Why Threads Are A Bad Idea

Download Report

Transcript Why Threads Are A Bad Idea

Why Threads Are A Bad Idea
(for most purposes)
John Ousterhout
Sun Microsystems Laboratories
[email protected]
http://www.sunlabs.com/~ouster
Introduction




Threads:
– Grew up in OS world (processes).
– Evolved into user-level tool.
– Proposed as solution for a variety of problems.
– Every programmer should be a threads programmer?
Problem: threads are very hard to program.
Alternative: events.
Claims:
– For most purposes proposed for threads, events are
better.
– Threads should be used only when true CPU
concurrency is needed.
Why Threads Are A Bad Idea
September 28, 1995, slide 2
What Are Threads?
Shared state
(memory, files, etc.)
Threads

General-purpose solution for managing concurrency.

Multiple independent execution streams.

Shared state.

Pre-emptive scheduling.

Synchronization (e.g. locks, conditions).
Why Threads Are A Bad Idea
September 28, 1995, slide 3
What Are Threads Used For?

Operating systems: one kernel thread for each user
process.

Scientific applications: one thread per CPU (solve
problems more quickly).

Distributed systems: process requests concurrently
(overlap I/Os).

GUIs:
– Threads correspond to user actions; can service
display during long-running computations.
– Multimedia, animations.
Why Threads Are A Bad Idea
September 28, 1995, slide 4
What's Wrong With Threads?
casual
wizards
all programmers
Visual Basic programmers
C programmers
C++ programmers
Threads programmers

Too hard for most programmers to use.

Even for experts, development is painful.
Why Threads Are A Bad Idea
September 28, 1995, slide 5
Why Threads Are Hard

Synchronization:
– Must coordinate access to shared data with locks.
– Forget a lock? Corrupted data.

Deadlock:
– Circular dependencies among locks.
– Each process waits for some other process: system
hangs.
thread 1
Why Threads Are A Bad Idea
lock A
lock B
thread 2
September 28, 1995, slide 6
Why Threads Are Hard, cont'd



Hard to debug: data dependencies, timing dependencies.
Threads break abstraction: can't design modules
independently.
Callbacks don't work with locks.
T1
T2
deadlock!
Module A
T1
calls
Module A
deadlock!
Module B
Module B
callbacks
sleep
wakeup
Why Threads Are A Bad Idea
T2
September 28, 1995, slide 7
Why Threads Are Hard, cont'd



Achieving good performance is hard:
– Simple locking (e.g. monitors) yields low concurrency.
– Fine-grain locking increases complexity, reduces
performance in normal case.
– OSes limit performance (scheduling, context switches).
Threads not well supported:
– Hard to port threaded code (PCs? Macs?).
– Standard libraries not thread-safe.
– Kernel calls, window systems not multi-threaded.
– Few debugging tools (LockLint, debuggers?).
Often don't want concurrency anyway (e.g. window
events).
Why Threads Are A Bad Idea
September 28, 1995, slide 8
Event-Driven Programming

One execution stream: no CPU
concurrency.

Register interest in events
(callbacks).

Event loop waits for events,
invokes handlers.

No preemption of event
handlers.

Handlers generally short-lived.
Why Threads Are A Bad Idea
Event
Loop
Event Handlers
September 28, 1995, slide 9
What Are Events Used For?

Mostly GUIs:
– One handler for each event (press button, invoke menu
entry, etc.).
– Handler implements behavior (undo, delete file, etc.).

Distributed systems:
– One handler for each source of input (socket, etc.).
– Handler processes incoming request, sends response.
– Event-driven I/O for I/O overlap.
Why Threads Are A Bad Idea
September 28, 1995, slide 10
Problems With Events




Long-running handlers make application nonresponsive.
– Fork off subprocesses for long-running things (e.g.
multimedia), use events to find out when done.
– Break up handlers (e.g. event-driven I/O).
– Periodically call event loop in handler (reentrancy adds
complexity).
Can't maintain local state across events (handler must
return).
No CPU concurrency (not suitable for scientific apps).
Event-driven I/O not always well supported (e.g. poor
write buffering).
Why Threads Are A Bad Idea
September 28, 1995, slide 11
Events vs. Threads

Events avoid concurrency as much as possible, threads
embrace:
– Easy to get started with events: no concurrency, no
preemption, no synchronization, no deadlock.
– Use complicated techniques only for unusual cases.
– With threads, even the simplest application faces the
full complexity.

Debugging easier with events:
– Timing dependencies only related to events, not to
internal scheduling.
– Problems easier to track down: slow response to button
vs. corrupted memory.
Why Threads Are A Bad Idea
September 28, 1995, slide 12
Events vs. Threads, cont'd

Events faster than threads on single CPU:
– No locking overheads.
– No context switching.

Events more portable than threads.

Threads provide true concurrency:
– Can have long-running stateful handlers without
freezes.
– Scalable performance on multiple CPUs.
Why Threads Are A Bad Idea
September 28, 1995, slide 13
Should You Abandon Threads?

No: important for high-end servers (e.g. databases).

But, avoid threads wherever possible:
– Use events, not threads, for GUIs,
distributed systems, low-end servers.
– Only use threads where true CPU
concurrency is needed.
– Where threads needed, isolate usage
in threaded application kernel: keep
most of code single-threaded.
Why Threads Are A Bad Idea
Event-Driven Handlers
Threaded Kernel
September 28, 1995, slide 14
Conclusions

Concurrency is fundamentally hard; avoid whenever
possible.

Threads more powerful than events, but power is
rarely needed.

Threads much harder to program than events; for
experts only.

Use events as primary development tool (both GUIs
and distributed systems).

Use threads only for performance-critical kernels.
Why Threads Are A Bad Idea
September 28, 1995, slide 15
Discussions

Slide 8 says "Threads not well supported:"These days
POSIX Threads library (Pthreads) is available even
on Windows! Does this weaken the Presenter
argument against threads?

I however disagree with Slide 15's point "Threads
more powerful than events, but power is rarely
needed". Cores are increasing. GHz/Core will
stabilize/decrease. To get the same performance in
future we will need threads! Daily Desktop
applications like encoding, rendering can always use
more power (more cores and more threads).
Comments?
Why Threads Are A Bad Idea
September 28, 1995, slide 16
Discussions

This paper was written in 1995. Today, 14 years later,
with the coming of CPU concurrency even in low-end
consumer machines, is the paper still relevant, or
should we be looking to use threads everywhere we
can?

If a malicious user tries to intercept the event, then
the system will be totally out of control. Am I right?

suppose there are several processors this time. When
each event loop runs at the respective processor, and
some of them use the same resources, then, the
synchronized problems are still there. Am I right?
Why Threads Are A Bad Idea
September 28, 1995, slide 17
Discussions

The paper claims that concurrency is fundamentally
hard and should be avoided whenever possible. Do we
buy this?

How does event mechanism operate in order to avoid
using a lock? How does it handle the situation that,
for example, two events occur at the same time, both
of which want to access the same address space?
Why Threads Are A Bad Idea
September 28, 1995, slide 18