View
226
Download
2
Tags:
Embed Size (px)
Citation preview
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.1
CS152Computer Architecture and Engineering
Lecture 21
Memory Systems (recap)Caches
April 21, 2003
John Kubiatowicz (www.cs.berkeley.edu/~kubitron)
lecture slides: http://inst.eecs.berkeley.edu/~cs152/
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.2
° The Five Classic Components of a Computer
° Today’s Topics: • Recap last lecture
• Simple caching techniques
• Many ways to improve cache performance
• Virtual memory?
Recap: The Big Picture: Where are We Now?
Control
Datapath
Memory
Processor
Input
Output
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.3
µProc60%/yr.(2X/1.5yr)
DRAM9%/yr.(2X/10 yrs)1
10
100
1000
198
0198
1 198
3198
4198
5 198
6198
7198
8198
9199
0199
1 199
2199
3199
4199
5199
6199
7199
8 199
9200
0
DRAM
CPU198
2
Processor-MemoryPerformance Gap:(grows 50% / year)
Per
form
ance
Time
“Moore’s Law”
Processor-DRAM Memory Gap (latency)
Recap: Who Cares About the Memory Hierarchy?
“Less’ Law?”
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.4
Recap: Memory Hierarchy: Why Does it Work? Locality!
° Temporal Locality (Locality in Time):=> Keep most recently accessed data items closer to the processor
° Spatial Locality (Locality in Space):=> Move blocks consists of contiguous words to the upper levels
Lower LevelMemoryUpper Level
MemoryTo Processor
From ProcessorBlk X
Blk Y
Address Space0 2^n - 1
Probabilityof reference
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.5
Recap: Static RAM Cell
6-Transistor SRAM Cell
bit bit
word(row select)
bit bit
word
° Write:1. Drive bit lines (bit=1, bit=0)
2.. Select row
° Read:1. Precharge bit and bit to Vdd or Vdd/2 => make sure equal!
2.. Select row
3. Cell pulls one line low
4. Sense amp on column detects difference between bit and bit
replaced with pullupto save area
10
0 1
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.6
Recap: 1-Transistor Memory Cell (DRAM)
° Write:• 1. Drive bit line
• 2.. Select row
° Read:• 1. Precharge bit line to Vdd/2
• 2.. Select row
• 3. Cell and bit line share charges
- Very small voltage changes on the bit line
• 4. Sense (fancy sense amp)
- Can detect changes of ~1 million electrons
• 5. Write: restore the value
° Refresh• 1. Just do a dummy read to every cell.
row select
bit
Trench Capacitor
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.7
Recap: Classical DRAM Organization (square)
° Row and Column Address Select 1 bit at a time° Act of reading refreshes one complete row
• Sense amps detect slight variations from VDD/2 and amplify them
row
decoder
rowaddress
Sense-AMPS, Column Selector & I/O Column
Address
data
RAM Cell Array
word (row) select
bit (data) lines
Each intersection representsa 1-T DRAM Cell
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.8
AD
OE_L
256K x 8DRAM9 8
WE_LCAS_LRAS_L
OE_L
A Row Address
WE_L
Junk
Read AccessTime
Output EnableDelay
CAS_L
RAS_L
Col Address Row Address JunkCol Address
D High Z Data Out
DRAM Read Cycle Time
Early Read Cycle: OE_L asserted before CAS_L Late Read Cycle: OE_L asserted after CAS_L
° Every DRAM access begins at:
• The assertion of the RAS_L
• 2 ways to read: early or late v. CAS
Junk Data Out High Z
Recap: Traditional “asynchronous” DRAM Read Timing
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.9
Recap: “Synchronous timing”: SDRAM timing for Lab6
° Micron 128M-bit dram (using 2Meg16bit4bank ver)• Row (12 bits), bank (2 bits), column (9 bits)
RAS(New Bank)
CAS End RASx
BurstREADCAS Latency
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.10
Processor
$
MEM
Memory
reference stream <op,addr>, <op,addr>,<op,addr>,<op,addr>, . . .
op: i-fetch, read, write
Optimize the memory system organizationto minimize the average memory access timefor typical workloads
Workload orBenchmarkprograms
The Art of Memory System Design
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.11
Impact of Memory Hierarchy on Algorithms
° Today CPU time is a function of (ops, cache misses)
° What does this mean to Compilers, Data structures, Algorithms?
