DOCSLIB.ORG

Dynamic Scheduling (OOO) Via Tomasulo's Approach

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Dynamic Scheduling (OOO) Via Tomasulo's Approach Lecture 16: Instruction Level Parallelism -- Dynamic Scheduling (OOO) via Tomasulo’s Approach CSE 564 Computer Architecture Summer 2017 Department of Computer Science and Engineering Yonghong Yan [email protected] www.secs.oakland.edu/~yan Topics for Instruction Level Parallelism § ILP Introduction, Compiler Techniques and Branch Prediction – 3.1, 3.2, 3.3 § Dynamic Scheduling (OOO) – 3.4, 3.5 and C.5, C.6 and C.7 (FP pipeline and scoreboard) § Hardware Speculation and Static Superscalar/VLIW – 3.6, 3.7 § Dynamic Scheduling, Multiple Issue and Speculation – 3.8, 3.9 § ILP Limitations and SMT – 3.10, 3.11, 3.12 2 Acknowledge and Copyright § Slides adapted from – UC Berkeley course “Computer Science 252: Graduate Computer Architecture” of David E. Culler Copyright(C) 2005 UCB – UC Berkeley course Computer Science 252, Graduate Computer Architecture Spring 2012 of John Kubiatowicz Copyright(C) 2012 UCB – Computer Science 152: Computer Architecture and Engineering, Spring 2016 by Dr. George Michelogiannakis from UC Berkeley § https://passlab.github.io/CSE564/copyrightack.html 3 Complex Pipelining: Motivation § Why would we want more than our in-order pipeline? Physical Physical Address Inst. Address Data Decode PC Cache D E + M Cache W Physical Memory Controller Physical Address Address Physical Address Main Memory (DRAM) 4 Complex Pipelining: Motivation Pipelining becomes complex when we want high performance in the presence of: § Long latency or partially pipelined floating-point units – Not all instructions are floating point or integer § Memory systems with variable access time – For example cache misses § Multiple arithmetic and memory units 5 Floating Point Representation § IEEE standard 754 Value = (-1)s * 1.mantissa * 2(exp-127) Exponent = 0 has special meaning 6 Floating-Point Unit (FPU) § Much more hardware than an integer unit – A simple FPU takes 150,000 gates. Verification complex. Some exceptions specific to floating point. – Integer FU to the order of thousands § Common to have several FPU’s – Some integer, some floating point § Common to have different types of FPU’s: Fadd, Fmul, Fdiv, … § An FPU may be pipelined, partially pipelined or not pipelined § To operate several FPU’s concurrently the FP register file needs to have more read and write ports 7 Unpipelined FP EXE Stage § FP takes loops to compute § Much longer clock period Single-cycle FPU is a bad idea 8 Latency and Interval § Latency – The number of intervening cycles between an instruction that produces a result and an instruction that uses the result. – Usually the number of stages after EX that an instruction produces a result » ALU Integer 0, Load latency 1 § Initiation or repeat interval – the number of cycles that must elapse between issuing two operations of a given type à structural hazards 9 Pipelined FP EXE § Increased stall for RAW hazards 10 Breaking Our Assumption of Integer Pipeline § The divide unit is not fully pipelined – structural hazards can occur » need to be detected and stall incurred. § The instructions have varying running times – the number of register writes required in a cycle can be > 1 § Instructions no longer reach WB in order – Write after write (WAW) hazards are possible » Note that write after read (WAR) hazards are not possible, since the register reads always occur in ID. § Instructions can complete in a different order than they were issued (out-of-order complete) – causing problems with exceptions § Longer latency of operations – stalls for RAW hazards will be more frequent. 