thermos
play

ThermOS System Support for Dynamic Thermal Management of Chip - PowerPoint PPT Presentation

Nov 23, 2022 •410 likes •824 views

22nd International Conference on Parallel Architectures and Compilation Techniques (PACT -22), 2013 September 9, 2013 Edinburgh, Scotland, UK ThermOS System Support for Dynamic Thermal Management of Chip Multi-Processors Filippo Sironi

  1. 22nd International Conference on Parallel Architectures and Compilation Techniques (PACT -22), 2013 September 9, 2013 Edinburgh, Scotland, UK ThermOS System Support for Dynamic Thermal Management of Chip Multi-Processors Filippo Sironi (sironi@elet.polimi.it) Martina Maggio, Riccardo Cattaneo, Giovanni F. Del Nero Donatella Sciuto, Marco D. Santambrogio 1

  2. DVFS is dangerous! (I know this is scary) temperature increase (°C) 20 10 swaptions @ 2.80 GHz ab @ 2.80 GHz 0 0 100 200 300 400 500 600 time (s) 2

  3. DVFS is dangerous! (I know this is scary) temperature increase (°C) 20 10 swaptions @ 2.80 GHz ab @ 2.80 GHz 0 0 100 200 300 400 500 600 time (s) 2

  4. DVFS is dangerous! (I know this is scary) DVFS from 2.80 to 2.13 GHz Δ 1 temperature increase (°C) 20 Δ 2 10 swaptions @ 2.80 GHz Δ 2 ab @ 2.80 GHz swaptions w/ DVFS ab w/ DVFS 0 0 100 200 300 400 500 600 time (s) 2

  5. DVFS is dangerous! (I know this is scary) DVFS from 2.80 to 2.13 GHz Δ 1 temperature increase (°C) 20 it may impair multi-programmed workloads... Δ 2 think about multi-tenant virtualization 10 infrastructures! swaptions @ 2.80 GHz Δ 2 ab @ 2.80 GHz swaptions w/ DVFS ab w/ DVFS 0 0 100 200 300 400 500 600 time (s) 2

  6. Idle cycle injection improves! temperature increase (°C) 20 10 swaptions @ 2.80 GHz ab @ 2.80 GHz 0 0 100 200 300 400 500 600 time (s) 3

  7. Idle cycle injection improves! temperature increase (°C) 20 10 swaptions @ 2.80 GHz ab @ 2.80 GHz 0 0 100 200 300 400 500 600 time (s) 3

  8. Idle cycle injection improves! Δ 1 temperature increase (°C) 20 10 swaptions @ 2.80 GHz ab @ 2.80 GHz swaptions w/ ThermOS ab w/ ThermOS 0 0 100 200 300 400 500 600 time (s) 3

  9. Outline • Why DTM • DTM in commodity CMPs • ThermOS • Related work • Conclusions and Future work 4

  10. Why DTM • Transistors per unit of area are still increasing (Moore’s law) • Power density is getting worse as lithography advances (failure of Dennard’s law) • High temperature impairs performance, energy efficiency, and reliability (Srinivasan et al. in ISCA’04 [3]) 5

  11. DTM in commodity CMPs • Commodity CMPs exploits DVFS • DVFS has chip-wide side effects • DVFS with core-wide side effects becomes costly as soon as the core count overcomes 2 (Kim et al. in HPCA’08 [8]) • Intel Haswell supports per-core DVFS but integrated voltage regulators may cause high temperature • Side effects are especially bad in shared environments (e.g., multi-tenant virtualized infrastructures) 6

  12. DTM in commodity CMPs • Commodity CMPs exploits DVFS • DVFS has chip-wide side effects • DVFS with core-wide side effects becomes costly as soon as the core count overcomes 2 (Kim et al. in HPCA’08 [8]) • Intel Haswell supports per-core DVFS but integrated voltage regulators may cause high temperature • Side effects are especially bad in shared environments (e.g., multi-tenant virtualized infrastructures) software-driven DTM of CMPs 6

  13. ThermOS • Linear discrete-time modeling of temperature dynamic • Commodity solution to measure temperature (i.e., DTSs and MSRs) • Formal feedback control for idle cycle determination • Idle cycle injection via operating system scheduling 7

  14. Modeling of temperature dynamic • Modeling approaches either have shortcomings (Wattch, Brooks and Martonosi in HPCA’01 [14]) or require too many information and become impractical (HotSpot, Skadron et al. in TACO’01 [1]) • No need to understand the full temperature dynamic: we need the dynamic near the temperature threshold 8

  15. Modeling of temperature dynamic 50 w/o ICI temperature increase (°C) 40 30 20 10 0 0 50 100 150 200 time (ms) 9

  16. Modeling of temperature dynamic 50 w/o ICI temperature increase (°C) 40 30 20 10 0 0 50 100 150 200 time (ms) 9

  17. Modeling of temperature dynamic 50 w/o ICI temperature increase (°C) w/ ICI 40 30 20 10 0 0 50 100 150 200 time (ms) 9

