CS137: Electronic Design Automation Day 17: November 11, 2005 Placement (Simulated Annealing…) 1 CALTECH CS137 Fall2005 -- DeHon Today • Placement • Improving Quality – Avoiding local minima • Techniques: – Simulated Annealing – Exhaustive (Branch-and-bound) 2 CALTECH CS137 Fall2005 -- DeHon Simulated Annealing • Physically motivated approach • Physical world has similar problems – – – – – objects/atoms seeking minimum cost arrangement at high temperature (energy) can move around at low temperature, no free energy to move cool quicklyfreeze in defects (weak structure) cool slowly allow to find minimum cost 3 CALTECH CS137 Fall2005 -- DeHon Key Benefit • Avoid Local Minima – Allowed to take locally non-improving moves in order to avoid being stuck 4 CALTECH CS137 Fall2005 -- DeHon Simulated Annealing • At high temperature can move around – not trapped to only make “improving” moves – free energy from temperature allows exploration of non-minimum states – avoid being trapped in local minima • As temperature lowers – less energy to take big, non-minimizing moves – more local / greedy moves 5 CALTECH CS137 Fall2005 -- DeHon Design Optimization Components: 1. “Energy” (Cost) function to minimize – represent entire state, drives system forward 2. Moves – local rearrangement/transformation of solution 3. Cooling schedule – – – initial temperature temperature steps (sequence) time at each temperature 6 CALTECH CS137 Fall2005 -- DeHon Basic Algorithm Sketch • Pick an initial solution • Set temperature (T) to initial value • while (T>Tmin) – for time at T • • • • pick a move at random compute Dcost if less than zero, accept else if (RND<e-Dcost/T), accept – update T 7 CALTECH CS137 Fall2005 -- DeHon Details • Initial Temperature – T0=Davg/ln(Paccept) • Cooling schedule – fixed ratio: T=lT • (e.g. l=0.85) – temperature dependent – function of both temperature and acceptance rate • example to come • Time at each temperature – fixed number of moves? – fixed number of rejected moves? – fixed fraction of rejected moves? CALTECH CS137 Fall2005 -- DeHon 8 Cost Function • Can be very general – Combine area, timing, energy, routability… • Should drive entire solution in right direction – reward each good move • Should be cheap to compute delta costs – e.g. FM – Ideally O(1) 9 CALTECH CS137 Fall2005 -- DeHon Example Cost Functions • Total Wire Length bbx bby – Linear, quadratic… • Bounding Box (semi-perimeter) – Surrogate for routed net length • Channel widths – probably wants to be more than just width • Cut width 10 CALTECH CS137 Fall2005 -- DeHon Bad Cost Functions • Update cost – rerun maze route on every move – rerun timing analysis – recalculate critical path delay • Drive toward solution: – size < threshold ? – Critical path delay 11 CALTECH CS137 Fall2005 -- DeHon VPR Wire Costs • VPR Bounding Box Cost i 1 qi bbx i bby i Nets Swartz, Betz, & Rose FPGA 1998 Original table: Cheng ICCAD 1994 12 CALTECH CS137 Fall2005 -- DeHon VPR Timing Costs • • • • Criticality(e)=1-Slack(e)/Dmax TCost(e)=Delay(e)*Criticality(e)CriticalityExp Keep all edge delays in a table Recompute Net Criticality at each Temperature Marquardt, Betz, & Rose FPGA2000 CALTECH CS137 Fall2005 -- DeHon 13 VPR Balance Wire and Time Cost DTCost DWCost DCost l 1 l OldTCost OldWCost Marquardt, Betz, & Rose FPGA2000 CALTECH CS137 Fall2005 -- DeHon 14 Initial Solution • Spectral Placement • Random • Constructive Placement – Fast placers start at lower temperature; assume constructive got global right. 15 CALTECH CS137 Fall2005 -- DeHon Moves • Swap two cells – Within some distance limit? (ex. to come) • swap regions – …rows, columns, subtrees • rotate cell (when feasible) • flip (mirror) cell • permute cell inputs (equivalent inputs) 16 CALTECH CS137 Fall2005 -- DeHon Variant • Allow non-legal solutions – capture badness in cost function – E.g. -- allow cells to overlap • Just make sure cost function makes very expensive as cool – settle out to legal solutions 17 CALTECH CS137 Fall2005 -- DeHon Variant: “Rejectionless” • Order moves by cost – compare FM • Pick random number first • Use random to define range of move costs will currently accept • Pick randomly within this range • Idea: never pick a costly move which will be rejected 18 CALTECH CS137 Fall2005 -- DeHon Theory • If stay long enough