11 Hazards and Forwarding for Longer- Latency Pipeline § H 12 Stalls of FP Operations § SPEC89 FP § Latency average § FP add, subtract, or convert – 1.7 cycles, or 56% of the latency (3 cycles). § Multiplies and divides – 2.8 and 14.2, respectively, or 46% and 59% of the corresponding latency. § Structural hazards for divides are rare – since the divide frequency is low. 13 Stalls per FP Operation § The total number of stalls per instruction – ranges from 0.65 for su2cor to 1.21 for doduc, with an average of 0.87. – FP result stalls dominate in all cases, with an average of 0.71 stalls per instruction, or 82% of the stalled cycles. 14 Problems Arising From Writes § If we issue one instruction per cycle, how can we avoid structural hazards at the writeback stage and out-of-order writeback issues? § WAW Hazards WAW Hazards 15 Complex In-Order Pipeline Inst. Data Decode GPRs PC Mem D X1 + X2 Mem X3 W § Delay writeback so all operations have same latency to W stage FPRs X1 X2 FAdd X3 W – Write ports never oversubscribed (one inst. in & one inst. out every cycle) – Stall pipeline on long latency operations, e.g., divides, X2 FMul X3 cache misses – Handle exceptions in-order at commit point Unpipeline How to prevent increased writeback latency FDiv X2 d divider X3 from slowing down single cycle integer Commit opera:ons? Bypassing Point 16 Floating-Point ISA § Interaction between floating-point datapath and integer datapath is determined by ISA § RISC-V ISA – separate register files for FP and Integer instructions » the only interaction is via a set of move/convert instructions (some ISA’s don’t even permit this) – separate load/store for FPR’s and GPR’s (general purpose registers) but both use GPR’s for address calculation – FP compares write integer registers, then use integer branch 17 Realistic Memory Systems Common approaches to improving memory performance: § Caches - single cycle except in case of a miss =>stall § Banked memory - multiple memory accesses => bank conflicts § split-phase memory operations (separate memory request from response), many in flight => out-of-order responses Latency of access to the main memory is usually much greater than one cycle and oHen unpredictable Solving this problem is a central issue in computer architecture 18 Multiple-Cycles MEM Stage § MIPS R4000 § IF: First half of instruction fetch; PC selection actually happens here, together with initiation of instruction cache access. § IS: Second half of instruction fetch, complete instruction cache access. § RF: Instruction decode and register fetch, hazard checking, and instruction cache hit detection. § EX: Execution, which includes effective address calculation, ALU operation, and branch-target computation and condition evaluation. § DF: Data fetch, first half of data cache access. § DS: Second half of data fetch, completion of data cache access. § TC: Tag check, to determine whether the data cache access hit. § WB: Write-back for loads and register-register operations. 19 2-Cycles Load Delay § 2 20 3-Cycle Branch Delay when Taken 21 Dynamic Scheduling § Data Hazards § Control Hazards 22 Types of Data Hazards Consider execuLng a sequence of rk <= ri op rj type of instrucLons Data-dependence r3 <= r1 op r2 Read-aHer-Write r5 <= r3 op r4 (RAW) hazard AnL-dependence r3 <= r1 op r2 Write-aHer-Read r1 <= r4 op r5 (WAR) hazard Output-dependence r3 <= r1 op r2 Write-aHer-Write r3 <= r6 op r7 (WAW) hazard 23 Register vs. Memory Dependence Data hazards due to register operands can be determined at the decode stage, but data hazards due to memory operands can be determined only after computing the effective address Store: M[r1 + disp1] <= r2 ! Load: r3 <= M[r4 + disp2]! ! Does (r1 + disp1) = (r4 + disp2) ? 