Presentation is loading. Please wait.

Presentation is loading. Please wait.

Device and Architecture Co-Optimization for FPGA Power Reduction Lerong Cheng, Phoebe Wong, Fei Li, Yan Lin, and Prof. Lei He EE Department, UCLA Partially.

Similar presentations


Presentation on theme: "Device and Architecture Co-Optimization for FPGA Power Reduction Lerong Cheng, Phoebe Wong, Fei Li, Yan Lin, and Prof. Lei He EE Department, UCLA Partially."— Presentation transcript:

1 Device and Architecture Co-Optimization for FPGA Power Reduction Lerong Cheng, Phoebe Wong, Fei Li, Yan Lin, and Prof. Lei He EE Department, UCLA Partially supported by NSF CAREER award CCR / and NSF grant CCR Address comments to

2 Outline Background and motivation Trace-based power and delay estimation Device and architecture co-optimization Conclusion

3 Evaluation of Conventional FPGA Architecture LUT size and cluster size have been evaluated for conventional FPGA  performance and area [Ahmed et al, ISFPGA’00]  power and performance [Li et al, ISFPGA ‘03]  Architecture tuning leads to 2.8X energy difference and 1.5X delay difference Logic block I/O pad Switch box Connection box Critical Path Delay (ns) Total FPGA Energy (nJ/cycle) (8, 7) (6, 7) (6, 6) (10, 5) (8, 5) (12, 4) (6, 5) (8, 4) (6, 4) (10, 4) (8, 6) (12, 5) (10, 6) (12, 6) (10, 7) (12, 7) (10, 3) (12, 3) (8, 3) (6, 3) Island style FPGA architecture Evaluation result

4 Evaluation of Low-Power FPGA Architecture Field programmable dual-vdd for power reduction [Lin et al, ISFPGA’05]  Applying field programmable dual Vdd reduces energy-delay product by 49% High Vdd Logic block Low Vdd logic block Vdd programmable logic block Conventional FPGA Vdd programmable FPGA

5 Evaluation Methodology Parasitic Extraction Cycle-accurate Power Simulator (Psim) Power Arch Spec Logic Optimization(SIS) Tech-Mapping (RASP) Timing-Driven Packing (TV-Pack) Placement & Routing (VPR) Delay Area Benchmark circuits

6 Impact of Device Tuning All the previous work only considers architecture tuning Device tuning leads to 84X power difference and 12X delay difference It is necessary to perform device tuning and architecture tuning simultaneously

7 Challenge of Device and Architecture Co-Optimization We consider the following architecture and device parameters during our co-optimization:  Architecture parameters: Cluster size (N) LUT size (K)  Device parameters: Supply voltage (Vdd) Threshold voltage (Vt) Hyper-architecture (hyper-arch) is the combination of the device and architecture parameters. Large number of hyper-arch combinations VPR and Psim are too slow to deal with such large number of experiments Need fast yet accurate power and delay estimation

8 Outline Back ground and motivation Trace-based power and delay estimation  Trace collection  Trace based power and delay model  Accuracy and efficiency verification of Trace-based estimator Device and architecture co-optimization Conclusion

9 Trace Collection VPR and Psim Ptrace Short circuit power ratio Circuit element statistics Switching activity Critical path structure Assume trace information will remain the same when device setting changes Area Trace

10 Trace Base Estimation (Ptrace) Framework Trace Ptrace Chip level delay, power, and area Circuit level delay and power Device independent Device dependent

11 Outline Back ground and motivation Trace-based power and delay estimation  Trace collection  Trace based power and delay model  Accuracy and efficiency verification of Trace-based estimator Device and architecture co-optimization Conclusion

12 Delay Model in VPR Delay is calculated for each path as  N i p is number of type i elements in the path and D i is delay of type i element  Delay of the logic elements is measured by SPICE simulation  Elmore delay is used for interconnect wire segments Critical path is the path with longest delay

13 Delay in Ptrace Obtain the path structure of a set of longest circuit paths Assume that when device setting changes, the new critical path is still among the set of longest paths. Delay computation: Trace information Device dependent parameters

14 Dynamic Power Model Psim  Switch power Switching activity is measured by timing simulation for each node S i is the average switching activity  Short circuit power α sc is calculated for each node Ptrace  Switch power  Short circuit power α sc is the average short circuit power ratio for the whole circuit Trace information Device dependent parameters

15 Static Power Model Psim  Without power gating  With power gating Ptrace  Without power gating  With power gating Trace information Device dependent parameters

16 Outline Back ground and motivation Trace-based power and delay estimation  Trace collection  Trace based power and delay model  Accuracy and efficiency verification of Trace-based estimator Device and architecture co-optimization Conclusion

17 Experiment Setting Collect trace using ITRS 70nm technology, but apply to both 100nm and 70nm technologies 20 MCNC benchmarks Assume each benchmark works in its highest possible frequency Power and delay are computed as geometric mean of 20 benchmarks. Evaluation range VddVtLUT size (K)Cluster size (N) 0.8~1.10.2~0.43~76~12

18 Accuracy Average power error is 3.4%. Average delay error is 6.4%.  Delay error is due to Ptrace ignores the impact of path branches that considered in VPR

19 Runtime VPR and Psim for one device setting  five days on eight 1.2GHz Intel Xeon servers Ptrace for 20 device settings  80 seconds on one 1.2GHz Intel Xeon server