at each cooling stage – will achieve tight error bound • If cool long enough – will find optimum 19 CALTECH CS137 Fall2005 -- DeHon Practice • Good results – ultimately, what most commercial tools use...what vpr uses… • Slow convergence • Tricky to pick schedules to accelerate convergence 20 CALTECH CS137 Fall2005 -- DeHon Range Limit • Want to tune so accepting 44% of the moves – Lam and Delosme DAC 1988 • VPR – Define Rlimit – defines maximum Dx and Dy accepted – Tune Rlimit to maintain acceptance rate – Rlimitnew=Rlimitold×(1-0.44+a) a is measured acceptance rate 21 CALTECH CS137 Fall2005 -- DeHon VPR Cooling Schedule • Moves at Temperature = cN4/3 • Temperature Update – Tnew=Told×g – Idea: advance slowly in good a range Betz, Rose, & Marquardt Kluwer 1999 22 CALTECH CS137 Fall2005 -- DeHon Range Limiting? • Eguro alternate [DAC 2005] – define P=D-M – Tune M to control a 23 CALTECH CS137 Fall2005 -- DeHon Range Limiting CALTECH CS137 Fall2005 -- DeHon Eurgo, Hauck, & Sharma DAC 2005 24 Big Hammer • Costly, but general • Works for most all problems – (part, placement, route, retime, schedule…) • Can have hybrid/mixed cost functions – as long as weight to single potential – (e.g. wire/time from VPR) • With care, can attack multiple levels – place and route • Ignores structure of problem – resignation to finding/understanding structure 25 CALTECH CS137 Fall2005 -- DeHon Optimal/Exhaustive 26 CALTECH CS137 Fall2005 -- DeHon • If you run simulated annealing long enough…. – It should converge to optimum • If you have enough monkeys typing at keyboards keyboards long enough – They’ll eventually produce the works of Shakespeare…. 27 CALTECH CS137 Fall2005 -- DeHon Brute Force? • If you are really going to give up on structure and explore the entire space – …there are more efficient ways to do it • …and maybe they’re not terrible – For small/modest problems – As our computers get faster 28 CALTECH CS137 Fall2005 -- DeHon Optimum Placement • Simplest case: – Gate/array (FPGA) w/ fixed cell locations • N locations • M cells • Try all permutations of N choose M N!/M! cases 29 CALTECH CS137 Fall2005 -- DeHon Improving • Prune off symmetry cases – Rotate 90, 180, 270 – Mirror X, Y, XY • Reject provably bad starts – (return to in a minute) 30 CALTECH CS137 Fall2005 -- DeHon Exhaustive Placement • More general: – Modules have variable size – Modules can be rotated/flipped… • To explore all cases: – For each module • For each orientation 31 CALTECH CS137 Fall2005 -- DeHon Branch-and-Bound • Consider dense 1D placement – Search space of all placements – Tree branch is choice of logical cell for physical cell position – Keep track of best solution so far as reach leaves – If partial solution worse than best solution, prune branch 32 CALTECH CS137 Fall2005 -- DeHon Pruning: 1D example • Reducing channel width – Have solution with width=10 – When find a partial solution with width>=10 • Can abort that branch • Reducing Delay – Have solution with delay=20 – When find a partial solution with delay>=20 • Can abort branch 33 CALTECH CS137 Fall2005 -- DeHon Viable • Only on small problems – But “small” growing with machine speed • Use for end-case in constructive – Flatten bottom of hierarchy – Maybe even in iteration/overlap relaxation 34 CALTECH CS137 Fall2005 -- DeHon Caldwell et. al. Results 35 CALTECH CS137 Fall2005 -- DeHon [TRCAD v19n11p1304-13] Runtime 36 CALTECH CS137 Fall2005 -- DeHon [TRCAD v19n11p1304-13] Runtime 37 CALTECH CS137 Fall2005 -- DeHon [TRCAD v19n11p1304-13] Faster? • Accounting and gain complexity – Make it linear time – But does make each update somewhat complex – Exhaustive case less bookkeeping • Not an atypical result… 38 CALTECH CS137 Fall2005 -- DeHon Summary • Simulated Annealing – use randomness to explore space – accept “bad” moves to avoid local minima – decrease tolerance over time • General purpose solution – costly in runtime • Small (sub)problems – May solve exhaustively – Can prune to accelerate… CALTECH CS137 Fall2005 -- DeHon 39 Admin • Class Monday • …but not W, F (finish assignment) 40 CALTECH CS137 Fall2005 -- DeHon Big Ideas: • Use randomness to explore large (nonconvex) space – Simulated Annealing • Use dominance to quickly skip over obviously bad solutions – Branch and Bound 41 CALTECH CS137 Fall2005 -- DeHon