24 Data Hazards: An Example I1 FDIV.D f6, f6, f4 I2 FLD f2, 45(x3) I3 FMUL.D f0, f2, f4 I4 FDIV.D f8, f6, f2 I5 FSUB.D f10, f0, f6 I6 FADD.D f6, f8, f2 RAW Hazards WAR Hazards WAW Hazards 25 Instruction Scheduling I1 FDIV.D f6, f6, f4 I1 I2 FLD f2, 45(x3) I FMULT.D f0, f2, f4 3 I2 I4 FDIV.D f8, f6, f2 I3 I5 FSUB.D f10, f0, f6 I FADD.D f6, f8, f2 6 I4 Valid orderings: I5 in-order I1 I2 I3 I4 I5 I6 out-of-order I 2 I1 I3 I4 I5 I6 I6 out-of-order I1 I2 I3 I5 I4 I6 26 Out-of-order Completion In-order Issue Latency I1 FDIV.D f6, f6, f4 4 I2 FLD f2, 45(x3) 1 I3 FMULT.D f0, f2, f4 3 I4 FDIV.D f8, f6, f2 4 I5 FSUB.D f10, f0, f6 1 I6 FADD.D f6, f8, f2 1 in-order comp 1 2 1 2 3 4 3 5 4 6 5 6 out-of-order comp 1 2 2 3 1 4 3 5 5 4 6 6 Underlines are completes 27 Dynamic Scheduling § Rearrange order of instructions to reduce stalls while maintaining data flow – Minimize RAW Hazards – Minimize WAW and WAR hazards via Register Renaming – Between registers and memory hazards § Advantages: – Compiler doesn’t need to have knowledge of microarchitecture – Handles cases where dependencies are unknown at compile time § Disadvantage: – Substantial increase in hardware complexity – Complicates exceptions 28 Dynamic Scheduling § Dynamic scheduling implies: – Out-of-order execution – Out-of-order completion § Creates more possibility for WAR and WAW hazards § Scoreboard: C.6 – CDC6600 in 1963 § Tomasulo’s Approach – Tracks when operands are available – Introduces register renaming in hardware » Minimizes WAW and WAR hazards 29 Register Renaming § Example: DIV.D F0,F2,F4 ADD.D F6,F0,F8 Anti-dependence on F8 S.D F6,0(R1) SUB.D F8,F10,F14 Output dependence on F6 MUL.D F6,F10,F8 30 Register Renaming § Example: DIV.D F0,F2,F4 ADD.D F6,F0,F8 DIV.D F0,F2,F4 S.D F6,0(R1) ADD.D F6,F0,F8 SUB.D F8,F10,F14 S.D S,0(R1) MUL.D F6,F10,F8 SUB.D T,F10,F14 MUL.D T,F10,T § Now only RAW hazards remain, which can be strictly ordered 31 Tomasulo Algorithm § For IBM 360/91 about 3 years after CDC 6600 (1966) § Goal: High Performance without special compilers § Differences between IBM 360 & CDC 6600 ISA – IBM has only 2 register specifiers/instr vs.
Load more

Recommended publications
  • IEEE Paper Template in A4
    Vasantha.N.S. et al, International Journal of Computer Science and Mobile Computing, Vol.6 Issue.6, June- 2017, pg. 302-306 Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology ISSN 2320–088X IMPACT FACTOR: 6.017 IJCSMC, Vol. 6, Issue. 6, June 2017, pg.302 – 306 Enhancing Performance in Multiple Execution Unit Architecture using Tomasulo Algorithm Vasantha.N.S.1, Meghana Kulkarni2 ¹Department of VLSI and Embedded Systems, Centre for PG Studies, VTU Belagavi, India ²Associate Professor Department of VLSI and Embedded Systems, Centre for PG Studies, VTU Belagavi, India 1 [email protected] ; 2 [email protected] Abstract— Tomasulo’s algorithm is a computer architecture hardware algorithm for dynamic scheduling of instructions that allows out-of-order execution, designed to efficiently utilize multiple execution units. It was developed by Robert Tomasulo at IBM. The major innovations of Tomasulo’s algorithm include register renaming in hardware. It also uses the concept of reservation stations for all execution units. A common data bus (CDB) on which computed values broadcast to all reservation stations that may need them is also present. The algorithm allows for improved parallel execution of instructions that would otherwise stall under the use of other earlier algorithms such as scoreboarding. Keywords— Reservation Station, Register renaming, common data bus, multiple execution unit, register file I. INTRODUCTION The instructions in any program may be executed in any of the 2 ways namely sequential order and the other is the data flow order. The sequential order is the one in which the instructions are executed one after the other but in reality this flow is very rare in programs.
