S ervice L equential

advertisement
Service Aggregated Linked Sequential Activities
SALSA Team
Geoffrey Fox
Xiaohong Qiu
Seung-Hee Bae
Huapeng Yuan
Indiana University
Technology Collaboration
George Chrysanthakopoulos
Henrik Frystyk Nielsen
Microsoft
Application Collaboration
Cheminformatics
Rajarshi Guha
David Wild
Bioinformatics
Haiku Tang
Demographics (GIS)
Neil Devadasan
IU Bloomington and IUPUI
GOALS: Increasing number of cores
accompanied by continued data deluge
Develop scalable parallel data mining
algorithms with good multicore and
cluster performance; understand
software runtime and parallelization
method. Use managed code (C#) and
package algorithms as services to
encourage broad use assuming
experts parallelize core algorithms.
CURRENT RESUTS: Microsoft CCR supports MPI,
dynamic threading and via DSS a Service model of
computing; detailed performance measurements
Speedups of 7.5 or above on 8-core systems for
“large problems” with deterministic annealed (avoid
local minima) algorithms for clustering, Gaussian
Mixtures, GTM (dimensional reduction) etc.
SALSA
N data points E(x) in D dim. space and Minimize F by EM
N
N
x 1
x 1
2 2
F  T
) ln{
g
(
k
)
exp[

0.5(
E
(
x
)

Y
(
k
))/ T/](Ts(k ))]
F
 aT( x
p(
x) ln{
exp[

(
E
(
x
)

Y
(
k
))
k 1  k 1
K
K
Deterministic Annealing Clustering (DAC)
• a(x) = 1/N or generally p(x) with  p(x) =1
• g(k)=1 and s(k)=0.5
• T is annealing temperature varied down from 
with final value of 1
• Vary cluster centerY(k)
• K starts at 1 and is incremented by algorithm
• My 4th most cited article but little used; probably
as no good software compared to simple K-means
SALSA
Deterministic Annealing Clustering of Indiana Census Data
Decrease temperature (distance scale) to discover more clusters
Distance Scale
Temperature0.5
N data points E(x) in D dim. space and Minimize F by EM
N
F  T  a ( x) ln{ k 1 g (k ) exp[ 0.5( E ( x)  Y (k )) 2 / (Ts(k ))]
K
x 1
Deterministic
Generative
Traditional
Topographic
Annealing
Gaussian
Clustering
Mapping
(GTM)
(DAC)
Deterministic
Annealing
Gaussian
mixture
models
GM
models
(DAGM)
• a(x) = 1/NMixture
or generally
p(x)
D/2 with  p(x) =1
• a(x) = 1 and g(k) = (1/K)(/2)
•and
Ass(k)=0.5
DAGM but set T=1 and fix K
•• g(k)=1
a(x)
=
1
• s(k) = 1/  and T = 1
• T is annealing
temperature
2)D/2}1/T
varied down from 
M W/(2(k)
•Y(k) •= g(k)={P
m=1DAGTM:

(X(k))
km m
Deterministic
Annealed
with
final
value
of
1
2
2/2 Gaussian)
• s(k)=
(k)
(taking
case
of(X-
spherical
• Choose
fixed

(X)
=
exp(
0.5
)
)
m
m
Generative
Topographic
Mapping
• Vary
cluster centerY(k)
but can
calculate
weight
T misand
annealing
temperature
varied
down
from

• Vary•W

but
fix
values
of
M
and
K
a
priori
2
• GTM
has several
natural
annealing
P
and
correlation
matrix
s(k)
=
(k)
(even
for space
k
with
final
value
of
1
•Y(k) E(x)versions
Wm are2 vectors
in
original
high
D
dimension
based
on eitherformulae
DAC orfor
DAGM:
matrix
(k)
)
using
IDENTICAL
•
Vary
Y(k)
P
and
(k)
• X(k) andunder
m areinvestigation
vectors
in 2 dimensional mapped space
k
Gaussian
• K startsmixtures
at 1 and is incremented by algorithm
•K starts at 1 and is incremented by algorithm
SALSA

We implement micro-parallelism using Microsoft CCR
(Concurrency and Coordination Runtime) as it supports both MPI rendezvous
and dynamic (spawned) threading style of parallelism
http://msdn.microsoft.com/robotics/

CCR Supports exchange of messages between threads using named ports
and has primitives like:
 FromHandler: Spawn threads without reading ports
 Receive: Each handler reads one item from a single port
 MultipleItemReceive: Each handler reads a prescribed number of items of
a given type from a given port. Note items in a port can be general
structures but all must have same type.
 MultiplePortReceive: Each handler reads a one item of a given type from
multiple ports.