• Quicksort: fastest comparison based sorting algorithm when keys fit in
memory
• Radix sort: also called “linear time” sortFor keys of fixed length and fixed radix a constant number of
passes over the data is sufficient independent of the number of keys
° “The Influence of Caches on the Performance of Sorting” by A. LaMarca and R.E. Ladner. Proceedings of the Eighth Annual ACM-SIAM Symposium on Discrete Algorithms, January, 1997, 370-379.
• For Alphastation 250, 32 byte blocks, direct mapped L2 2MB cache, 8 byte keys, from 4000 to 4000000
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.12
Quicksort vs. Radix as vary number keys: Instructions
Job size in keys
Instructions/key
Radix sort
Quicksort
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.13
Quicksort vs. Radix as vary number keys: Instrs & Time
Time
Job size in keys
Instructions
Radix sort
Quicksort
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.14
Quicksort vs. Radix as vary number keys: Cache misses
Cache misses
Job size in keys
Radix sort
Quicksort
What is proper approach to fast algorithms?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.15
Example: 1 KB Direct Mapped Cache with 32 B Blocks° For a 2 ** N byte cache:
• The uppermost (32 - N) bits are always the Cache Tag
• The lowest M bits are the Byte Select (Block Size = 2M)
• One cache miss, pull in complete “Cache Block” (or “Cache Line”)
Cache Index
0
1
2
3
:
Cache Data
Byte 0
0431
:
Cache Tag Example: 0x50
Ex: 0x01
0x50
Stored as partof the cache “state”
Valid Bit
:31
Byte 1Byte 31 :
Byte 32Byte 33Byte 63 :Byte 992Byte 1023 :
Cache Tag
Byte Select
Ex: 0x00
9Block address
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.16
Set Associative Cache° N-way set associative: N entries for each Cache Index
• N direct mapped caches operates in parallel
° Example: Two-way set associative cache• Cache Index selects a “set” from the cache
• The two tags in the set are compared to the input in parallel
• Data is selected based on the tag result
Cache Data
Cache Block 0
Cache TagValid
:: :
Cache Data
Cache Block 0
Cache Tag Valid
: ::
Cache Index
Mux 01Sel1 Sel0
Cache Block
CompareAdr Tag
Compare
OR
Hit
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.17
Disadvantage of Set Associative Cache
° N-way Set Associative Cache versus Direct Mapped Cache:
• N comparators vs. 1• Extra MUX delay for the data• Data comes AFTER Hit/Miss decision and set selection
° In a direct mapped cache, Cache Block is available BEFORE Hit/Miss:
• Possible to assume a hit and continue. Recover later if miss.
Cache Data
Cache Block 0
Cache Tag Valid
: ::
Cache Data
Cache Block 0
Cache TagValid
:: :
Cache Index
Mux 01Sel1 Sel0
Cache Block
CompareAdr Tag
Compare
OR
Hit
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.18
Example: Fully Associative
° Fully Associative Cache• Forget about the Cache Index
• Compare the Cache Tags of all cache entries in parallel
• Example: Block Size = 32 B blocks, we need N 27-bit comparators
° By definition: Conflict Miss = 0 for a fully associative cache
:
Cache Data
Byte 0
0431
:
Cache Tag (27 bits long)
Valid Bit
:
Byte 1Byte 31 :
Byte 32Byte 33Byte 63 :
Cache Tag
Byte Select
Ex: 0x01
=
=
=
=
=
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.19
° Should be reading Chapter 7 of your book ° Second midterm: Monday May 5th
• Pipelining - Hazards, branches, forwarding, CPI calculations- (may include something on dynamic scheduling)
• Memory Hierarchy• Possibly something on I/O (see where we get in lectures)• Possibly something on power (Broderson Lecture)
° Lab 6: Hopefully all is well!• How about a “fake processor” written in Verilog to debug memory
system?