  18. Modeling of temperature dynamic 50 w/o ICI temperature increase (°C) w/ ICI 40 30 20 40 10 80 90 100 110 0 0 50 100 150 200 time (ms) 9

  19. Modeling of temperature dynamic 50 w/o ICI temperature increase (°C) w/ ICI 40 30 T(k + 1) = a T(k) + b I(k) 20 40 10 80 90 100 110 0 0 50 100 150 200 time (ms) 9

  20. Linear discrete-time thermal model: offline estimation • Low overhead but requires the model to be conservative • Linear regression over 70% of a dataset of over 1.5 million of {temperature_next, temperature, idle} tuples; different regressions yields 95% prediction accuracy over the remaining 30% of the dataset • Estimated variances of a and b parameters is almost negligible 10

  21. Formal feedback control • Proportional-Integral (PI) controller • proportional term to capture the dependency from the current error (i.e., expected minus current temperature) • integral term to get the dependency from past errors • Synthesis of a “stable by definition” controller • Robust to estimation errors of the b parameter 11

  22. Formal feedback control idle = previous idle + • Proportional-Integral (PI) controller • proportional term to capture the dependency from the current error A current error - (i.e., expected minus current temperature) • integral term to get the dependency from past errors • Synthesis of a “stable by definition” controller B previous error • Robust to estimation errors of the b parameter 11

  23. Formal feedback control I(k) = I(k - 1) + • Proportional-Integral (PI) controller • proportional term to capture the dependency from the current error e(k) (1 - p) / b - (i.e., expected minus current temperature) • integral term to get the dependency from past errors • Synthesis of a “stable by definition” controller e(k - 1) a (1 - p) / b • Robust to estimation errors of the b parameter 11

  24. Idle cycle injection • Do not affect the scheduling of high-priority and vital tasks (e.g., real-time task and kernel tasks) • Exploit task scheduling and cpuidle (Pallipadi et al. in Linux Symposium’07 [10]) and is not invasive thanks to the use of the dynamic tick code • Alternative solutions are suboptimal from either a software engineering or an effectiveness stand point 12

  25. ThermOS _ T e I% I T C A P + - T S 13

  26. ThermOS _ T e I% I T C A P + - T S one feedback controller per core 13

  27. ThermOS 10 ms of control period _ T e I% I T C A P + - T S one feedback controller per core 13

  28. ThermOS max. 80% of control period 10 ms of control period _ T e I% I T C A P + - T S one feedback controller per core 13

  29. Evaluation platform • 4-core Intel Xeon (Nehalem) • From 1.60 GHz to 2.8 GHz • C0, C1E (3 us latency), C3 (20 us latency plus other overheads), and C6 (200 us latency plus many other overheads) C-states • Ambient temperature about (20 Celsius plus/minus 1) • Idle temperature about (28-32 Celsius plus/minus 1 depending on the core) • Modified Linux kernel 3.4 • PARSEC 2.1 benchmark 14

  30. Thermal profile temperature increase (°C) 50 40 swaptions @ core 0/2 swaptions @ core 1 swaptions @ core 3 30 470 480 490 500 510 520 530 time (s) 15

  31. Thermal profile temperature increase (°C) 50 temperature is not symmetric in CMPs 40 swaptions @ core 0/2 swaptions @ core 1 swaptions @ core 3 30 470 480 490 500 510 520 530 time (s) 15

  32. Research questions • Can ThermOS constraint the temperature and selectively affect applications in a multi- programmed workload? • How much ThermOS is efficient w.r.t. state of the art solutions? 16

  33. Management of multi-programmed workloads 55 temperature increase (°C) 50 45 40 37 swaptions @ core 0/2 35 swaptions @ core 1 swaptions @ core 3 30 510 520 530 540 550 560 570 time (s) 17

  34. Management of multi-programmed workloads 55 temperature increase (°C) 50 core 3: 91% in C0, 5% in C1E, 4% in C3 45 core 2: 91% in C0, 6% in C1E, 3% in C3 core 0: 91% in C0, 7% in C1E, 2% in C3 40 core 1: 92% in C0, 8% in C1E, 0% in C3 37 swaptions @ core 0/2 35 swaptions @ core 1 swaptions @ core 3 30 510 520 530 540 550 560 570 time (s) 17

  35. State of the art solutions • Dimetrodon (Bailis et al. in DAC’11 [4]) • Probabilistic feedforward control inside the FreeBSD 7.2 task scheduler • We swipe the idle quantum/probability configuration space • VFS • We statically select the following frequencies (and the associated voltages): 2.79, 2.66, 2.53, 2.39, 2.26, 2.13 GHz 18

  36. Efficiency with multi-programmed workloads 100 performance (%) 90 80 Dimetrodon ThermOS VFS 70 0 10 20 30 40 temperature decrease (%) 19

  37. Efficiency with multi-programmed workloads 100 dynamic power is proportional to C V**2 f performance (%) 90 as the supply voltage approaches its threshold 80 DVFS will loose most of its efficiency Dimetrodon ThermOS VFS 70 0 10 20 30 40 temperature decrease (%) 19

More recommend