CCR has fewer primitives than MPI but can implement MPI collectives
efficiently

Use DSS (Decentralized System Services) built in terms of CCR for service
model

DSS has ~35 µs and CCR a few µs overhead
SALSA
MPI Exchange Latency in µs (20-30 µs computation between messaging)
Machine
Intel8c:gf12
(8 core
2.33 Ghz)
(in 2 chips)
Intel8c:gf20
(8 core
2.33 Ghz)
Intel8b
(8 core
2.66 Ghz)
AMD4
(4 core
2.19 Ghz)
Intel(4 core)
OS
Runtime
Grains
Parallelism
MPI Latency
Redhat
MPJE(Java)
Process
8
181
MPICH2 (C)
Process
8
40.0
MPICH2:Fast
Process
8
39.3
Nemesis
Process
8
4.21
MPJE
Process
8
157
mpiJava
Process
8
111
MPICH2
Process
8
64.2
Vista
MPJE
Process
8
170
Fedora
MPJE
Process
8
142
Fedora
mpiJava
Process
8
100
Vista
CCR (C#)
Thread
8
20.2
XP
MPJE
Process
4
185
Redhat
MPJE
Process
4
152
mpiJava
Process
4
99.4
MPICH2
Process
4
39.3
XP
CCR
Thread
4
16.3
XP
CCR
Thread
4
25.8
Fedora
Messaging CCR versus MPI
C# v. C v. Java
SALSA
Intel8b: 8 Core
(μs)
1
2
3
4
7
8
1.58
2.44
3
2.94
4.5
5.06
Shift
2.42
3.2
3.38
5.26
5.14
Two Shifts
4.94
5.9
6.84
14.32
19.44
3.96
4.52
5.78
6.82
7.18
Shift
4.46
6.42
5.86
10.86
11.74
Exchange As Two
Shifts
7.4
11.64
14.16
31.86
35.62
6.94
11.22
13.3
18.78
20.16
Pipeline
Dynamic
Spawned
Threads
Pipeline
Rendezvous
MPI style
Number of Parallel Computations
CCR Custom
Exchange
2.48
SALSA
30
Time Microseconds
AMD Exch
25
AMD Exch as 2 Shifts
AMD Shift
20
15
10
5
Stages (millions)
0
0
2
4
6
8
10
Overhead (latency) of AMD4 PC with 4 execution threads on MPI style Rendezvous
Messaging for Shift and Exchange implemented either as two shifts or as custom CCR
pattern
70
Time Microseconds
60
Intel Exch
50
Intel Exch as 2 Shifts
Intel Shift
40
30
20
10
Stages (millions)
0
0
2
4
6
8
10
Overhead (latency) of Intel8b PC with 8 execution threads on MPI style Rendezvous
Messaging for Shift and Exchange implemented either as two shifts or as custom
CCR pattern
1.6
Scaled
Intel 8b Vista C# CCR 1 Cluster
1.5
10,000
Runtime
1.4
500,000
1.3
Divide runtime
by
Grain Size n
. # Clusters K
1.2
50,000
Datapoints
per thread
1.1
1
a)
1
2
3
4
5
6
Number of Threads (one per core)
7
8
1
Scaled
Runtime
Intel 8b Vista C# CCR 80 Clusters
50,000
10,000
0.95
500,000
0.9
Datapoints
per thread
0.85
0.8
b)
1
2
3
4
5
8 cores (threads)
and 1 cluster
show memory
bandwidth effect
6
Number of Threads (one per core)
7
8
80 clusters show
cache/memory
bandwidth effect
Multicore Matrix Multiplication
(dominant linear algebra in GTM)
Speedup = Number of cores/(1+f)
f = (Sum of Overheads)/(Computation per core)
10,000.00
Execution Time
Seconds 4096X4096 matrices
Computation  Grain Size n . # Clusters K
Overheads are
Synchronization: small with CCR
Load Balance: good
Memory Bandwidth Limit:  0 as K  
Cache Use/Interference: Important
Runtime Fluctuations: Dominant large n, K
All our “real” problems have f ≤ 0.05 and
speedups on 8 core systems greater than 7.6
1 Core
1,000.00
Parallel Overhead
 1%
8 Cores
100.00
Block Size
10.00
1
0.14
10
100
1000
10000
Parallel GTM Performance
0.12
Fractional
Overhead
f
0.1
0.08
0.06
4096 Interpolating Clusters
0.04
0.02
1/(Grain Size n)
0
0
0.002
n = 500
0.004
0.006
0.008
0.01
100
0.012
0.014
0.016
0.018
0.02
50 SALSA
0.1
Std Dev Intel 8a XP C# CCR
Runtime 80 Clusters
0.075
500,000
10,000
0.05
50,000
0.025
Datapoints
per thread
0
b)
0
1
2
3
4
5
6
7
Number of Threads (one per core)
8
synchronization
0.006
Std Dev Intel 8c Redhat C Locks
Runtime 80 Clusters
10,000
0.004
50,000
500,000
0.002
Datapoints
per thread
0
b)
1
2
3
4
5
6
Number of Threads (one per core)
This is
average of
standard
deviation of
run time of
the 8 threads
between
messaging
7
8
points





Early implementations of our clustering algorithm
showed large fluctuations due to the cache line
interference effect (false sharing)
We have one thread on each core each calculating a sum
of same complexity storing result in a common array A
with different cores using different array locations
Thread i stores sum in A(i) is separation 1 – no memory
access interference but cache line interference
Thread i stores sum in A(X*i) is separation X
Serious degradation if X < 8 (64 bytes) with Windows


