stall c store b buffer a mem a ros s kaxiras isca 18 los
play

Stall c Store b Buffer a Mem A. Ros & S. Kaxiras ISCA18 - PowerPoint PPT Presentation

Aug 19, 2023 •95 likes •843 views

Goal Problem Solution Challenge Results Conclusions pdf/portada N ON -S PECULATIVE S TORE C OALESCING IN T OTAL S TORE O RDER Alberto Ros 1 Stefanos Kaxiras 2 1 Universidad de Murcia aros@ditec.um.es 2 Uppsala University

  1. Goal Problem Solution Challenge Results Conclusions pdf/portada N ON -S PECULATIVE S TORE C OALESCING IN T OTAL S TORE O RDER Alberto Ros 1 Stefanos Kaxiras 2 1 Universidad de Murcia aros@ditec.um.es 2 Uppsala University stefanos.kaxiras@it.uu.se June 4th, 2018 A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 1

  2. Goal Problem Solution Challenge Results Conclusions O UTLINE Goal : Coalescing stores into a single write operation to reduce processor stalls and memory accesses A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 2

  3. Goal Problem Solution Challenge Results Conclusions O UTLINE Goal : Coalescing stores into a single write operation to reduce processor stalls and memory accesses Problem : Coalescing can break store order (and programming intuition for x86 processors) A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 2

  4. Goal Problem Solution Challenge Results Conclusions O UTLINE Goal : Coalescing stores into a single write operation to reduce processor stalls and memory accesses Problem : Coalescing can break store order (and programming intuition for x86 processors) Solution : Atomicity gives the illusion of order but is hard to scale and costly (e.g., centralized, transactional) A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 2

  5. Goal Problem Solution Challenge Results Conclusions O UTLINE Goal : Coalescing stores into a single write operation to reduce processor stalls and memory accesses Problem : Coalescing can break store order (and programming intuition for x86 processors) Solution : Atomicity gives the illusion of order but is hard to scale and costly (e.g., centralized, transactional) Challenge : To perform multiple stores atomically without speculation (without rollback) and in a distributed manner A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 2

  6. Goal Problem Solution Challenge Results Conclusions O UTLINE Goal : Coalescing stores into a single write operation to reduce processor stalls and memory accesses Problem : Coalescing can break store order (and programming intuition for x86 processors) Solution : Atomicity gives the illusion of order but is hard to scale and costly (e.g., centralized, transactional) Challenge : To perform multiple stores atomically without speculation (without rollback) and in a distributed manner We present the first solution to this challenge A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 2

  7. Goal Problem Solution Challenge Results Conclusions O UTLINE 1 G OAL : C OALESCING STORES 2 P ROBLEM : C OALESCING BREAKS STORE ORDER 3 S OLUTION : A TOMICITY 4 C HALLENGE : D ISTRIBUTED , NON - SPECULATIVE ATOMICITY 5 R ESULTS 6 C ONCLUSIONS A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 3

  8. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc st a Store Program Buffer Order a Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  9. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc st b Store Program b Buffer Order a Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  10. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc Store Program b Buffer Order a Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  11. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc Store Program b Buffer Order a First-In First-Out (FIFO) a Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  12. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc Store Program b Buffer Order First-In First-Out (FIFO) a b Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  13. Goal Problem Solution Challenge Results Conclusions S TORE BUFFER AND T OTAL S TORE O RDER Proc Store Program Buffer Order a Total Store Order (TSO) b Mem Store operations in current x86 processors A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 4

  14. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc st c c Store b Buffer a Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  15. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc � Stall c Store b Buffer a Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  16. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc � Stall c Store b Buffer a Individual writes � Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  17. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc st c Coalescing b Store Buffer a, c Stores to the same cache line (same color in the example) Mem can coalesce in a single write A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  18. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc � No stall Coalescing b Store Buffer a, c Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  19. Goal Problem Solution Challenge Results Conclusions L IMITATIONS AND THE SOLUTION OF COALESCING Proc � No stall Coalescing b Store Buffer a, c Coalesced writes � Mem Performance and energy improvements A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 5

  20. Goal Problem Solution Challenge Results Conclusions O UTLINE 1 G OAL : C OALESCING STORES 2 P ROBLEM : C OALESCING BREAKS STORE ORDER 3 S OLUTION : A TOMICITY 4 C HALLENGE : D ISTRIBUTED , NON - SPECULATIVE ATOMICITY 5 R ESULTS 6 C ONCLUSIONS A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 6