• Make sure to try writes to individual words and reads from adjacent words in the cache
• Careful about 2-clocks for DRAM module
- I have sent out a couple of messages about this: you should be able to single-step your processor while DRAM continues to run at full speed….
Administrative Issues
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.20
° Compulsory (cold start or process migration, first reference): first access to a block
• “Cold” fact of life: not a whole lot you can do about it
• Note: If you are going to run “billions” of instruction, Compulsory Misses are insignificant
° Capacity:• Cache cannot contain all blocks access by the program
• Solution: increase cache size
° Conflict (collision):• Multiple memory locations mapped
to the same cache location
• Solution 1: increase cache size
• Solution 2: increase associativity
° Coherence (Invalidation): other process (e.g., I/O) updates memory
A Summary on Sources of Cache Misses
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.21
Design options at constant cost
Direct Mapped N-way Set Associative Fully Associative
Compulsory Miss
Cache Size
Capacity Miss
Coherence Miss
Big Medium Small
Note:If you are going to run “billions” of instruction, Compulsory Misses are insignificant (except for streaming media types of programs).
Same Same Same
Conflict Miss High Medium Zero
Low Medium High
Same Same Same
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.22
° Q1: Where can a block be placed in the upper level? (Block placement)
° Q2: How is a block found if it is in the upper level? (Block identification)
° Q3: Which block should be replaced on a miss? (Block replacement)
° Q4: What happens on a write? (Write strategy)
Recap: Four Questions for Caches and Memory Hierarchy
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.23
° Block 12 placed in 8 block cache:• Fully associative, direct mapped, 2-way set associative• S.A. Mapping = Block Number Modulo Number Sets
0 1 2 3 4 5 6 7Blockno.
Fully associative:block 12 can go anywhere
0 1 2 3 4 5 6 7Blockno.
Direct mapped:block 12 can go only into block 4 (12 mod 8)
0 1 2 3 4 5 6 7Blockno.
Set associative:block 12 can go anywhere in set 0 (12 mod 4)
Set0
Set1
Set2
Set3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Block-frame address
1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3Blockno.
Q1: Where can a block be placed in the upper level?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.24
° Direct indexing (using index and block offset), tag compares, or combination
° Increasing associativity shrinks index, expands tag
Blockoffset
Block AddressTag Index
Q2: How is a block found if it is in the upper level?
Set Select
Data Select
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.25
° Easy for Direct Mapped
° Set Associative or Fully Associative:• Random
• LRU (Least Recently Used)
Associativity: 2-way 4-way 8-way
Size LRU Random LRU Random LRU Random
16 KB 5.2% 5.7% 4.7% 5.3% 4.4% 5.0%
64 KB 1.9% 2.0% 1.5% 1.7% 1.4% 1.5%
256 KB 1.15% 1.17% 1.13% 1.13% 1.12% 1.12%
Q3: Which block should be replaced on a miss?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.26
° Write through—The information is written to both the block in the cache and to the block in the lower-level memory.
° Write back—The information is written only to the block in the cache. The modified cache block is written to main memory only when it is replaced.
• is block clean or dirty?
° Pros and Cons of each?• WT: read misses cannot result in writes
• WB: no writes of repeated writes
° WT always combined with write buffers so that don’t wait for lower level memory
Q4: What happens on a write?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.27
° A Write Buffer is needed between the Cache and Memory
• Processor: writes data into the cache and the write buffer
• Memory controller: write contents of the buffer to memory
° Write buffer is just a FIFO:• Typical number of entries: 4
• Must handle bursts of writes
• Works fine if: Store frequency (w.r.t. time) << 1 / DRAM write cycle
ProcessorCache
Write Buffer
DRAM
Write Buffer for Write Through
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.28
Write Buffer Saturation
° Store frequency (w.r.t. time) > 1 / DRAM write cycle• If this condition exist for a long period of time (CPU cycle time too
quick and/or too many store instructions in a row):
- Store buffer will overflow no matter how big you make it
- The CPU Cycle Time <= DRAM Write Cycle Time
° Solution for write buffer saturation:• Use a write back cache
• Install a second level (L2) cache: (does this always work?)