    [Show full text]
  • Computer Science 246 Computer Architecture Spring 2010 Harvard University
    Computer Science 246 Computer Architecture Spring 2010 Harvard University Instructor: Prof. David Brooks [email protected] Dynamic Branch Prediction, Speculation, and Multiple Issue Computer Science 246 David Brooks Lecture Outline • Tomasulo’s Algorithm Review (3.1-3.3) • Pointer-Based Renaming (MIPS R10000) • Dynamic Branch Prediction (3.4) • Other Front-end Optimizations (3.5) – Branch Target Buffers/Return Address Stack Computer Science 246 David Brooks Tomasulo Review • Reservation Stations – Distribute RAW hazard detection – Renaming eliminates WAW hazards – Buffering values in Reservation Stations removes WARs – Tag match in CDB requires many associative compares • Common Data Bus – Achilles heal of Tomasulo – Multiple writebacks (multiple CDBs) expensive • Load/Store reordering – Load address compared with store address in store buffer Computer Science 246 David Brooks Tomasulo Organization From Mem FP Op FP Registers Queue Load Buffers Load1 Load2 Load3 Load4 Load5 Store Load6 Buffers Add1 Add2 Mult1 Add3 Mult2 Reservation To Mem Stations FP adders FP multipliers Common Data Bus (CDB) Tomasulo Review 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 LD F0, 0(R1) Iss M1 M2 M3 M4 M5 M6 M7 M8 Wb MUL F4, F0, F2 Iss Iss Iss Iss Iss Iss Iss Iss Iss Ex Ex Ex Ex Wb SD 0(R1), F0 Iss Iss Iss Iss Iss Iss Iss Iss Iss Iss Iss Iss Iss M1 M2 M3 Wb SUBI R1, R1, 8 Iss Ex Wb BNEZ R1, Loop Iss Ex Wb LD F0, 0(R1) Iss Iss Iss Iss M Wb MUL F4, F0, F2 Iss Iss Iss Iss Iss Ex Ex Ex Ex Wb SD 0(R1), F0 Iss Iss Iss Iss Iss Iss Iss Iss Iss M1 M2
    [Show full text]
  • Dynamic Scheduling
    Dynamically-Scheduled Machines ! • In a Statically-Scheduled machine, the compiler schedules all instructions to avoid data-hazards: the ID unit may require that instructions can issue together without hazards, otherwise the ID unit inserts stalls until the hazards clear! • This section will deal with Dynamically-Scheduled machines, where hardware- based techniques are used to detect are remove avoidable data-hazards automatically, to allow ‘out-of-order’ execution, and improve performance! • Dynamic scheduling used in the Pentium III and 4, the AMD Athlon, the MIPS R10000, the SUN Ultra-SPARC III; the IBM Power chips, the IBM/Motorola PowerPC, the HP Alpha 21264, the Intel Dual-Core and Quad-Core processors! • In contrast, static multiple-issue with compiler-based scheduling is used in the Intel IA-64 Itanium architectures! • In 2007, the dual-core and quad-core Intel processors use the Pentium 5 family of dynamically-scheduled processors. The Itanium has had a hard time gaining market share.! Ch. 6, Advanced Pipelining-DYNAMIC, slide 1! © Ted Szymanski! Dynamic Scheduling - the Idea ! (class text - pg 168, 171, 5th ed.)" !DIVD ! !F0, F2, F4! ADDD ! !F10, F0, F8 !- data hazard, stall issue for 23 cc! SUBD ! !F12, F8, F14 !- SUBD inherits 23 cc of stalls! • ADDD depends on DIVD; in a static scheduled machine, the ID unit detects the hazard and causes the basic pipeline to stall for 23 cc! • The SUBD instruction cannot execute because the pipeline has stalled, even though SUBD does not logically depend upon either previous instruction! • suppose the machine architecture was re-organized to let the SUBD and subsequent instructions “bypass” the previous stalled instruction (the ADDD) and proceed with its execution -> we would allow “out-of-order” execution! • however, out-of-order execution would allow out-of-order completion, which may allow RW (Read-Write) and WW (Write-Write) data hazards ! • a RW and WW hazard occurs when the reads/writes complete in the wrong order, destroying the data.
