages


Why not let them compete?
Result: unified memory systems – file/VM
18
File Access Efficiency

read(fd, buf, size)




Buffer in process’s memory
Data exists in two places – filesystem cache &
process’s memory
Known as “double buffering”
Various scenarios


Many processes read same file
Process wants only parts of a file, but doesn’t
know which parts in advance
19
Result: Memory-Mapped Files
Process A Process B Process C
File
File
File
Process A Process B Process C
Map
Map
Map
File
20
Lazy Versus Eager

Eager: do things right away




read(fd, buf, size) – returns # bytes read
Bytes must be read before read completes
What happens if size is big?
Lazy: do them as they’re needed




mmap(…) – returns pointer to mapping
Mapping must exist before mmap completes
When/how are bytes read?
What happens if size is big?
21
Semantics: How Things Behave

What happens when





Two process obtain data (read or mmap)
One process modifies data
Two processes obtain data (read or mmap)
A third process modifies data
The two processes access the data
22
Being Too Smart…
Assume a unified VM/File scheme
 You’ve implemented perfect Global LRU


What happens on a filesystem “dump”?
23
Amdahl’s Law
Gene Amdahl (IBM, then Amdahl)
 Noticed the bottlenecks to speedup
 Assume speedup affects one
component
 New time =
(1-not affected) + affected/speedup
 In other words, diminishing returns

24
NT x86 Virtual Address Space Layouts
00000000
Application code
Globals
Per-thread stacks
DLL code
3-GB user space
7FFFFFFF
80000000
Kernel & exec
HAL
Boot drivers
C0000000 Process page tables
Hyperspace
C0800000
System cache
Paged pool
Nonpaged pool
FFFFFFFF
BFFFFFFF
C0000000
1-GB system space
FFFFFFFF
25
Virtual Address Space in Win95 and Win98
00000000
User accessible
7FFFFFFF
80000000 Shared, process-writable
(DLLs, shared memory,
Win16 applications)
C0000000
FFFFFFFF
Unique per process
(per application),
user mode
Systemwide
user mode
Win95 and Win98
Operating system
(Ring 0 components)
Systemwide
kernel mode
26
Details with VM Management

Create a process’s virtual address space




Allocate page table entries (reserve in NT)
Allocate backing store space (commit in NT)
Put related info into PCB
Destroy a virtual address space



Deallocate all disk pages (decommit in NT)
Deallocate all page table entries (release in NT)
Deallocate all page frames
27
Page States (NT)








Active: Part of a working set and a PTE points to it
Transition: I/O in progress (not in any working sets)
Standby: Was in a working set, but removed.
A PTE points to it, not modified and invalid.
Modified: Was in a working set, but removed.
A PTE points to it, modified and invalid.
Modified no write: Same as modified but no write
back
Free: Free with non-zero content
Zeroed: Free with zero content
Bad: hardware errors
28
Dynamics in NT VM
Demand
zero fault
Page in or allocation
Standby
list
Process
“Soft”
working faults
set
Modified
writer
Free
list
Zero
thread
Zero
list
Bad
list
Modified
list
Working set
replacement
29
Shared Memory

How to destroy a
virtual address space?



How to swap out/in?



Link all PTEs
Reference count

.
.
.
..
.
Page table
Process 1
Link all PTEs
Operation on all entries
How to pin/unpin?

w
Link all PTEs
Reference count
.
.
.
w
.
.
.
..
.
Physical
pages
Page table
Process 2
30
Copy-On-Write





Child’s virtual address
space uses the same
page mapping as
parent’s
Make all pages read-only
Make child process ready
On a read, nothing
happens
On a write, generates an
access fault



map to a new page
frame
copy the page over
restart the instruction
r
r
.
.
.
..
.
Page table
Parent process
r
r
.
.
.
..
.
.
.
.
Physical
pages
Page table
Child process
31
Issues of Copy-On-Write

How to destroy an address space


How to swap in/out?


Same as shared memory case?
Same as shared memory
How to pin/unpin

Same as shared memory
32