ProcessorCache
Write Buffer
DRAM
ProcessorCache
Write Buffer
DRAML2Cache
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.29
° Write-Buffer Issues: Could introduce RAW Hazard with memory!
• Write buffer may contain only copy of valid data Reads to memory may get wrong result if we ignore write buffer
° Solutions:• Simply wait for write buffer to empty before servicing reads:
- Might increase read miss penalty (old MIPS 1000 by 50% )• Check write buffer contents before read (“fully associative”);
- If no conflicts, let the memory access continue- Else grab data from buffer
° Can Write Buffer help with Write Back?• Read miss replacing dirty block
- Copy dirty block to write buffer while starting read to memory• CPU stall less since restarts as soon as do read
RAW Hazards from Write Buffer!
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.30
° Assume: a 16-bit write to memory location 0x0 and causes a miss
• Do we allocate space in cache and possibly read in the block?
- Yes: Write Allocate
- No: Not Write Allocate
Cache Index
0
1
2
3
:
Cache Data
Byte 0
0431
:
Cache Tag Example: 0x00
Ex: 0x00
0x50
Valid Bit
:
31
Byte 1Byte 31 :
Byte 32Byte 33Byte 63 :Byte 992Byte 1023 :
Cache Tag
Byte Select
Ex: 0x00
9
Write-miss Policy: Write Allocate versus Not Allocate
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.31
° Set of Operations that must be supported• read: data <= Mem[Physical Address]• write: Mem[Physical Address] <= Data
° Determine the internal register transfers° Design the Datapath° Design the Cache Controller
Physical Address
Read/Write
Data
Memory“Black Box”
Inside it has:Tag-Data Storage,Muxes,Comparators, . . .
CacheController
CacheDataPathAddress
Data In
Data Out
R/WActive
ControlPoints
Signalswait
How Do you Design a Memory System?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.32
Impact on Cycle Time
IR
PCI -Cache
D Cache
A B
R
T
IRex
IRm
IRwb
miss
invalid
Miss
Cache Hit Time:directly tied to clock rateincreases with cache sizeincreases with associativity
Average Memory Access time = Hit Time + Miss Rate x Miss Penalty
Time = IC x CT x (ideal CPI + memory stalls)
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.33
Options to reduce AMAT:
1. Reduce the miss rate,
2. Reduce the miss penalty, or
3. Reduce the time to hit in the cache.
Time = IC x CT x (ideal CPI + memory stalls)
Average Memory Access time = Hit Time + (Miss Rate x Miss Penalty) =
(Hit Rate x Hit Time) + (Miss Rate x Miss Time)
Improving Cache Performance: 3 general options
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.34
1. Reduce the miss rate,
2. Reduce the miss penalty, or
3. Reduce the time to hit in the cache.
Improving Cache Performance
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.35
Cache Size (KB)
Mis
s R
ate
per
Typ
e
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
1 2 4 8
16
32
64
12
8
1-way
2-way
4-way
8-way
Capacity
Compulsory
Conflict
3Cs Absolute Miss Rate (SPEC92)
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.36
Cache Size (KB)
Mis
s R
ate
per
Typ
e
0
0.02
0.04
0.06
0.08
0.1
0.12
0.141 2 4 8
16
32
64
12
8
1-way
2-way
4-way
8-way
Capacity
Compulsory
Conflict
miss rate 1-way associative cache size X = miss rate 2-way associative cache size X/2
2:1 Cache Rule
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.37
Cache Size (KB)
Mis
s R
ate
per
Typ
e
0%
20%
40%
60%
80%
100%
1 2 4 8
16
32
64
12
8
1-way
2-way4-way
8-way
Capacity
Compulsory
Conflict
3Cs Relative Miss Rate
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.38
Block Size (bytes)
Miss Rate
0%
5%
10%
15%
20%
25%
16
32
64
12
8
25
6
1K
4K
16K
64K
256K
1. Reduce Misses via Larger Block Size
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.39
° 2:1 Cache Rule: • Miss Rate DM cache size N Miss Rate 2-way cache size N/2
° Beware: Execution time is only final measure!• Will Clock Cycle time increase?