    [Show full text]
  • United States Patent (19) 11 Patent Number: 5,680,565 Glew Et Al
    USOO568.0565A United States Patent (19) 11 Patent Number: 5,680,565 Glew et al. 45 Date of Patent: Oct. 21, 1997 (54) METHOD AND APPARATUS FOR Diefendorff, "Organization of the Motorola 88110 Super PERFORMING PAGE TABLE WALKS INA scalar RISC Microprocessor.” IEEE Micro, Apr. 1996, pp. MCROPROCESSOR CAPABLE OF 40-63. PROCESSING SPECULATIVE Yeager, Kenneth C. "The MIPS R10000 Superscalar Micro INSTRUCTIONS processor.” IEEE Micro, Apr. 1996, pp. 28-40 Apr. 1996. 75 Inventors: Andy Glew, Hillsboro; Glenn Hinton; Smotherman et al. "Instruction Scheduling for the Motorola Haitham Akkary, both of Portland, all 88.110" Microarchitecture 1993 International Symposium, of Oreg. pp. 257-262. Circello et al., “The Motorola 68060 Microprocessor.” 73) Assignee: Intel Corporation, Santa Clara, Calif. COMPCON IEEE Comp. Soc. Int'l Conf., Spring 1993, pp. 73-78. 21 Appl. No.: 176,363 "Superscalar Microprocessor Design" by Mike Johnson, Advanced Micro Devices, Prentice Hall, 1991. 22 Filed: Dec. 30, 1993 Popescu, et al., "The Metaflow Architecture.” IEEE Micro, (51) Int. C. G06F 12/12 pp. 10-13 and 63–73, Jun. 1991. 52 U.S. Cl. ........................... 395/415: 395/800; 395/383 Primary Examiner-David K. Moore 58) Field of Search .................................. 395/400, 375, Assistant Examiner-Kevin Verbrugge 395/800, 414, 415, 421.03 Attorney, Agent, or Firm-Blakely, Sokoloff, Taylor & 56 References Cited Zafiman U.S. PATENT DOCUMENTS 57 ABSTRACT 5,136,697 8/1992 Johnson .................................. 395/586 A page table walk is performed in response to a data 5,226,126 7/1993 McFarland et al. 395/.394 translation lookaside buffer miss based on a speculative 5,230,068 7/1993 Van Dyke et al.
    [Show full text]
  • Identifying Bottlenecks in a Multithreaded Superscalar
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by KITopen Identifying Bottlenecks in a Multithreaded Sup erscalar Micropro cessor Ulrich Sigmund and Theo Ungerer VIONA DevelopmentGmbH Karlstr D Karlsruhe Germany University of Karlsruhe Dept of Computer Design and Fault Tolerance D Karlsruhe Germany Abstract This pap er presents a multithreaded sup erscalar pro cessor that p ermits several threads to issue instructions to the execution units of a wide sup erscalar pro cessor in a single cycle Instructions can simul taneously b e issued from up to threads with a total issue bandwidth of instructions p er cycle Our results show that the threaded issue pro cessor reaches a throughput of instructions p er cycle Intro duction Current micropro cessors utilize instructionlevel parallelism by a deep pro cessor pip eline and by the sup erscalar technique that issues up to four instructions p er cycle from a single thread VLSItechnology will allow future generations of micropro cessors to exploit instructionlevel parallelism up to instructions p er cycle or more However the instructionlevel parallelism found in a conventional instruction stream is limited The solution is the additional utilization of more coarsegrained parallelism The main approaches are the multipro cessor chip and the multithreaded pro ces pro cessors on a sor The multipro cessor chip integrates two or more complete single chip Therefore every unit of a pro cessor is duplicated and used indep en dently of its copies on the chip
    [Show full text]
  • Sections 3.2 and 3.3 Dynamic Scheduling – Tomasulo's Algorithm
