financial portfolio management with java on steroids - jax finance 2016
TRANSCRIPT
Marcus Gründler | aixigo AG
Financial Portfolio Management on Steroids
About meMarcus Gründler@marcusgruendler
Head of Portfolio Management SystemsArchitect at aixigo AG, Germany
www.aixigo.de
JAX Finance 2016 - London
Agenda
• Financial portfolio management
• Hardware
• Java memory
• Programming patterns
• Scaling
What exactly is
Financial PortfolioManagement?
Portfolio Management
• Transactions, prices, securities
• Financial algorithms
• Historical analysis
• Time series
Portfolio Management
• No extreme low latency
• High data throughput (1 mio rec/sec)
• Low response times (100ms/10,000 rec)
• Very large datasets
Ishigh performancecomputingpossible withJava?
Yes ...
... read the fine print!
Matrix Sum23 101 2 34 88 120 4
44 12 234 211 112 189 11
33 1 86 201 3 11 22
65 32 62 22 34 15 67
43 178 105 138 192 38 41
11 58 35 25 27 16 21
Row major access
Matrix Sum
23 101 2 34 88 120 4
44 12 234 211 112 189 11
33 1 86 201 3 11 22
65 32 62 22 34 15 67
43 178 105 138 192 38 41
11 58 35 25 27 16 21
Column major access
Matrix Sum(10,000 x 10,000 elements)
ops/sec
ops/sec
ops/sec
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Matrix Sum
Row major access
Column major access
Tool Support - JMH
• OpenJDK JMH (Java Microbenchmark Harness)
• Eliminates measurement (in)accuracy
• Statistically robust measurements
• Maven and Jenkins support
http://openjdk.java.net/projects/code-tools/jmh/
Memory Access
CPU
RAM
Cached Memory Access
CPU
RAM
Cache
Multi Level Caches
CPU
RAM
L1 L2 L3
Latency NumbersCPU cycles Time Size
L1 latency ~ 4-5 cycles 1.5 ns 32 KBL2 latency ~ 12 cycles 3.5 ns 256 KBL3 latency ~ 36 cycles 10.6 ns 8-40 MBRAM latency ~ 230 cycles 67.6 ns 256 GB2KB over 1GBit 20,000 ns1 MB from RAM 250,000 nsDisk seek 10,000,000 nsRoundtrip US-EU-US 150,000,000 nsIntel i7-4770 (Haswell), 3.4 GHz Sources:http://www.7-cpu.com/cpu/Haswell.html
http://norvig.com/21-days.html#answers
Latency NumbersCPU cycles Time Size
L1 latency ~ 4-5 cycles 1.5 ns 32 KBL2 latency ~ 12 cycles 3.5 ns 256 KBL3 latency ~ 36 cycles 10.6 ns 8-40 MBRAM latency ~ 230 cycles 67.6 ns 256 GB2KB over 1GBit 20,000 ns1 MB from RAM 250,000 nsDisk seek 10,000,000 nsRoundtrip US-EU-US 150,000,000 nsIntel i7-4770 (Haswell), 3.4 GHz Sources:http://www.7-cpu.com/cpu/Haswell.html
http://norvig.com/21-days.html#answers
Cache-oblivious Algorithms
• Optimized for minimal memory transfer
• All computation with L1 cache
• Cache-“oblivious“: no knowledge about
cache hierarchy
• Keeps CPU permanently „under pressure“
• Empowers cache prefetching
Plain Matrix Transpose
1 2 34 5 6
...
1 4...2 5
637
7
Cache-oblivious Transpose
Matrix Transpose(4096 x 4096 elements)
ops/sec
Matrix Transpose(4096 x 4096 elements)
Cores
ops/sec
• Memory access patterns matter
• Avoid main memory jumps
• Algorithms should support prefetching
How muchmemorydo weconsume?
How large is an object?
double = 8 byteDouble = ? byte
BigDecimal = ? byte
java.lang.Double
double = 8 byteDouble = ? byte
BigDecimal = ? byte
0 4 8 12 16 20 24 28
Padding8 bytealignment
Objectheader
„Mark word“• HashCode• Flags
Class pointer
Data
java.lang.Double
double = 8 byteDouble = ? byte
BigDecimal = ? byte
0 4 8 12 16 20 24 28
Objectheader
„Mark word“• HashCode• Flags
Class pointer
Data
double = 8 byteDouble = 24 byte
BigDecimal = ? byte
double array0 4 8 12 16 20 24 28
„Mark word“• HashCode• Flags
Class pointer
Data 0 Data 1Arraylength
...
double = 8 byteDouble = 24 byte
BigDecimal = ? byte
java.lang.Double
double = 8 byteDouble = 24 byte
BigDecimal = ? byte
BigDecimal0 4 8 12 16 20 24 28 32 36 40
ref to BigInteger
ref to int[ ]
Σ 40 byte
Σ 40 byte
Σ >16 byte>96 byte
double = 8 byteDouble = 24 byte
BigDecimal = >96 byte
BigDecimal
Tool Support - JOL• OpenJDK tool – JOL (Java Object Layout)
• Insight into memory layout
• Heap dump analysis
• Exact memory usage
• Graphical layout visualization
• Maven module
http://openjdk.java.net/projects/code-tools/jol/
Tool Support - JOL
java -jar jol-cli.jar \internals java.math.BigDecimal
or
java –cp jol-cli.jar:my-own.jar \org.openjdk.jol.Main \internals foo.MyClass
Tool Support - JOL
java.math.BigDecimal object internals:OFFSET SIZE TYPE DESCRIPTION VALUE
0 12 (object header) N/A12 4 int BigDecimal.scale N/A16 8 long BigDecimal.intCompact N/A24 4 int BigDecimal.precision N/A28 4 BigInteger BigDecimal.intVal N/A32 4 String BigDecimal.stringCache N/A36 4 (loss due to the next object alignment)
Instance size: 40 bytes (estimated, the sample instance is not available)Space losses: 0 bytes internal + 4 bytes external = 4 bytes total
Data Locality
56 + 520 = 576 bytes 64 Bytes
Memory LayoutTransactionPlain[100]
TransactionCompact[100]
276 kB
• Keep data compact
• Think about data types
• Keep memory allocations low
• Keep garbage collection rate low
And...
Which patternsshould Iuse?
Tree Model
Flattened Data Model
Streaming Data Access
!𝑧𝑦𝑥
4
%𝑦𝑥
1%𝑥𝑥
3
2
Decoupled Algorithms
• Prefer primitives over classes
• Prefer arrays over object trees
• Process one array at a time
What aboutclusterreplication?
MVCC(Multi Version Concurrency Control)
V1
V2
V3
V4
View
1Vi
ew 1
+2Vi
ew 1
+2+3
View
1+2
+3+4
Compact Data
V1
V2
V3
V4
View
1Vi
ew 1
+2Vi
ew 1
+2+3
View
1+2
+3+4
Off heap RAMOn disk
From Disk to RAM
V1
V2
V3
V1 V2 V3
MappedByteBuffer
Data Distribution
Apache Kafka
Data Distribution
Data Distribution
Summary
• Large speedups through cache optimized
algorithms
• Memory layout is crucial
• Scale by cache replication
@marcusgruendler
Thankyou!