Note A is a double (8 bytes)
Less interference effect with Linux – especially Red Hat
Machine
OS
Run
Time
Intel8b
Intel8b
Intel8b
Intel8b
Intel8a
Intel8a
Intel8a
Intel8c
AMD4
AMD4
AMD4
AMD4
AMD4
AMD4
Vista
Vista
Vista
Fedora
XP CCR
XP Locks
XP
Red Hat
WinSrvr
WinSrvr
WinSrvr
XP
XP
XP
C# CCR
C# Locks
C
C
C#
C#
C
C
C# CCR
C# Locks
C
C# CCR
C# Locks
C



Time µs versus Thread Array Separation (unit is 8 bytes)
1
4
8
1024
Mean Std/
Mean
Std/
Mean Std/
Mean Std/
Mean
Mean
Mean
Mean
8.03
.029
3.04
.059
0.884 .0051
0.884 .0069
13.0
.0095 3.08
.0028
0.883 .0043
0.883 .0036
13.4
.0047 1.69
.0026
0.66
.029
0.659 .0057
1.50
.01
0.69
.21
0.307 .0045
0.307 .016
10.6
.033
4.16
.041
1.27
.051
1.43
.049
16.6
.016
4.31
.0067
1.27
.066
1.27
.054
16.9
.0016 2.27
.0042
0.946 .056
0.946 .058
0.441 .0035 0.423
.0031
0.423 .0030
0.423 .032
8.58
.0080 2.62
.081
0.839 .0031
0.838 .0031
8.72
.0036 2.42
0.01
0.836 .0016
0.836 .0013
5.65
.020
2.69
.0060
1.05
.0013
1.05
.0014
8.05
0.010
2.84
0.077
0.84
0.040
0.840 0.022
8.21
0.006
2.57
0.016
0.84
0.007
0.84
0.007
6.10
0.026
2.95
0.017
1.05
0.019
1.05
0.017
Note measurements at a separation X of 8 and X=1024 (and values between 8 and
1024 not shown) are essentially identical
Measurements at 7 (not shown) are higher than that at 8 (except for Red Hat which
shows essentially no enhancement at X<8)
As effects due to co-location of thread variables in a 64 byte cache line, align the
array with cache boundaries
Parallel Generative Topographic Mapping GTM
Reduce dimensionality preserving
topology and perhaps distances
Here project to 2D
GTM Projection of PubChem:
10,926,94 compounds in 166
dimension binary property space takes
4 days on 8 cores. 64X64 mesh of GTM
clusters interpolates PubChem. Could
usefully use 1024 cores! David Wild will
use for GIS style 2D browsing interface
to chemistry
PCA
GTM
Linear PCA v. nonlinear GTM on 6 Gaussians in 3D
PCA is Principal Component Analysis
GTM Projection of 2 clusters
of 335 compounds in 155
SALSA
dimensions
“Main Thread” and Memory M
MPI/CCR/DSS
From other nodes
MPI/CCR/DSS
From other nodes
0
m0
1
m1
2
m2
3
m3
4
m4
5
m5
6
m6
7
m7
Subsidiary threads t with memory mt

Use Data Decomposition as in classic distributed memory
but use shared memory for read variables. Each thread
uses a “local” array for written variables to get good cache
performance

Multicore and Cluster use same parallel algorithms but
different runtime implementations; algorithms are
 Accumulate matrix and vector elements in each process/thread
 At iteration barrier, combine contributions (MPI_Reduce)
 Linear Algebra (multiplication, equation solving, SVD)
SALSA


Micro-parallelism uses low latency CCR threads or
MPI processes
Services can be used where loose coupling natural
Input data
 Algorithms

 PCA
 DAC GTM GM DAGM DAGTM – both for complete algorithm
and for each iteration
 Linear Algebra used inside or outside above
 Metric embedding MDS, Bourgain, Quadratic Programming ….
 HMM, SVM ….

User interface: GIS (Web map Service) or equivalent
SALSA
Average run time (microseconds)
350
DSS Service Measurements
300
250
200
150
100
50
0
1
10
100
1000
10000
Round trips

Measurements of Axis 2 shows about 500 microseconds – DSS is 10 times better
20

This class of data mining does/will parallelize well on current/future multicore
nodes

Several engineering issues for use in large applications
 How to take CCR in multicore node to cluster (MPI or cross-cluster CCR?)
 Need high performance linear algebra for C# (PLASMA from UTenn)
 Access linear algebra services in a different language?
 Need equivalent of Intel C Math Libraries for C# (vector arithmetic – level 1
BLAS)
 Service model to integrate modules

Need access to a ~ 128 node Windows cluster

Future work is more applications; refine current algorithms such as DAGTM

New parallel algorithms
 Clustering with pairwise distances but no vectorspaces
 Bourgain Random Projection for metric embedding
 MDS Dimensional Scaling with EM-like SMACOF and deterministicannealing
 Support use of Newton’s Method (Marquardt’s method) as EM alternative
 Later HMM and SVM
SALSA
Download