NoC Router Estimation Andrew B Kahng Bill Lin and Siddhartha Nath UCSD CSE and ECE Departments abk billlin sinath engucsdedu Outline Motivation ID: 806261
Download The PPT/PDF document "Explicit Modeling of Control and Data fo..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Explicit Modeling of Control and Data for Improved NoC Router Estimation
Andrew B.
Kahng
+*
, Bill Lin
*
and Siddhartha
Nath
+
UCSD
CSE
+
and ECE
*
Departments
{
abk
,
billlin
,
sinath
}@
eng.ucsd.edu
Slide2Outline
Motivation
Our work: Overview
Methodology
Flit-level power estimationSummary
2
Slide3NoC Modeling So Far… (ORION)
3
Arbiter
XBAR
BUF I
BUFE
BUFW
BUFN
BUFS
Link
Link
Link
Link
SRC
Link
Link
Link
Link
SINK
ORION1.0
(
2002)
6NOR + 2INV + DFF
ORION2.0 (2009)
6NOR + 2INV + DFF
Leakage power
Clock power
Slide4What Is The Problem?
RTL code mismatch
Logic transformation and technology mapping mismatch4
Arbiter
XBAR
BUF I
BUFE
BUFW
BUFN
BUFS
Link
Link
Link
Link
SRC
Link
Link
Link
Link
SINK
6NOR + 2INV + DFF
Slide5How Bad Is It?
Router RTL generators:
Netmaker – Cambridge, UKStanford NoC - Stanford
460%
89%
Why such large errors?
Assumed logic template inaccurate
Control logic not modeled
Implementation details missing
5
Slide6Motivation
Our work: Overview
Methodology
Flit-level power estimation
Summary Outline
6
Slide7P - #Ports
V - #VCs
B - #BUFs
F – Flit-width
Key idea: No assumed logic template
Component models derived from actual RTL synthesized with cell libraries
We Propose: Step 1
Derive router component block parametric models
from
post-synthesis netlists
P
V
B
F
#
Instances102832
330082832
211252832825
~P
2
~P
2
PV
BF
# Instances
528
16
40052
832
8255
28641673
~F
XBAR ~ P
2F7
Slide8We Propose: Step 2
Automatic fitting of models with
post-P&R power and area8
XBAR ~ P
2
F
P
V
B
F
Area
5
28161439.9
528322916.0
528645867.4
828327465.1
LSQR
XBAR
area = a1.P
2F + a0
Key idea: Capture implementation
details using automatic regression fit
Characterization
performed only once and usable for multiple design space explorations
Slide9Motivation
Our work: Overview
Methodology
Flit-level power estimation
Summary Outline
9
Slide10Model Development
Two RTL generators:
Netmaker (Cambridge, UK)Stanford NoCSP&R tools:
Cadence RC & Synopsys DC for hierarchical synthesis to analyze each blockCadence SOC Encounter for P&R
NoC
router RTL generators
Impl
params
: Clock Frequency
µArch
params
: P, V, B, F
Synthesis and P&R: DC/RC, SOCE
Analysis of blocks: XBAR, SW & VC arbiter, Input & Output buffers
New models for each component block
Component
ModelXBARP
2FSWVC9(P2V2 + P2 + PV – P)InBUF
180PV + 2PVBF + 2P2VB + 3PVB + 5P2B + P2 + PF + 15POutBUF25P + 80PV
CLKCTRL0.02(SWVC + InBUF + OutBUF)
10
Slide11Overall Methodology
Manual
Quick and easyMisses implementation details
Basic
Regression fit
Manual
Estimates for gate count
ORION_NEW models
LSQR
Technology
Library
Cell area
Cell leakage
Pin cap.
Internal
e
nergy
Area
Power: leakage, internal, switching
Post P&R
data per block
Std. cell count
& area
Leakage power
Internal power
Switching power
LSQR
Accurate (captures
implementation details)
One-
tim
e overhead
(generation
of
P&R
training data points)
11
Slide12POWER
6.5x
reduction
Results: Area And Power
12
AREA
4x
reduction
Methodology scales across technologies, router RTL generators
Slide13Motivation
Our work: Overview
Methodology
Flit-level power estimation
Summary Outline
13
Slide14Flit-level Power Estimation
Dynamic power estimation using flit-level bit encodings
Have integrated with full-system NoC simulator (GARNET)
Post-P&R router netlist
Testbench
Gate-level simulation
VCD
Power analysis
Power Report
Regression fit
ORION_NEW models
Flit-level power model
GARNET
gem5
Flit-level power estimates
14
Slide15Results: Flit-level Power
Accurate estimation of flit-level dynamic power
15
3.6x
reduction
Slide16Motivation
Our work: Overview
Methodology
Flit-level power estimation
Summary Outline
16
Slide17Summary
New hybrid modeling methodology:
relax the template mindsetExplicitly models control and data signalsCaptures RTL and implementation details
Using proposed parametric regression methodology, worst-case estimation errors reduced by a factor of
6.5x from ORION2.0 for power4x from ORION2.0 for areaWe propose an application of our methodology for flit-level dynamic power modeling and integration with GARNET3.6x worst-case error reduction in dynamic power estimation
Ongoing: Non-parametric modeling of post-P&R power and area
17
Slide18Thank You !
18
Slide19Back
up
19
Slide20Regression analysis approach
Multi-step regression fit
Step 1: Fit instances of each router component with post-layout instance counts
a1
. Instsmodel <component> + a0 = Inststool
<component>
Step 2a: Fit area of each router component with post-layout area
b
1
.
Insts
R
model <component> + b
0 = Areatool <component>InstsRmodel
<component> = a1. Instsmodel <component> + a0
Step 2b: Fit power of each router component with post-layout power (leakage, internal, switching separately)
{c
5, d5, e5}. InstsRmodel
XBAR + {c4, d4, e4}.InstsR model SWVC + {c3
, d3, e3}.InstsRmodel InBUF + {c2, d2, e2}.Insts
Rmodel OutBUF + {c1, d1, e1}.InstsRmodel CLKCTRL + {c
0, d0, e0} = {Pleak tool,P
int tool, PSW tool}
20
Slide21Related work
Architecture templates
ORION2.0Gate-level analytical modelsParametric regressionPre- and post-layout power estimationRTL simulations
Non-parametric regressionMARS
NoC
Modeling
Regression model
Parametric
Non-parametric
ORION_NEW + regression; flit-level
Circuit model
Arch templates
Analytical
Significant
Departure:
Relax
the “template” mindset
Control
Tool
21
Slide22Results
Avg. estimation error in
# instances
reduced from
109.5%
to
8.8%
Avg. estimation error in
area
reduced to
9.8%
Avg
estimation error in
power reduced to 4.58%22