    EEF011 Computer Architecture 計算機結構 Sections 3.2 and 3.3 Dynamic Scheduling – Tomasulo’s Algorithm 吳俊興 高雄大學資訊工程學系 October 2004 A Dynamic Algorithm: Tomasulo’s Algorithm • For IBM 360/91 (before caches!) – 3 years after CDC • Goal: High Performance without special compilers • Small number of floating point registers (4 in 360) prevented interesting compiler scheduling of operations – This led Tomasulo to try to figure out how to get more effective registers — renaming in hardware! • Why Study 1966 Computer? • The descendants of this have flourished! – Alpha 21264, HP 8000, MIPS 10000, Pentium III, PowerPC 604, … Example to eleminate WAR and WAW by register renaming • Original DIV.D F0, F2, F4 ADD.D F6, F0, F8 S.D F6, 0(R1) SUB.D F8, F10, F14 MUL.D F6, F10, F8 WAR between ADD.D and SUB.D, WAW between ADD.D and MUL.D (Due to that DIV.D needs to take much longer cycles to get F0) • Register renaming DIV.D F0, F2, F4 ADD.D S, F0, F8 S.D S, 0(R1) SUB.D T, F10, F14 MUL.D F6, F10, T Tomasulo Algorithm • Register renaming provided – by reservation stations, which buffer the operands of instructions waiting to issue – by the issue logic • Basic idea: – a reservation station fetches and buffers an operand as soon as it is available, eliminating the need to get the operand from a register (WAR) – pending instructions designate the reservation station that will provide their input (RAW) – when successive writes to a register overlap in execution, only the last one is actually used to update the register (WAW) As instructions are issued, the register specifiers for pending operands are renamed to the names of the reservation station, which provides register renaming • more reservation stations than real registers Properties of Tomasulo Algorithm 1.
    [Show full text]
  • Multiple Instruction Issue and Completion Per Clock Cycle Using Tomasulo’S Algorithm – a Simple Example
    Multiple Instruction Issue and Completion per Clock Cycle Using Tomasulo’s Algorithm – A Simple Example Assumptions: . The processor has in-order issue but execution may be out-of-order as it is done as soon after issue as operands are available. Instructions commit as they finish execu- tion. There is no speculative execution on Branch instructions because out-of-order com- pletion prevents backing out incorrect results. The reason for the out-of-order com- pletion is that there is no buffer between results from the execution and their com- mitment in the register file. This restriction is just to keep the example simple. It is possible for at least two instructions to issue in a cycle. The processor has two integer ALU’s with one Reservation Station each. ALU op- erations complete in one cycle. We are looking at the operation during a sequence of arithmetic instructions so the only Functional Units shown are the ALU’s. NOTES: Customarily the term “issue” is the transition from ID to the window for dy- namically scheduled machines and dispatch is the transition from the window to the FU’s. Since this machine has an in-order window and dispatch with no speculation, I am using the term “issue” for the transition to the reservation stations. It is possible to have multiple reservation stations in front of a single functional unit. If two instructions are ready for execution on that unit at the same clock cycle, the lowest station number executes. There is a problem with exception processing when there is no buffering of results being committed to the register file because some register entries may be from instructions past the point at which the exception occurred.
