790 likes | 1.03k Views
Clocked Storage Elements for High-Performance and Low-Power Systems The book under the same title is published by J. Wiley Pub. Co. Vojin G. Oklobdzija* June 23th, 2003 Presentation given at: EPFL Lausanne, Switzerland *Advanced Computer System Engineering Laboratory
E N D
Clocked Storage Elements for High-Performance and Low-Power SystemsThe book under the same title is published by J. Wiley Pub. Co. Vojin G. Oklobdzija* June 23th, 2003 Presentation given at: EPFL Lausanne, Switzerland *Advanced Computer System Engineering Laboratory University of California Davis Presentations available at: http://www.ece.ucdavis.edu/acsel
Outline • Why working on Clocked Storage Elements ? • M-S Latch is not a Flip-Flop ! • How do we compare them ? • What are the relevant parameters ? • What is an appropriate setup ? • What do we use in high-performance microprocessors ? • How do they compare ? • What should we do for low-power ? • How do they compare ? • What next ? Ideas, Suggestions, Insights Prof. V.G. Oklobdzija, University of California
Importance Prof. V.G. Oklobdzija, University of California
ISSCC-2002 Clock trends in high-performance systems Prof. V.G. Oklobdzija, University of California
Courtesy: Doug Carmean, Intel Corp, Hot-Chips-13 presentation Prof. V.G. Oklobdzija, University of California
Why working on Clocked Storage Elements ? Example: In a 2.0 GHZ processor T=500pS • Typically clocked storage element D-Q delay is in the order of 100-150pS • If one can design a faster CSE: e.g. 80-100pS D-Q, this represents 10-15% performance improvement • If in addition one can absorb 20pS of clock uncertainties and embedd one level of logic – this can yield up to 20% performance improvement • Try to achieve 10-20% performance improvement by introducing new features in the architecture ! • This is sufficient to turn an architect into a circuit designer ! Prof. V.G. Oklobdzija, University of California
Basic Definitions Prof. V.G. Oklobdzija, University of California
Clock Generation and Distribution Non-idealities • Jitter • Jitter is a temporal variation of the clock signal manifested as uncertainty of consecutive edges of a periodic clock signal. • It is caused by temporal noise events • Manifested as: - cycle-to-cycle or short-term jitter, tJS - long-term jitter, tJL • Characteristic of clock generation system • Skew • Is a time difference between temporally-equivalent or concurrent edges of two periodic signals • Manifests as SE-to-SE fluctuation of clock arrival at the same time instance • Characteristic of clock distribution system • Caused by spatial variations in signal propagation Prof. V.G. Oklobdzija, University of California
Clock Uncertainties Prof. V.G. Oklobdzija, University of California
Difference between Latch and Flip-Flop Prof. V.G. Oklobdzija, University of California
After the transition of the clock data can not change Latch is “transparent” Difference between Latch and Flip-Flop Prof. V.G. Oklobdzija, University of California
Two-Phase Clocking with Two-Phase Double Latch Prof. V.G. Oklobdzija, University of California
Two-Phase Clocking with One-Phase Double Latch Some people refer to this latch arrangement as: “negative edge Flip-Flop” ! Prof. V.G. Oklobdzija, University of California
How can one recognize the difference without knowing what is inside the “black-box” ? Flip-Flop and M-S Latch Arrangement Prof. V.G. Oklobdzija, University of California
F-F and M-S Latch: Difference Experiment: Failed ! Prof. V.G. Oklobdzija, University of California
No Clock Pulse Capturing Latch Flip-Flop M-S Latch F-F and M-S Latch: Difference Structural Difference: S R Prof. V.G. Oklobdzija, University of California
PG Theory of Operation: Sn+1 Prof. V.G. Oklobdzija, University of California
R S Flip-Flop: Example-2 D=0 pulse D=1 SAFF DEC Alpha 21264 (Madden & Bowhill, 1990, Matsui 1994) Prof. V.G. Oklobdzija, University of California