• Hill [1988] suggested hit time for 2-way vs. 1-way external cache +10%, internal + 2%
2. Reduce Misses via Higher Associativity
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.40
° Assume CCT = 1.10 for 2-way, 1.12 for 4-way, 1.14 for 8-way vs. CCT direct mapped
Cache Size Associativity
(KB) 1-way 2-way 4-way 8-way
1 2.33 2.15 2.07 2.01
2 1.98 1.86 1.76 1.68
4 1.72 1.67 1.61 1.53
8 1.46 1.48 1.47 1.43
16 1.29 1.32 1.32 1.32
32 1.20 1.24 1.25 1.27
64 1.14 1.20 1.21 1.23
128 1.10 1.17 1.18 1.20
(Red means A.M.A.T. not improved by more associativity)
Example: Avg. Memory Access Time vs. Miss Rate
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.41
To Next Lower Level InHierarchy
DATATAGS
One Cache line of DataTag and Comparator
One Cache line of DataTag and Comparator
One Cache line of DataTag and Comparator
One Cache line of DataTag and Comparator
3. Reducing Misses via a “Victim Cache”
° How to combine fast hit time of direct mapped yet still avoid conflict misses?
° Add buffer to place data discarded from cache
° Jouppi [1990]: 4-entry victim cache removed 20% to 95% of conflicts for a 4 KB direct mapped data cache
° Used in Alpha, HP machines
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.42
° E.g., Instruction Prefetching• Alpha 21064 fetches 2 blocks on a miss
• Extra block placed in “stream buffer”
• On miss check stream buffer
° Works with data blocks too:• Jouppi [1990] 1 data stream buffer got 25% misses from 4KB cache; 4
streams got 43%
• Palacharla & Kessler [1994] for scientific programs for 8 streams got 50% to 70% of misses from 2 64KB, 4-way set associative caches
° Prefetching relies on having extra memory bandwidth that can be used without penalty
• Could reduce performance if done indiscriminantly!!!
4. Reducing Misses by Hardware Prefetching
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.43
° Data Prefetch• Load data into register (HP PA-RISC loads)
• Cache Prefetch: load into cache (MIPS IV, PowerPC, SPARC v. 9)
• Special prefetching instructions cannot cause faults;a form of speculative execution
° Issuing Prefetch Instructions takes time• Is cost of prefetch issues < savings in reduced misses?
• Higher superscalar reduces difficulty of issue bandwidth
5. Reducing Misses by Software Prefetching Data
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.44
° McFarling [1989] reduced caches misses by 75% on 8KB direct mapped cache, 4 byte blocks in software
° Instructions• Reorder procedures in memory so as to reduce conflict misses
• Profiling to look at conflicts(using tools they developed)
° Data• Merging Arrays: improve spatial locality by single array of compound elements
vs. 2 arrays
• Loop Interchange: change nesting of loops to access data in order stored in memory
• Loop Fusion: Combine 2 independent loops that have same looping and some variables overlap
• Blocking: Improve temporal locality by accessing “blocks” of data repeatedly vs. going down whole columns or rows
6. Reducing Misses by Compiler Optimizations
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.45
1. Reduce the miss rate,
2. Reduce the miss penalty, or
3. Reduce the time to hit in the cache.
Improving Cache Performance (Continued)
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.46
0. Reducing Penalty: Faster DRAM / Interface
° New DRAM Technologies • RAMBUS - same initial latency, but much higher bandwidth
• Synchronous DRAM
• TMJ-RAM (Tunneling magnetic-junction RAM) from IBM??
