COMP 206: Computer Architecture and Implementation

1
COMP 206Computer Architecture and Implementation

• Montek Singh
• Wed., Sep 18, 2002
• Topic Pipelining -- Intermediate Concepts
• (Multicycle Operations Exceptions)

2
Pipelining Multicycle Operations

• Assume five-stage pipeline
• Third stage (execution) has two functional units
  E1 and E2
• Instruction goes through either E1 or E2, but not
  both
• E1 and E2 are not pipelined
• Stage delay of E1 2 cycles
• Stage delay of E2 4 cycles
• No buffering on inputs of E1 and E2
• Stage delay of other stages 1 cycle
• Consider an instruction sequence of five
  instructions
• Instructions 1, 3, 5 need E1
• Instructions 2, 4 need E2

3
Space-Time Diagram Multicycle Operations

• Out-of-order completion
• 3 finishes before 2, and 5 finishes before 4
• Instructions may be delayed after entering the
  pipeline because of structural hazards
• Instructions 2 and 4 both want to use E2 unit at
  same time
• Instruction 4 stalls in ID unit
• This causes instruction 5 to stall in IF unit

4
Floating-Point Operations in MIPS
IF
ID
EX
Out-of-order completion has ramifications
for exceptions
WAW hazards possible WAR hazards not possible
Longer operation latency implies more
frequent stalls for RAW hazards
MEM
Structural hazard instructions have varying
running times
WB
Structural hazard not fully pipelined
5
Structural Hazard on WB Unit

• This is worst-case scenario max steady-state
  number of write ports is 1
• Dont replicate resources detect and serialize
  access as needed
• Early resolution
• Track use of WB in ID stage (using shift
  register), stall instructions there
• reservation register
• Simplifies pipeline control all stalls occur in
  ID
• adds shift register and write-conflict logic
• Late resolution
• Stall instructions at entry to MEM or WB stage
• Complicates pipeline control (two stall locations)

6
WAW Hazards

• WAW hazard arises only when no instruction
  between ADD.D and L.D uses result computed by
  ADD.D
• Adding an instruction like ADD.D F8,F2,F4
  before L.D would stall pipeline enough for RAW
  hazard to avoid WAW hazard
• Can happen through a branch/trap (example in HP3,
  Section A.9)
• Rare situation, but must still handle correctly
• Hazard resolution
• Delay the issue of L.D until ADD.D enters MEM
• Cancel write of ADD.D

7
RAW Hazards

• Longer delays of FP operations increases number
  of stalls in response to RAW hazards
• Two methods for reducing stalls
• Compiler could have moved instruction D between
  instructions M and A, which would allow D to
  complete earlier or hardware could detect this
  possibility and issue instruction D out of order
• ID stage is a bottleneck because instructions
  wait their for their operands to be available
  could add buffers (reservation stations) to
  functional units and let instructions await their
  operands there

8
Responsibilities of ID (all stalls in ID)

• Three sets of checks
• Structural hazards
• Check for availability of FP unit
• Ensure WB unit will be available when needed
• RAW hazards
• Stall current instruction until its source
  registers are not listed as pending registers in
  a pipeline register that will not be available
  when current instruction needs the result
• WAW hazards
• If any instruction in adder, divider, or
  multiplier has same register destination as
  current instruction, stall current instruction
• Hazards between FP and integer instructions
• Integer and FP instructions use disjoint sets of
  registers, except for FP-integer register moves
• FP load-stores can conflict with integer
  load-stores in MEM stage

9
MIPS R4000 Floating-Point Pipeline
10
Instruction Mixes in FP Pipeline Adds Only
Cant initiate another add on cycle 2 Conflict
here

• Forbidden latencies 1 and 2
• Steady-state utilization (cycles 4 through 18)
• (57)/(815) 35/120 29.17
• Total utilization (cycles 1 through 19)
• (5572)/(819) 42/152 27.63

Cant initiate another add on cycle 3 Conflict
here
11
FP Pipeline Multiplies Only

• Collision vector
• 1 indicates forbidden latency
• 0 indicates allowed latency
• Steady-state utilization (cycles 5-24)
• (510)/(820) 50/160 31.25
• Total utilization (cycles 1-28)
• (55105)/(828) 60/224 26.79

Multiply
12
FP Pipeline Adds and Multiplies

• Note out-of-order
• completion
• Steady-state utilization
• (cycles 6-21)
• (417)/(816) 68/128
• 53.13
• Total utilization
• (1241722)/(828)
• 85/224 37.95

13
Interrupts, Faults, or Exceptions

• Synchronous, coerced interrupts that occur within
  instructions and after which execution must
  resume are the hardest to implement
• See Figure A.27 in HP3