F-F Derivation using Delayed Clock Equivalent to: Prof. V.G. Oklobdzija, University of California
Systematically Derived ET FF N. Nedovic, V. G. Oklobdzija, “Dynamic Flip-Flop with Improved Power”, ICCD 2000, Sept. 2000 Prof. V.G. Oklobdzija, University of California
Flip-Flop: Example (HLFF, H. Partovi) Prof. V.G. Oklobdzija, University of California
Flip-Flop: Example (HLFF, H. Partovi) Prof. V.G. Oklobdzija, University of California
Timing and Power metrics Prof. V.G. Oklobdzija, University of California
Delay • Sum of setup time U and Clk-Q delay is the only true measure of the performance with respect to the system speed • T = TClk-Q + TLogic + Tsetup+ Tskew T TD-Q=TClk-Q + TSetup TClk-Q TLogic TSetup Prof. V.G. Oklobdzija, University of California
Delay vs. Setup/Hold Times Sampling Window Prof. V.G. Oklobdzija, University of California
Timing Characteristics Prof. V.G. Oklobdzija, University of California
Absorbing Clock Uncertainties Prof. V.G. Oklobdzija, University of California
Hybrid Latch Flip-Flop Skew absorption Partovi et al, ISSCC’96 Prof. V.G. Oklobdzija, University of California
Power Consumption • All power related to the SE can be divided into: • Input power • Data power (PD) • Clock power (PCLK) • Internal power (PINT) • Load power (PLOAD) • PLOAD can be merged into PINT • Internal power is a function of • data activity ratio () – number of captured data transitions with respect to number of clock transitions (max=100%) • no activity (0000… and 1111…) • maximum activity (0101010..) • average activity (random sequence) • Glitching activity • Delay is (minimum D-Q) • Clk-Q + setup time Prof. V.G. Oklobdzija, University of California
State Element Performance Metrics It is always possible trade power for speed Common metrics: • Power-Delay Product (PDP) • Misleading measure • Good only if measured at constant frequency = EDP • EDP - Energy-Delay Product (EDP) • More accurate measure • ED2P – Energy-Delay2-Product • A new measure, being justified by new results (Hofstee, Nowka, IBM) Prof. V.G. Oklobdzija, University of California
PDP, EDP Comparison High Voltage Low Voltage Slow Corner Prof. V.G. Oklobdzija, University of California
Design & optimization tradeoffs • Opposite Goals • Minimal Total power consumption • Minimal Delay • Power-Delay tradeoff • Minimize Power-Delay product (PDPtot) @ f=const. Opt. Opt. Opt. Prof. V.G. Oklobdzija, University of California
Clocked Storage Elements:Examples Prof. V.G. Oklobdzija, University of California
Simulation Conditions: • Power Supply Voltage: VDD=1.8V nominal • Temperature T=27°C nominal • Technology: 0.18m Fujitsu • Fan-Out of 4 Delay = 75pS • Transistor Widths • Minimal 0.36m • Maximal 10m • Load: 14 minimal inverters in the technology used • Clock frequency: 500MHz (250MHz for Dual-Egde) • Data/Clock slopes of ideal signal 100ps Prof. V.G. Oklobdzija, University of California
Transmission Gate MS Latch • Two staticized transmission gate transparent latches • Direct path D-Q consists of two transmission gates and two regenerative inverters • Two-phase clock • Advantage: symmetric high-to-low and low-to-high transitions are achievable • Disadvantage: large cost associated with two-phase clock distribution PowerPC 603 (Gerosa, JSSC 12/94) • Comments: • Very low internal power. • Large Total Power due to clock and data load Prof. V.G. Oklobdzija, University of California
C2MOS MS Latch • Forward path consists of two clocked inverters - parts of C2MOS latches • Degradation of speed due to pMOS stacks • Degradation in speed due to non-ideal 2-phase clock • Large clock power (if not buffered locally) Y. Suzuki, “Clocked CMOS Calculator Circuitry”, IEEE J. Solid-State Circuits, Dec. 1973 Prof. V.G. Oklobdzija, University of California