• Merged DRAM/Logic - IRAM project here at Berkeley
° Better BUS interfaces
° CRAY Technique: only use SRAM
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.47
° A Write Buffer is needed between the Cache and Memory• Processor: writes data into the cache and the write buffer
• Memory controller: write contents of the buffer to memory
° Write buffer is just a FIFO:• Typical number of entries: 4
• Works fine if:Store frequency (w.r.t. time) << 1 / DRAM write cycle
• Must handle burst behavior as well!
ProcessorCache
Write Buffer
DRAM
1. Reducing Penalty: Read Priority over Write on Miss
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.48
° Write-Buffer Issues: Could introduce RAW Hazard with memory!• Write buffer may contain only copy of valid data
Reads to memory may get wrong result if we ignore write buffer
° Solutions:• Simply wait for write buffer to empty before servicing reads:
- Might increase read miss penalty (old MIPS 1000 by 50% )• Check write buffer contents before read (“fully associative”);
- If no conflicts, let the memory access continue- Else grab data from buffer
° Can Write Buffer help with Write Back?• Read miss replacing dirty block
- Copy dirty block to write buffer while starting read to memory• CPU stall less since restarts as soon as do read
RAW Hazards from Write Buffer!
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.49
° Don’t wait for full block to be loaded before restarting CPU
• Early restart—As soon as the requested word of the block arrives, send it to the CPU and let the CPU continue execution
• Critical Word First—Request the missed word first from memory and send it to the CPU as soon as it arrives; let the CPU continue execution while filling the rest of the words in the block. Also called wrapped fetch and requested word first
• DRAM FOR LAB 6 can do this in burst mode! (Check out sequential timing)
° Generally useful only in large blocks,
° Spatial locality a problem; tend to want next sequential word, so not clear if benefit by early restart
block
2. Reduce Penalty: Early Restart and Critical Word First
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.50
° Non-blocking cache or lockup-free cache allow data cache to continue to supply cache hits during a miss
• requires F/E bits on registers or out-of-order execution• requires multi-bank memories
° “hit under miss” reduces the effective miss penalty by working during miss vs. ignoring CPU requests
° “hit under multiple miss” or “miss under miss” may further lower the effective miss penalty by overlapping multiple misses
• Significantly increases the complexity of the cache controller as there can be multiple outstanding memory accesses
• Requires muliple memory banks (otherwise cannot support)• Penium Pro allows 4 outstanding memory misses
3. Reduce Penalty: Non-blocking Caches
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.51
° For in-order pipeline, 2 options:• Freeze pipeline in Mem stage (popular early on: Sparc, R4000)
IF ID EX Mem stall stall stall … stall Mem Wr IF ID EX stall stall stall … stall stall Ex Wr
• Use Full/Empty bits in registers + MSHR queue
- MSHR = “Miss Status/Handler Registers” (Kroft)Each entry in this queue keeps track of status of outstanding memory requests to one complete memory line.– Per cache-line: keep info about memory address.– For each word: register (if any) that is waiting for result.– Used to “merge” multiple requests to one memory line
- New load creates MSHR entry and sets destination register to “Empty”. Load is “released” from pipeline.
- Attempt to use register before result returns causes instruction to block in decode stage.
- Limited “out-of-order” execution with respect to loads. Popular with in-order superscalar architectures.
° Out-of-order pipelines already have this functionality built in… (load queues, etc).
What happens on a Cache miss?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.52
° FP programs on average: AMAT= 0.68 -> 0.52 -> 0.34 -> 0.26
° Int programs on average: AMAT= 0.24 -> 0.20 -> 0.19 -> 0.19
° 8 KB Data Cache, Direct Mapped, 32B block, 16 cycle miss
Hit Under i Misses
Av
g.
Me
m.
