CRAFTING
A READY-TO-GO STM

Guy Korland
Tel Aviv University
Process 1

a = acc.get()
a = a + 100
acc.set(a)

Process 2
b = acc.get()
b = b + 50
acc.set(b)

... Lost Update! ...
Process 1

Process 2

lock(A)
lock(B)

lock(B)
lock(A)

... Deadlock! ...
Process 1

Process 2

atomic{
a = acc.get()
a = a + 100
acc.set(a)
}

atomic{
b = acc.get()
b = b + 50
acc.set(b)
}

... WIIIII ...
DSTM2
MAURICE HERLIHY ET AL, A FLEXIBLE FRAMEWORK … [OOPSLA06]

@atomic public interface INode{
int getValue ();
void setValue (int value );
}
Factory<INode> factory = Thread.makeFactory(INode.class );
final INode node = factory.create();
factory result = Thread.doIt(new Callable<Boolean>() {
public Boolean call () {
return node.setValue(value);
} });
JVSTM
JOÃO CACHOPO AND ANTÓNIO RITO-SILVA, VERSIONED BOXES AS THE BASIS
FOR MEMORY TRANSACTIONS [SCOOL05]

public class Account{
private VBox<Long> balance = new VBox<Long>();

public @Atomic void withdraw(long amount) {
balance.put (balance.get() - amount);
}
}
DEUCE STM - API
Guy Korland, Nir Shavit and Pascal Felber, “Noninvasive Java
Concurrency with Deuce STM”, [MultiProg '10]

public class Bank{
private double commission = 10;

@Atomic
public void transaction( Account ac1, Account ac2, double amount){
ac1.balance -= (amount + commission);
ac2.balance += amount;
}
}
DEUCE STM - OVERVIEW
BENCHMARKS
(SUN ULTRASPARC T2 PLUS – 2 X QUAD X 8HT)
BENCHMARKS
(AZUL – VEGA2 – 2 X 48)
BENCHMARK - THE DARK SIDE
1.2
1
0.8
0.6

Lock
Deuce

0.4
0.2
0

1

2

3

4

5

6

7

8

9

10
OVERHEAD
• Contention – Retries, Aborts, Contention Manager …
• STM Algorithm – Data structures, optimistic, pessimistic…
• Semantic – Consistency model, Privatization…
• Instrumented Memory access – Linear overhead on every
read/write
STATIC ANALYSIS
OPTIMIZATIONS
1. Avoiding instrumentation of accesses to
immutable and transaction-local memory.
2. Avoiding lock acquisition and releases for
thread-local memory.
3. Avoiding readset population in
read-only transactions.
NEW STATIC ANALYSIS
OPTIMIZATIONS
Yehuda Afek, Guy Korland, and Arie Zilberstein,
“Lowering STM Overhead with Static Analysis”, LCPC'10

1. Reduce amount of instrumented memory reads using
load elimination.
2. Reduce amount of instrumented memory writes using
scalar promotion.
3. Avoid writeset lookups for memory
not yet written to.
4. Avoid writeset record keeping for memory that
will not be read.
5. Reduce false conflicts by Transaction re-scoping.
6. …
BENCHMARKS – K-MEANS
STM FRIENDLY LIBRARIES
• Most of the developers/application only use Data Structures.
@Atomic(retries=64)
public void transaction( Map map1, Map map2, String key){
map2.put(key, map1.remove(key));
}

 We should build a set of “STM Friendly” libraries.
Maurice Herlihy, Eric Koskinen: Transactional boosting:
a methodology for highly-concurrent transactional objects. PPOPP’08
WWW.DEUCESTM.ORG

Crafting a Ready-to-Go STM

  • 1.
    CRAFTING A READY-TO-GO STM GuyKorland Tel Aviv University
  • 4.
    Process 1 a =acc.get() a = a + 100 acc.set(a) Process 2 b = acc.get() b = b + 50 acc.set(b) ... Lost Update! ...
  • 5.
  • 7.
    Process 1 Process 2 atomic{ a= acc.get() a = a + 100 acc.set(a) } atomic{ b = acc.get() b = b + 50 acc.set(b) } ... WIIIII ...
  • 9.
    DSTM2 MAURICE HERLIHY ETAL, A FLEXIBLE FRAMEWORK … [OOPSLA06] @atomic public interface INode{ int getValue (); void setValue (int value ); } Factory<INode> factory = Thread.makeFactory(INode.class ); final INode node = factory.create(); factory result = Thread.doIt(new Callable<Boolean>() { public Boolean call () { return node.setValue(value); } });
  • 10.
    JVSTM JOÃO CACHOPO ANDANTÓNIO RITO-SILVA, VERSIONED BOXES AS THE BASIS FOR MEMORY TRANSACTIONS [SCOOL05] public class Account{ private VBox<Long> balance = new VBox<Long>(); public @Atomic void withdraw(long amount) { balance.put (balance.get() - amount); } }
  • 11.
    DEUCE STM -API Guy Korland, Nir Shavit and Pascal Felber, “Noninvasive Java Concurrency with Deuce STM”, [MultiProg '10] public class Bank{ private double commission = 10; @Atomic public void transaction( Account ac1, Account ac2, double amount){ ac1.balance -= (amount + commission); ac2.balance += amount; } }
  • 12.
    DEUCE STM -OVERVIEW
  • 13.
    BENCHMARKS (SUN ULTRASPARC T2PLUS – 2 X QUAD X 8HT)
  • 14.
  • 17.
    BENCHMARK - THEDARK SIDE 1.2 1 0.8 0.6 Lock Deuce 0.4 0.2 0 1 2 3 4 5 6 7 8 9 10
  • 19.
    OVERHEAD • Contention –Retries, Aborts, Contention Manager … • STM Algorithm – Data structures, optimistic, pessimistic… • Semantic – Consistency model, Privatization… • Instrumented Memory access – Linear overhead on every read/write
  • 20.
    STATIC ANALYSIS OPTIMIZATIONS 1. Avoidinginstrumentation of accesses to immutable and transaction-local memory. 2. Avoiding lock acquisition and releases for thread-local memory. 3. Avoiding readset population in read-only transactions.
  • 21.
    NEW STATIC ANALYSIS OPTIMIZATIONS YehudaAfek, Guy Korland, and Arie Zilberstein, “Lowering STM Overhead with Static Analysis”, LCPC'10 1. Reduce amount of instrumented memory reads using load elimination. 2. Reduce amount of instrumented memory writes using scalar promotion. 3. Avoid writeset lookups for memory not yet written to. 4. Avoid writeset record keeping for memory that will not be read. 5. Reduce false conflicts by Transaction re-scoping. 6. …
  • 22.
  • 23.
    STM FRIENDLY LIBRARIES •Most of the developers/application only use Data Structures. @Atomic(retries=64) public void transaction( Map map1, Map map2, String key){ map2.put(key, map1.remove(key)); }  We should build a set of “STM Friendly” libraries. Maurice Herlihy, Eric Koskinen: Transactional boosting: a methodology for highly-concurrent transactional objects. PPOPP’08
  • 25.