20 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization  Energy and delay tradeoff  ED and area tradeoff  Comparison between classes  Comparison between device tuning and architecture tuning Conclusion

21 Architectures Classes to be Evaluated Hyper-architecture classes Baseline case  Vdd suggested by ITRS  Architecture same as Xilinx Virtex-II™.  Vt optimized by our method with respect to the above architecture and Vdd Hyper-arch classes Vt Homo-VtHomogeneous Vt Hetero-VtHeterogeneous Vt Homo-Vt+GHomogeneous Vt + Power Gating Hetero-Vt+GHeterogeneous Vt + Power Gating VddVtLUT size (K)Cluster size (N)

22 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization  Energy and delay tradeoff  ED and area tradeoff  Comparison between classes  Comparison between device tuning and architecture tuning Conclusion

23 Energy and Delay Tradeoff Dominant hyper-arch  Hyper-arch B is inferior to A if A has less energy and smaller delay than B.  Dominant hyper-archs (dom-arch) are the hyper-archs that are NOT inferior to any other hyper-archs.

24 Energy and Delay Tradeoff Hetero-Vt can reduce power Power gating reduces more leakage power than hetero-Vt Hetero-Vt has less impact when power gating is applied

25 Min-ED Hyper-Arch Hyper-arch classes Vdd (V) CVt (V)IVt (V)(N, K)ED (nJ·ns) ED reduction % Baseline (8,4) Homo-Vt (6,7) Hetero-Vt (8,4) Homo-Vt+G (12,4) Hetero-Vt+G (8,4) To achieve the best energy and delay tradeoff, we find out the hyper-arch with the minimum energy and delay product (ED)  Compared to the baseline, the min-ED hyper-arch of the conventional FPGA (Homo-Vt) reduces ED by 13.4%  For the Hetero-Vt class, ED is reduced by 20.5%  If power gating is applied, ED can be reduced by up to 59.0%

26 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization  Energy and delay tradeoff  ED and area tradeoff  Comparison between classes  Comparison between device tuning and architecture tuning Conclusion

27 ED and area Tradeoff Architecture tuning has great impact on area. To achieve the best area and ED tradeoff, we find the hyper-arch with the minimum product of area, energy and delay (AED)

28 ED Area Tradeoff for Classes without Power Gating Compared to the min-ED hyper arch, the min-AED hyper-arch significantly reduce area with a small ED increase

29 Sleep Transistor Size Tuning When Power gating is applied, sleep transistors may increase area The larger the sleep transistor size, the smaller the delay Sleep transistor size tuning:  Area overhead introduced by sleep transistors of logic blocks is negligible.  We consider 2X, 4X, 7X and 10X PMOS as sleep transistor for switch buffer

30 ED Area Tradeoff for Classes with Power Gating The area reduction achieved by device and architecture co-optimization compensates the area overhead introduced by sleep transistors

31 Min-AED Hyper-Arch Vdd (V) CVt (V) IVt (V) (N,K) Sleep transistor size ED (nJ·ns) Normalized area AED reduction % Baseline (8,4) Homo-Vt (6,4) Hetero-Vt (12, 4) Hetero-Vt+G (12, 4) Hetero-Vt+G (12, 4) Compared to the baseline, the min-AED hyper-arch in the conventional FPGA class can reduce area by 20% and ED by 12.3% In the Hetero-Vt class, ED is reduced by 20.8% and area is reduced by 23% compared to the baseline If power gating is applied, ED is reduced by 54.6% and area is reduced by 8.3%

32 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization  Energy and delay tradeoff  ED and area tradeoff  Comparison between classes  Comparison between device tuning and architecture tuning Conclusion

33 Comparison Between Classes in Similar Performance Range Homo-VtHetero-Vt VddVt(N, K)E (nJ)D (ns) ED (nJ·ns) VddCVtIVt(N, K)E (nJ)D (ns)ED (nJ·ns) , , , , , , Homo-Vt+GHetero-Vt+G VddVt(N, K)E (nJ)D (ns) ED (nJ·ns) VddCVtIVt(N, K)E (nJ)D (ns)ED (nJ·ns) , , , , , , Vt for logic block is lower than Vt for interconnect Vt for classes with power gating is lower

34 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization  Energy and delay tradeoff  ED and area tradeoff  Comparison between classes  Comparison between device tuning and architecture tuning Conclusion

35 Dom-Archs under Different Device Settings For a given device setting architecture tuning changes delay and energy in a smaller range Device tuning has a much more impact on delay and energy

36 Outline Back ground and motivation Trace-based power and delay estimation Device and architecture co-optimization Conclusion

37 Conclusion and Discussion Trace-based estimator provides efficient and accurate FPGA power and delay estimation  Average power error is 3.4% and average delay error is 6.1% Device and architecture co-optimization reduces ED by 20.5% and area by 23.3% when there is no power gating With power gating, device and architecture co-optimization reduces ED by 54.6% and area by 8.3% Device tuning has a more significant impact on delay and power than architecture tuning does In recent research, Ptrace has been extended to consider leakage and timing yield with process variations

38


Download ppt "Device and Architecture Co-Optimization for FPGA Power Reduction Lerong Cheng, Phoebe Wong, Fei Li, Yan Lin, and Prof. Lei He EE Department, UCLA Partially."

Similar presentations


Ads by Google