  21. Goal Problem Solution Challenge Results Conclusions T HE PROBLEM OF COALESCING STORES Proc Coalescing b Store Buffer a, c Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 7

  22. Goal Problem Solution Challenge Results Conclusions T HE PROBLEM OF COALESCING STORES Proc Coalescing b Store Buffer a, c First green ⇒ c overtakes b Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 7

  23. Goal Problem Solution Challenge Results Conclusions T HE PROBLEM OF COALESCING STORES Proc Coalescing b Store Buffer a, c First green ⇒ c overtakes b First blue ⇒ b overtakes a Mem A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 7

  24. Goal Problem Solution Challenge Results Conclusions T HE PROBLEM OF COALESCING STORES Proc Coalescing b Store Buffer a, c First green ⇒ c overtakes b First blue ⇒ b overtakes a Mem Store order? � In the paper: A new litmus test that captures a TSO violation when breaking store order A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 7

  25. Goal Problem Solution Challenge Results Conclusions O UTLINE 1 G OAL : C OALESCING STORES 2 P ROBLEM : C OALESCING BREAKS STORE ORDER 3 S OLUTION : A TOMICITY 4 C HALLENGE : D ISTRIBUTED , NON - SPECULATIVE ATOMICITY 5 R ESULTS 6 C ONCLUSIONS A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 8

  26. Goal Problem Solution Challenge Results Conclusions A TOMICITY : I LLUSION OF STORE ORDER Proc Coalescing b Store Buffer a, c Store order ⇒ Atomicity Mem Coalescing forms atomic write groups A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 9

  27. Goal Problem Solution Challenge Results Conclusions F ORMING ATOMIC WRITE GROUPS st c b a Match younger write to the same cache line A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 10

  28. Goal Problem Solution Challenge Results Conclusions F ORMING ATOMIC WRITE GROUPS st c b b a, c a Writes to a, b, and c are indivisible A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 10

  29. Goal Problem Solution Challenge Results Conclusions F ORMING ATOMIC WRITE GROUPS st c st d b b b a, c a, c a A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 10

  30. Goal Problem Solution Challenge Results Conclusions F ORMING ATOMIC WRITE GROUPS st c st d st e d b b b b a, c a, c a, c a A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 10

  31. Goal Problem Solution Challenge Results Conclusions F ORMING ATOMIC WRITE GROUPS st c st d st e d d b, e b b b b a, c a, c a, c a, c a Writes to a, b, c, d, and e are indivisible A. Ros & S. Kaxiras ISCA’18 @ Los Angeles, CA, USA June 4th, 2018 10

Recommend

Dynamic Memory Dependence Predication Zhaoxiang Jin and Soner nder ISCA-2018, Los Angeles

Dynamic Memory Dependence Predication Zhaoxiang Jin and Soner nder ISCA-2018, Los Angeles

Dynamic Memory Dependence Predication Zhaoxiang Jin and Soner nder ISCA-2018, Los Angeles 6/19/2018 Background 1. Store instructions do not update the cache until they are retired (too late). 2. Store queue is implemented to keep the

607 views • 25 slides
An introduction to Isca Dr Stephen I. Thomson Plan Part 1: What are these practical

An introduction to Isca Dr Stephen I. Thomson Plan Part 1: What are these practical

An introduction to Isca Dr Stephen I. Thomson Plan Part 1: What are these practical sessions for? What code will we be using? Isca What is Isca - components, options? Part 2: How do we run Isca / configure it / modify it?

792 views • 45 slides
IT 4500 : Information Security Secure Programming Dr Joe Francom Buffer Overflow What it is? A

IT 4500 : Information Security Secure Programming Dr Joe Francom Buffer Overflow What it is? A

IT 4500 : Information Security Secure Programming Dr Joe Francom Buffer Overflow What it is? A buffer overflow occurs when a program or process tries to store more data in a buffer (temporary data storage area) than it was intended to hold.