SAFF: Strong Arm 110 • Staticized Sense Amplifier Flip-Flop • Weak nMOS keeps set/reset signals low • Second stage – non-clocked SR latch • Additional NMOS transistor causes slightly increased power consumption and delay degradation • Bad timing characteristics due to the latching stage. Signal propagates through three stages. • Unbalanced rising and falling time of the output signals (speed degraded by 40%) Prof. V.G. Oklobdzija, University of California
Modified SAFF • The first stage is unchanged sense amplifier • Second stage is sized to provide maximum switching speed • Driver transistors are large • Keeper transistors are small and disengaged during transitions Nikolic, Oklobdzija, Stojanovic ISSCC ‘99 V. Stojanovic, US Patent No. 6,232,810 Prof. V.G. Oklobdzija, University of California
Systematicaly Derived SAFF: Example-2 • New pulse-generating stage • Inverters decoupling gates from MN3, MN4 • MN5, MN6 provide leakage current paths • Second stage is unchanged Nikolic, Oklobdzija, ESSCIRC’99 V. Stojanovic, US Patent No. 6,232,810. Prof. V.G. Oklobdzija, University of California
Sense Amplifier-based Flip-Flop (SAbFF) • Emerged as a workaround for SAFF drawbacks • floating nodes (keeping the Sb, Rb nodes low with additional transistors parallel to data-controlled transistors) • symmetric second stage (push-pull realization) • Internal signals still experience transition on every clock cycle V. Stojanovic, US Patent No. 6,232,810. Prof. V.G. Oklobdzija, University of California
Comparison with other SAFFs Nikolic, Oklobdzija, ESSCIRC’99 800 CMOS, nominal corner, Leff = 0.18m, VDD = 1.8V, T = 25C, load on both outputs 700 Falling Egde SAFF Clk-Output Delay [ps] w/NOR 600 500 Rising Egde 400 SAFF w/NAND 300 Rising Egde 200 SAFF Falling Egde SAFF Rising Egde SAFF this work 100 this work 0 0 50 100 150 200 250 Load [fF] Prof. V.G. Oklobdzija, University of California
Conditional Capture Flip-Flop (CCFF) 0.18m Fujitsu; f = 500MHz; VDD = 1.8V; Data activity 50% • Principle of Operation • Suppress any transition in flip-flop if the input to be captured is equal to previous output value • Double-ended realization • FF functionality achieved by producing clock pulse • Static operation by use of keepers • Second stage is pass-transistor latch • Comments • Contention with keepers causes larger first stage • Large power consumption despite conditional signaling B. S. Kong, et all, ISSCC 2000 Prof. V.G. Oklobdzija, University of California
Partovi’s HLFF • Hybrid Latch-Flip-Flop combination • Negative set-up time of -80pS • Robustness to clock skew and fast clocking Our simulations show AMD K-6, Partovi, ISSCC’96 • Gains • speed (negative setup time) • robustness to clock skew • Drawbacks • sensitivity to clock slope • relatively high internal power (due to precharge) Prof. V.G. Oklobdzija, University of California
Semi-Dynamic Flip-Flop • Hybrid combination used in UltraSPARC-III • Very fast circuit ( 173pS Clk-Q delay .18u technology, 1.8V, 27oC ) • Problem D=Q=1: Our simulations shows F. Klass, VLSI Circuits’98 • Negative setup time • Feature of small penalty for embedded logic • Relatively high internal power consumption and clock load Prof. V.G. Oklobdzija, University of California
Transmission Gate Flip-Flop (TGFF) • Two transmission gates define transparency window • Time window with non precharge-evaluate structure • Low input activity => low output activity • Comments: • Two transmission gates increase delay • Noticeable data power Prof. V.G. Oklobdzija, University of California
Comparison Prof. V.G. Oklobdzija, University of California
Overall Results 4 fo4 2 fo4 Prof. V.G. Oklobdzija, University of California
Overall Results Prof. V.G. Oklobdzija, University of California
Overall Results Prof. V.G. Oklobdzija, University of California
Conventional Clk-Q vs. minimum D-Q • Hidden positive setup time • Degradation of Clk-Q Older 0.22u comparison results Prof. V.G. Oklobdzija, University of California