    [Show full text]
  • Advanced Processor Designs
    Advanced processor designs We’ve only scratched the surface of CPU design. Today we’ll briefly introduce some of the big ideas and big words behind modern processors by looking at two example CPUs. —The Motorola PowerPC, used in Apple computers and many embedded systems, is a good example of state-of-the-art RISC technologies. —The Intel Itanium is a more radical design intended for the higher-end systems market. April 9, 2003 ©2001-2003 Howard Huang 1 General ways to make CPUs faster You can improve the chip manufacturing technology. — Smaller CPUs can run at a higher clock rates, since electrons have to travel a shorter distance. Newer chips use a “0.13µm process,” and this will soon shrink down to 0.09µm. — Using different materials, such as copper instead of aluminum, can improve conductivity and also make things faster. You can also improve your CPU design, like we’ve been doing in CS232. — One of the most important ideas is parallel computation—executing several parts of a program at the same time. — Being able to execute more instructions at a time results in a higher instruction throughput, and a faster execution time. April 9, 2003 Advanced processor designs 2 Pipelining is parallelism We’ve already seen parallelism in detail! A pipelined processor executes many instructions at the same time. All modern processors use pipelining, because most programs will execute faster on a pipelined CPU without any programmer intervention. Today we’ll discuss some more advanced techniques to help you get the most out of your pipeline. April 9, 2003 Advanced processor designs 3 Motivation for some advanced ideas Our pipelined datapath only supports integer operations, and we assumed the ALU had about a 2ns delay.
    [Show full text]
  • Computer Architecture Out-Of-Order Execution
    Computer Architecture Out-of-order Execution By Yoav Etsion With acknowledgement to Dan Tsafrir, Avi Mendelson, Lihu Rappoport, and Adi Yoaz 1 Computer Architecture 2013– Out-of-Order Execution The need for speed: Superscalar • Remember our goal: minimize CPU Time CPU Time = duration of clock cycle × CPI × IC • So far we have learned that in order to Minimize clock cycle ⇒ add more pipe stages Minimize CPI ⇒ utilize pipeline Minimize IC ⇒ change/improve the architecture • Why not make the pipeline deeper and deeper? Beyond some point, adding more pipe stages doesn’t help, because Control/data hazards increase, and become costlier • (Recall that in a pipelined CPU, CPI=1 only w/o hazards) • So what can we do next? Reduce the CPI by utilizing ILP (instruction level parallelism) We will need to duplicate HW for this purpose… 2 Computer Architecture 2013– Out-of-Order Execution A simple superscalar CPU • Duplicates the pipeline to accommodate ILP (IPC > 1) ILP=instruction-level parallelism • Note that duplicating HW in just one pipe stage doesn’t help e.g., when having 2 ALUs, the bottleneck moves to other stages IF ID EXE MEM WB • Conclusion: Getting IPC > 1 requires to fetch/decode/exe/retire >1 instruction per clock: IF ID EXE MEM WB 3 Computer Architecture 2013– Out-of-Order Execution Example: Pentium Processor • Pentium fetches & decodes 2 instructions per cycle • Before register file read, decide on pairing Can the two instructions be executed in parallel? (yes/no) u-pipe IF ID v-pipe • Pairing decision is based… On data
    [Show full text]
  • Tomasulo's Algorithm
    Out-of-Order Execution Several implementations • out-of-order completion • CDC 6600 with scoreboarding • IBM 360/91 with Tomasulo’s algorithm & reservation stations • out-of-order completion leads to: • imprecise interrupts • WAR hazards • WAW hazards • in-order completion • MIPS R10000/R12000 & Alpha 21264/21364 with large physical register file & register renaming • Intel Pentium Pro/Pentium III with the reorder buffer Autumn 2006 CSE P548 - Tomasulo 1 Out-of-order Hardware In order to compute correct results, need to keep track of: • which instruction is in which stage of the pipeline • which registers are being used for reading/writing & by which instructions • which operands are available • which instructions have completed Each scheme has different hardware structures & different algorithms to do this Autumn 2006 CSE P548 - Tomasulo 2 1 Tomasulo’s Algorithm Tomasulo’s Algorithm (IBM 360/91) • out-of-order execution capability plus register renaming Motivation • long FP delays • only 4 FP registers • wanted common compiler for all implementations Autumn 2006 CSE P548 - Tomasulo 3 Tomasulo’s Algorithm Key features & hardware structures • reservation stations • distributed hazard detection & execution control • forwarding to eliminate RAW hazards • register renaming to eliminate WAR & WAW hazards • deciding which instruction to execute next • common data bus • dynamic memory disambiguation Autumn 2006 CSE P548 - Tomasulo 4 2 Hardware for Tomasulo’s Algorithm Autumn 2006 CSE P548 - Tomasulo 5 Tomasulo’s Algorithm: Key Features Reservation