221 views • 3 slides
From Array Domains to Abstract Interpretation Under Store-Buffer-Based Memory Models Thibault

From Array Domains to Abstract Interpretation Under Store-Buffer-Based Memory Models Thibault

From Array Domains to Abstract Interpretation Under Store-Buffer-Based Memory Models Thibault Suzanne, Antoine Min Static Analysis: 23rd International Symposium, SAS 2016 September, 2016, Edinburgh, UK Pirmin Schmid Seminar Software

759 views • 50 slides
1 Store Buffer Design Example Memory Dependence Any load instruction receives the memory Store

1 Store Buffer Design Example Memory Dependence Any load instruction receives the memory Store

Register and Memory Dependences Store: SW Rt, A(Rs) LW Rt, A(Rs) 1. 1. Calculate effective Calculate effective Lecture 10: Memory Dependence memory address memory address Detection and Speculation dependent on Rs dependent on Rs

479 views • 3 slides
Disks, Memories & Buffer Management The two offices of memory are collection and

Disks, Memories & Buffer Management The two offices of memory are collection and

Disks, Memories & Buffer Management The two offices of memory are collection and distribution. - Samuel Johnson CS3223 - Storage 1 What does a DBMS Store? Relations Actual data Indexes Data structures to speed up

292 views • 28 slides
Buffer Manager Each relation is can be mapped onto many files (each file containing data from one

Buffer Manager Each relation is can be mapped onto many files (each file containing data from one

Buffer Management The database buffer is the mediator between the basic file system and the tuple-oriented file system. The buffer managers task is to make the pages addressable in main memory and to coordinate the writing of pages to disk

1.12k views • 8 slides
2 Measured CPU Cycles 30% BUFFER POOL 28% LOCKING RECOVERY 30% 12% REAL WORK OLTP THROUGH

2 Measured CPU Cycles 30% BUFFER POOL 28% LOCKING RECOVERY 30% 12% REAL WORK OLTP THROUGH

2 Measured CPU Cycles 30% BUFFER POOL 28% LOCKING RECOVERY 30% 12% REAL WORK OLTP THROUGH THE LOOKING GLASS, AND WHAT WE FOUND THERE 3 SIGMOD, pp. 981-992, 2008. H-STORE: A HIGH-PERFORMANCE, DISTRIBUTED MAIN MEMORY TRANSACTION

137 views • 12 slides
Introduction Buffer Overflows Buffer overflows were the most common form of security

Introduction Buffer Overflows Buffer overflows were the most common form of security

Introduction Buffer Overflows Buffer overflows were the most common form of security vulnerability in the 90s. Buffer overflows dominate in the area of remote network penetration vulnerabilities. CS 465 They represent one of the

651 views • 4 slides
Buffer Overflows with Content 2 A Process Stack Buffer Overflow Common Techniques employed

Buffer Overflows with Content 2 A Process Stack Buffer Overflow Common Techniques employed

1 Buffer Overflows with Content 2 A Process Stack Buffer Overflow Common Techniques employed in buffer overflow exploits to create backdoors Execution of additional network services via the INETD daemon The addition of new users

497 views • 30 slides
External buffer Raslan Darawsheh Mellanox External buffer First was introduced by Olivier

External buffer Raslan Darawsheh Mellanox External buffer First was introduced by Olivier

x External buffer Raslan Darawsheh Mellanox External buffer First was introduced by Olivier in his presentation in 2016. The support for External buffer has been there since DPDK 18.05 2 External buffer: Contd 3 External

544 views • 12 slides
Lab 2: Buffer Overflows Fengwei Zhang SUSTech CS 315 Computer Security 1 Buffer Overflows

Lab 2: Buffer Overflows Fengwei Zhang SUSTech CS 315 Computer Security 1 Buffer Overflows

Lab 2: Buffer Overflows Fengwei Zhang SUSTech CS 315 Computer Security 1 Buffer Overflows One of the most common vulnerabilities in software Programming languages commonly associated with buffer overflows including C and C++

1.14k views • 17 slides
Buffer Overflows and Defenses CS 419: Computer Security Lecture 10 and 11 Buffer Overflow a

Buffer Overflows and Defenses CS 419: Computer Security Lecture 10 and 11 Buffer Overflow a

Buffer Overflows and Defenses CS 419: Computer Security Lecture 10 and 11 Buffer Overflow a very common attack mechanism from 1988 Morris Worm to Code Red, Slammer, Sasser and many others prevention techniques known still of major concern

873 views • 53 slides
Sapporo Sapporo Namba Namba Shinjuku Shinjuku Store Store Store Store West Store West

Sapporo Sapporo Namba Namba Shinjuku Shinjuku Store Store Store Store West Store West

Sapporo Sapporo Namba Namba Shinjuku Shinjuku Store Store Store Store West Store West Store Securities Code: 3048 B A C Yurakucho Yurakucho Fujisawa Fujisawa Kashiwa Kashiwa Store Store Store Store Store Store D F

587 views • 32 slides
TinyOS Determine when Fill message Specify Pass buffer message buffer Network Communication

TinyOS Determine when Fill message Specify Pass buffer message buffer Network Communication

