1. Exploring Branch Prediction
Aditya Akella
Shuchi Chawla
Jia-Yu Pan

2. A Framework for Branch Prediction

3. Dividing Streams First level: per branch substream Finer decomposition
Branch pattern history
Global pattern history
Global path history
1011…
1110…
(b1,1)(b2,1)(b2,1)…
Loop1:
addq r1, r2, r3
addq r2, r4, r5
loop 2:
subq r2,r3,r5
beq r5, loop2
bne r5, loop1
Executes twice

4. Path vs Pattern
As expected, path history provides better correlation

5. Predictors Static vs Dynamic Adapt to changing bias Biased stream:
Majority prediction
Adapt to changing bias
Damping: 2 (or more) – bit predictor
Is more better?
Static is better on a large number of branches
But, dynamic far outperforms static on the rest
Adaptive schemes should exploit this

6. Implementation Issues: Limited space
Many branches map to same stream: Aliasing
Mostly a destructive effect
More aggressive schemes may not be necessarily better

7. gshare GAs  6 bits branch identifier 6 bits history
concatenate
6 bits history
12 bits stream identifier
12 bits branch identifier
12 bit history 
12 bits stream identifier
truncate to 1024 streams

8. Implementation Issues: Limited space
Many branches map to same stream: Aliasing
Mostly a destructive effect
More aggressive schemes may not be necessarily better

9. Cross-procedure Correlation
Some static prediction schemes “forget” history on returning from a procedure call
Fun_call:
if (x>0) return(1);
else return(0);
Main:
a = fun_call();
if (a>0) goto…
Direction taken by the branch inside the procedure completely specifies the direction taken by the branch outside

10. Branch Prediction Using Data Values
Traditional approach
Local history information
Global path and history information
Reducing table interference (better indexing)
Use PCs and branch outcomes as input
Do not contain all the information
Misprediction!!

11. How can one improve prediction of such branches?
Use Data Values!!!!!!!!!!!!

12. How to Use Data Values Speculative branch execution
Second scheme was chosen
Lower-latency prediction
Some branches can use combined predictions
Solution avoids data value prediction
Using data values directly

13. Design Problems… (1) Large number of data values to store…
Soln – store their difference
Branch Difference Predictor (BDP)
Backing Predictor predicts most cases
REP predicts the small remaining subset
Fringe cases with exceptional outcomes
VHT provides ‘tag-check’.

14. Design Problems… (2) Delay in updating data values
Out-of-order execution, pipeline latencies
BDC -- most recently committed branch difference (indexed by branch PC)
BCT -- count of the outstanding instances
Staleness of the data value
Operation
Branch Fetch : Increment counter
Branch Commit : Reset counter and update difference cache
REP replacement policy
Replace least successful entries
Randomly ignore a chunk of mispredictions

15. Search for a good predictor design
Paper:
“A Language for Describing Predictors and its Application to Automatic Synthesis”

16. Designing a predictor Components of a predictor
Counters and History
More?
Formulate as a search problem
How?

17. Search the design space
Representation
Feedback Loop Model + Primitives
P[w,d](I;U) => parse tree
Feedback/
Update (U) d
Input
Index (I)
Prediction (P) w

18. Examples Onebit[d](PC;T) Counter[n,d](I;T) Twobit[d](PC;T)
=P[1,d](PC; T)
Counter[n,d](I;T)
= P[n,d](I; if T then P+1 else P-1)
Twobit[d](PC;T)
=MSB(Counter[2,d](PC;T))

19. Genetic Programming Process Operations on parse trees
Population, Evaluation, Selection and populate next generation
Operations on parse trees
Replication, Crossover, Mutation, Encapsulation

20. Discussion Comparable good prediction accuracy
Complex but new components
Rediscover counters and local/global histories
Predictors for other purposes, e.g. indirect jump target prediction

21. End
Comments?
Thank you.