14
Precise Interrupts (Sequential Processor)

• When interrupt occurs, state of interrupted
  process is saved, including PC ( u), registers,
  and memory
• Interrupt is precise if the following three
  conditions hold
• All instructions preceding u have been executed,
  and have modified the state correctly
• All instructions following u are unexecuted, and
  have not modified the state
• If the interrupt was caused by an instruction, it
  was caused by instruction u, which is either
  completely executed (overflow) or completely
  unexecuted (VM page fault)
• Precise interrupts are desirable if software is
  to fix up error that caused interrupt and
  execution has to be resumed
• Easy for external interrupts, could be complex
  and costly for internal
• Imperative for some interrupts (VM page faults,
  IEEE FP standard)

15
Problems on Sequential Processors

• Long-running instructions
• Not enough to be able to restore state, must make
  progress from interrupt to interrupt
• Example MVC on IBM 360 copies 256 bytes
• No virtual memory, so interrupts not allowed to
  stop MVC
• Example MVC on IBM 370 copies 256 bytes
• Has virtual memory, so first access all pages
  involved after that, no interrupts allowed
• Example MVCL on IBM 370 copies up to 224 bytes
• Has VM two addresses and length are in registers
• Registers saved and restored on interrupts
  (making progress)

• Instruction modifies state early, then causes an
  interrupt
• State change must be undone
• Example First operand of VAX instruction uses
  autodecrement addressing mode, which writes a
  register. Trying to access second operand causes
  a page fault. Since instruction execution cannot
  be completed, we must restore the register
  written by autodecrement to its original value

16
Interrupts in MIPS Pipeline

• How do we stop and restart execution on an
  interrupt to keep it precise?
• What problems do delayed branches cause?
• What happens if multiple exceptions occur in the
  pipeline?
• Can exceptions occur out-of-order?
• What problems do multi-cycle instructions cause?

17
MIPS Integer Pipeline, Single Interrupt

• Force TRAP instruction in pipeline on next IF
• Turn off all writes for faulting instruction and
  subsequent instructions
• After exception-handling routine in OS receives
  control, save PC of faulting instruction
• When exception has been handled, the RFE
  instruction reloads PC and restarts sequential
  instruction execution

18
Complications with Delayed Branches

• Suppose instruction 2 causes an exception (e.g.,
  a page fault) after the taken branch completes
  (determining that the branch outcome is true)
• Instruction 2 cannot complete
• Neither can instruction u
• On restart, we do not have sequential execution
• We must remember two PC values 2 and u

19
Complications with Multiple Exceptions

• At same cycle, LW takes a data page fault and ADD
  takes an arithmetic exception
• On an unpipelined machine, LWs exception would
  occur first
• Handle the page fault
• Restart execution
• ADD will cause arithmetic exception to reoccur
  handle it then

20
Complications with Out-of-order Exceptions

• LW takes data page fault, ADD takes instruction
  page fault
• Relative timing differs between unpipelined and
  pipelined machines
• To maintain precise interrupts, we need to
  consider both when they occur and the
  instructions that caused them
• Post exceptions in exception status vector, turn
  off state modifications, and check vector in WB
  unit

21
Complications with Multicycle Operations

• Instructions are independent (no hazards) and
  therefore issue immediately
• Differences in running times causes out-of-order
  termination
• DIVF throws arithmetic exception late in its
  execution
• At that point, ADDF and SUBF have both completed
  execution and destroyed one of their operands
• Can we maintain precise interrupts under these
  conditions?

22
FP Pipeline Exceptions Solns. 1 and 2

• Settle for imprecise interrupts (CRAY, with
  checkpointing)
• Done on Alpha 21064 and 21164, IBM Power-1 and
  Power-2, MIPS R8000 by supporting a fast
  imprecise mode and a slow precise mode
• Not an option if you have to support virtual
  memory or IEEE floating point standard
• Software finishes certain instructions (SPARC)
• Keep enough state around for trap handler to
  create a precise sequence for exception and
  finish work for some instruction stages
• Only FP instructions cause this problem

23
FP Pipeline Exceptions Solns. 3 and 4

• Stalling (MIPS R2000/3000, MIPS R4000, Pentium)
• An instruction is allowed to issue only if it is
  certain that all the instructions before the
  issuing instruction will complete without causing
  an exception
• To prevent excessive stalling, FP units must
  decide on possibility of exceptions early in
  pipeline
• General methods (PowerPC 620, MIPS R10000)
• Reorder buffer, history file, future file
• An instruction is allowed to finalize its writes
  only when all previously issued instructions are
  complete
• More naturally used in connection with ILP
  (Chapter 4)
• Significant complexity (to be discussed later)