    [Show full text]
  • CSE 586 Computer Architecture Lecture 4
    Highlights from last week CSE 586 Computer Architecture • ILP: where can the compiler optimize – Loop unrolling and software pipelining – Speculative execution (we’ll see predication today) Lecture 4 • ILP: Dynamic scheduling in a single issue machine – Scoreboard Jean-Loup Baer – Tomasulo’s algorithm http://www.cs.washington.edu/education/courses/586/00sp CSE 586 Spring 00 1 CSE 586 Spring 00 2 Highlights from last week (c’ed) -- Highlights from last week (c’ed) – Scoreboard Tomasulo’s algorithm • The scoreboard keeps a record of all data dependencies • Decentralized control • The scoreboard keeps a record of all functional unit • Use of reservation stations to buffer and/or rename occupancies registers (hence gets rid of WAW and WAR hazards) • The scoreboard decides if an instruction can be issued • Results –and their names– are broadcast to reservations • The scoreboard decides if an instruction can store its result stations and register file • Implementation-wise, scoreboard keeps track of which • Instructions are issued in order but can be dispatched, registers are used as sources and destinations and which executed and completed out-of-order functional units use them CSE 586 Spring 00 3 CSE 586 Spring 00 4 Highlights from last week (c’ed) Multiple Issue Alternatives • Register renaming: avoids WAW and WAR hazards • Superscalar (hardware detects conflicts) • Is performed at “decode” time to rename the result register – Statically scheduled (in order dispatch and hence execution; cf DEC Alpha 21164) • Two basic implementation schemes – Dynamically scheduled (in order issue, out of order dispatch and – Have a separate physical register file execution; cf MIPS 10000, IBM Power PC 620 and Intel Pentium – Use of reorder buffer and reservation stations (cf.
    [Show full text]
  • Tomasulo's Algorithm
    Lecture-12 (Tomasulo’s Algorithm) CS422-Spring 2018 Biswa@CSE-IITK Another Dynamic One: Tomasulo’s Algorithm • For IBM 360/91 about 3 years after CDC 6600 (1966) • Goal: High Performance without special compilers • Differences between IBM 360 & CDC 6600 ISA – IBM has only 2 register specifiers/instruction vs. 3 in CDC 6600 – IBM has 4 FP registers vs. 8 in CDC 6600 – IBM has memory-register ops • Why Study? lead to Alpha 21264, HP 8000, MIPS 10000, Pentium II, PowerPC 604, … CS422: Spring 2018 Biswabandan Panda, CSE@IITK 2 Tomasulo’s Organization From Mem FP Op FP Registers Queue Load Buffers Load1 Load2 Load3 Load4 Load5 Store Load6 Buffers Add1 Add2 Mult1 Add3 Mult2 Reservation To Mem Stations FP adders FP multipliers CS422: Spring 2018 Biswabandan Panda, CSE@IITK 3 Tomasulo vs Scoreboard • Control & buffers distributed with Function Units (FU) vs. centralized in scoreboard; – FU buffers called “reservation stations”; have pending operands • Registers in instructions replaced by values or pointers to reservation stations(RS); called register renaming ; – avoids WAR, WAW hazards – More reservation stations than registers, so can do optimizations compilers can’t • Results to FU from RS, not through registers, over Common Data Bus that broadcasts results to all FUs • Load and Stores treated as FUs with RSs as wells CS422: Spring 2018 Biswabandan Panda, CSE@IITK 4 Reservation Station Components Op: Operation to perform in the unit (e.g., + or –) Vj, Vk: Value of Source operands – Store buffers has V field, result to be stored Qj, Qk: Reservation stations producing source registers (value to be written) – Note: No ready flags as in Scoreboard; Qj,Qk=0 => ready – Store buffers only have Qi for RS producing result Busy: Indicates reservation station or FU is busy Register result status—Indicates which functional unit will write each register, if one exists.
    [Show full text]