Inter-Node Communication General idea: Sender: TinyOS Determine when Fill message Specify Pass buffer message buffer Network Communication buffer with data Recipients to OS can be reused Computer Network Receiver:

176 views • 3 slides
Week 03 Lectures PostgreSQL Buffer Manager 1/95 PostgreSQL buffer manager: provides a shared

Week 03 Lectures PostgreSQL Buffer Manager 1/95 PostgreSQL buffer manager: provides a shared

Week 03 Lectures PostgreSQL Buffer Manager 1/95 PostgreSQL buffer manager: provides a shared pool of memory buffers for all backends all access methods get data from disk via buffer manager Buffers are located in a large region of shared

505 views • 25 slides
Sensors CSCI545 Introduction to Robotics Hadi Moradi Previous Lecture DC motors DC

Sensors CSCI545 Introduction to Robotics Hadi Moradi Previous Lecture DC motors DC

Sensors CSCI545 Introduction to Robotics Hadi Moradi Previous Lecture DC motors DC motors Inefficient Operating voltage Operating current Stall current Stall torque Stall torque Gearing up and down Gear

826 views • 37 slides
Riparian Buffer or Riparian Forest Buffer Equivalency Demonstration and Offsetting Water

Riparian Buffer or Riparian Forest Buffer Equivalency Demonstration and Offsetting Water

Implementation of Act 162 of 2014 Riparian Buffer or Riparian Forest Buffer Equivalency Demonstration and Offsetting Water Resources Advisory Committee August 12, 2015 Tom Wolf, Governor John Quigley, Secretary Agenda 1. Overview of Act 162

530 views • 37 slides
Chapter 10 Buffer Overflow Buffer Overflow Common attack mechanism first wide use by

Chapter 10 Buffer Overflow Buffer Overflow Common attack mechanism first wide use by

Chapter 10 Buffer Overflow Buffer Overflow Common attack mechanism first wide use by the Morris Worm in 1988 Prevention techniques known NX bit, stack canaries, ASLR Still of major concern Recent examples:

354 views • 21 slides
Buffer Trees Lars Arge. The Buffer Tree: A New Technique for Optimal I/O Algorithms . In

Buffer Trees Lars Arge. The Buffer Tree: A New Technique for Optimal I/O Algorithms . In

Buffer Trees Lars Arge. The Buffer Tree: A New Technique for Optimal I/O Algorithms . In Proceedings of Fourth Workshop on Algorithms and Data Structures (WADS), Lecture Notes in Computer Science Vol. 955, Springer-Verlag, 1995, 334-345. 1

518 views • 27 slides
AKC Education Webinar Series: Social Media Resources Buffer: What Is It? Buffer is a software

AKC Education Webinar Series: Social Media Resources Buffer: What Is It? Buffer is a software

AKC Education Webinar Series: Social Media Resources Buffer: What Is It? Buffer is a software application for the web and mobile, designed to manage accounts in social networks, by providing the means for a user to schedule posts to Twitter,

506 views • 35 slides
More Vulnerabilities (buffer overreads, format string, integer overflow, heap overflows) Chester

More Vulnerabilities (buffer overreads, format string, integer overflow, heap overflows) Chester

More Vulnerabilities (buffer overreads, format string, integer overflow, heap overflows) Chester Rebeiro Indian Institute of Technology Madras Buffer Overreads 2 Buffer Overread Example 3 Buffer Overread Example len read from command line

906 views • 76 slides
Investigation of Airfoil Aero Acoustics due to Different Stall Mechanism using Large Eddy

Investigation of Airfoil Aero Acoustics due to Different Stall Mechanism using Large Eddy

Investigation of Airfoil Aero Acoustics due to Different Stall Mechanism using Large Eddy Simulation (LES) Voo Keng Soon Vince Lim Nee Sheng Winson Tan Chun Hern DSO National Laboratories 05 Nov 2012 Slide 1 Brief Review of Available

526 views • 24 slides
Y U MEETING NG A AGE GEND NDA Fis isca cal Y Year ear-End P nd Planni nning ng

Y U MEETING NG A AGE GEND NDA Fis isca cal Y Year ear-End P nd Planni nning ng

Y U MEETING NG A AGE GEND NDA Fis isca cal Y Year ear-End P nd Planni nning ng COV OVID-19 19 Em Emergency Pay & & Time T Tra racking T Template Adob dobe Si Sign Doc Docs pres esen ented b by A/

988 views • 41 slides

More recommend