Presentation is loading. Please wait.

Presentation is loading. Please wait.

11 Online Computing and Predicting Architectural Vulnerability Factor of Microprocessor Structures Songjun Pan Yu Hu Xiaowei Li {pansongjun, huyu,

Similar presentations


Presentation on theme: "11 Online Computing and Predicting Architectural Vulnerability Factor of Microprocessor Structures Songjun Pan Yu Hu Xiaowei Li {pansongjun, huyu,"— Presentation transcript:

1 11 Online Computing and Predicting Architectural Vulnerability Factor of Microprocessor Structures Songjun Pan Yu Hu Xiaowei Li {pansongjun, huyu, lxw}@ict.ac.cn Key Laboratory of Computer System and Architecture, Institute of Computing Technology, Chinese Academy of Sciences Nov. 17, 2009 PRDC2009

2 2 Outline Background and Motivation AVF Computing and Predicting Experimental Results Conclusions PRDC2009 Nov. 17

3 3 Soft Errors in Microprocessors PRDC2009 Nov. 17 Soft errors – –Caused by neutrons and alpha particles Technology Scaling – –Smaller transistors – –Lower threshold voltage Higher soft error rate It is important to analyze the vulnerability of different structures to soft errors

4 4 Reliability Analysis Method Compute Architectural Vulnerability Factor (AVF) for different structures [Mukerjee, MICRO’03] – –Reflect the vulnerability of a structure to soft errors AVF: The probability a soft error in a structure would result in an external visible failure – –AVF = 0% branch predictor – –AVF ≈ 100% program counter – –Higher AVF, higher vulnerable to soft errors AVF computing is important! PRDC2009 Nov. 17

5 5 Offline AVF Computing Methods Offline methods [Mukerjee, MICRO’03] – –Analyze ACE bit/un-ACE bit Sim-SODA method [Fu, Workshop of ISCA’06] – –Compute AVF for different structures (IQ, ROB, register file) – –Accurate AVF result – –Guide reliability estimation at early design stage PRDC2009 Nov. 17

6 6 Motivation of Our Work Traditional protection schemes (AR-SMT, SRT) for soft errors result in a high performance overhead [Mukerjee, ISCA’02] AVF varies significantly across different workloads and individual structures [Li, DSN’05] The AVF information can be used to guide the protection of microprocessors PRDC2009 Nov. 17

7 7 Online Methods Are Needed PRDC2009 Nov. 17 Dynamically tuning, make a trade-off between reliability and performance Offline methods are not enough Online AVF computing methods are needed –Computing AVF during program execution Time Active protection scheme AVF th AVF AVF>AVF th

8 8 Our Contributions Propose an occupancy-based online AVF computing method Predict the AVF based on history information to guide reliability design Demonstrate the efficiency of our method PRDC2009 Nov. 17

9 9 … Schematic Diagram Online AVF computing and predicting architecture Front-End … … … Load/Store Queue Issue Queue Extra Bit Reorder Buffer Extra Bit … Register File FU AVF Computing& Predicting Activated Protection scheme AVF>AVF th PRDC2009 Nov. 17

10 10 Occupancy-based online AVF computing – –The percentage of entries have been taken during a program execution Occupancy-based Method Reorder Buffer Occupancy IN OUT 4/95/93/9 PRDC2009 Nov. 17 0/9 Cycle 4Cycle 1Cycle 2Cycle 3

11 11 Key Observations Efficient to get the occupancy information during program execution Assuming the occupancy of a structure as the AVF of that structure This method takes all the bits as ACE bits, which results in a conservative AVF We need to further refine the AVF result PRDC2009 Nov. 17

12 12 Refine AVF Computing PRDC2009 Nov. 17 Instruction types –NOP instructions (NOP): not affect the program output –Dynamic dead instructions (DDI) Not affect the program output, BUT Need a long time to differentiate –ACE instructions (ACE): affect the program output Refine AVF result –Exclude NOP instructions –Counter dynamic dead instructions as ACE instructions

13 13 Our Online Computing Method PRDC2009 Nov. 17 AVF IN OUT NOP Reorder Buffer 3/9 Flag Bit 0 2/9 Cycle 4Cycle 1 Cycle 2Cycle 3 Interval Computing AVF online at interval granularity

14 14 AVF Predicting PRDC2009 Nov. 17 Activate a protection scheme when AVF > AVF th Predicting the next interval’s AVF based on the history AVF information –Algorithm 1: last-value based –Algorithm 2: average of the latest three interval’s value interval Interval length: 1000 cycles Time latest interval next interval L1L1 L2L2 L3L3 N1N1

15 15 Overall Flowchart Decode NOP ? Record Occupancy End of interval AVF computing and predicting FLAG=1 Cycle++ FLAG=0 PRDC2009 Nov. 17 No Yes No Activate Protection scheme AVF>AVF th AVF<AVF th Repeat

16 16 Experimental Setup Simulated machine configurations –4 integer ALUs, 2 integer multipliers, 2 float ALUs –IQ/ROB/LSQ 20/80/64 entries –Hybrid, 4K global + 2-level 1K local + 4K choice branch predictor –64KB 2-way L1 data cache, 2MB direct mapped L2 cache Workload –SPEC2000 Integer benchmark suite –Simulate 100M instructions starting from each SimPoint. PRDC2009 Nov. 17

17 17 Experimental Results (1/4) PRDC2009 Nov. 17 Online computed AVF for IQ 、 ROB, and LSQ

18 18 Experimental Results (2/4) PRDC2009 Nov. 17 AVF Results for different configurations Different Configurations - IQ/ROB/LSQ entries –Base : 20/80/64 –20/10/64 40/80/64 20/80/8 Config2 Config4 Config3

19 19 Experimental Results (3/4) AVF results with different predicting algorithms Algorithm 2: higher prediction accuracy PRDC2009 Nov. 17 IQROB

20 20 Experimental Results (4/4) AVF for IQ, ROB, and LSQ during executing crafty and gap PRDC2009 Nov. 17 craftygap AVF th

21 21 Conclusions We propose an occupancy-based method to compute and predict AVF online Our method can compute AVF efficiently, the difference between our method and an offline method are 0.10, 0.01, and 0.039 respectively Our method is also independent of the microprocessor configurations. Our method combines AVF to activate protection scheme, ensuring high reliability while with less performance overhead. PRDC2009 Nov. 17

22 22 Thanks! Q&A PRDC2009 Nov. 17

23 23 Backup Slices

24 24 T = 3ACE% = 0/4T = 2ACE% = 1/4 Vulnerability of a structure ACE bit and un-ACE bit T = 1ACE% = 2/4 Average number of ACE bits in a cycle Total number of bits in the structure = T = 4ACE% = 3/4 ( 2 + 1 + 0 + 3 ) / 4 ( 2 + 1 + 0 + 3 ) / 44 = PRDC2009 Nov. 17

25 25 Interval length PRDC2009 Nov. 17 Choose an appropriate interval length – –Single cycle / length of the entire application Interval length: 1000 cycles

26 26 Refine AVF Computing PRDC2009 Nov. 17 Instruction types –NOP instructions (NOP) / Dynamic dead instructions (DDI) / ACE instructions (ACE) NOP 10.7% DDI 15.8% ACE 73.5%


Download ppt "11 Online Computing and Predicting Architectural Vulnerability Factor of Microprocessor Structures Songjun Pan Yu Hu Xiaowei Li {pansongjun, huyu,"

Similar presentations


Ads by Google