Acce
ss T
ime
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
eqntott
espresso
xlisp
compress
mdljsp2
ear
fpppp
tomcatv
swm256
doduc
su2cor
wave5
mdljdp2
hydro2d
alvinn
nasa7
spice2g6
ora
0->1
1->2
2->64
Base
Integer Floating Point
“Hit under n Misses”
0->11->22->64Base
Value of Hit Under Miss for SPEC
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.53
° L2 Equations
AMAT = Hit TimeL1 + Miss RateL1 x Miss PenaltyL1
Miss PenaltyL1 = Hit TimeL2 + Miss RateL2 x Miss PenaltyL2
AMAT = Hit TimeL1 +
Miss RateL1 x (Hit TimeL2 + Miss RateL2 x Miss PenaltyL2)
° Definitions:• Local miss rate— misses in this cache divided by the total number of
memory accesses to this cache (Miss rateL2)
• Global miss rate—misses in this cache divided by the total number of memory accesses generated by the CPU (Miss RateL1 x Miss RateL2)
• Global Miss Rate is what matters
4. Reduce Penalty: Second-Level Cache
Proc
L1 Cache
L2 Cache
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.54
° Reducing Miss Rate1. Reduce Misses via Larger Block Size
2. Reduce Conflict Misses via Higher Associativity
3. Reducing Conflict Misses via Victim Cache
4. Reducing Misses by HW Prefetching Instr, Data
5. Reducing Misses by SW Prefetching Data
6. Reducing Capacity/Conf. Misses by Compiler Optimizations
Reducing Misses: which apply to L2 Cache?
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.55
Relative CPU Time
Block Size
11.11.21.31.41.51.61.71.81.9
2
16 32 64 128 256 512
1.361.28 1.27
1.34
1.54
1.95
° 32KB L1, 8 byte path to memory
L2 cache block size & A.M.A.T.
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.56
1. Reduce the miss rate,
2. Reduce the miss penalty, or
3. Reduce the time to hit in the cache:
Lower Associativity (+victim caching or 2nd-level cache)?
Multiple cycle Cache access (e.g. R4000)
Harvard Architecture
Careful Virtual Memory Design (rest of lecture!)
Improving Cache Performance (Continued)
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.57
° Sample Statistics:• 16KB I&D: Inst miss rate=0.64%, Data miss rate=6.47%• 32KB unified: Aggregate miss rate=1.99%
° Which is better (ignore L2 cache)?• Assume 33% loads/store, hit time=1, miss time=50• Note: data hit has 1 stall for unified cache (only one port)
AMATHarvard=(1/1.33)x(1+0.64%x50)+(0.33/1.33)x(1+6.47%x50) = 2.05AMATUnified=(1/1.33)x(1+1.99%x50)+(0.33/1.33)X(1+1+1.99%x50)= 2.24
ProcI-Cache-1
Proc
UnifiedCache-1
UnifiedCache-2
D-Cache-1
Proc
UnifiedCache-2
Example: Harvard Architecture
Unified
Harvard Architecture
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.58
Summary #1/ 2:
° The Principle of Locality:• Program likely to access a relatively small portion of the address
space at any instant of time.
- Temporal Locality: Locality in Time
- Spatial Locality: Locality in Space
° Three (+1) Major Categories of Cache Misses:• Compulsory Misses: sad facts of life. Example: cold start misses.
• Conflict Misses: increase cache size and/or associativity.Nightmare Scenario: ping pong effect!
• Capacity Misses: increase cache size
• Coherence Misses: Caused by external processors or I/O devices
° Cache Design Space• total size, block size, associativity
• replacement policy
• write-hit policy (write-through, write-back)
• write-miss policy
4/21/03 ©UCB Spring 2003 CS152 / Kubiatowicz
Lec21.59
Summary #2 / 2: The Cache Design Space
° Several interacting dimensions• cache size
• block size
• associativity
• replacement policy
• write-through vs write-back
• write allocation
° The optimal choice is a compromise• depends on access characteristics
- workload
- use (I-cache, D-cache, TLB)
• depends on technology / cost
° Simplicity often wins
Associativity
Cache Size
Block Size
Bad
Good
Less More
Factor A Factor B