Lecture 2: Technology Trends
Prof. Randy H. Katz Computer Science 252 Spring 1996
RHK.S96 1 Original Food Chain Picture
Big Fishes Eating Little Fishes
RHK.S96 2 1985 Computer Food Chain
Mainframe Work- PC Mini- station computer
Supercomputer Mini- supercomputer
RHK.S96 3 Mini- Mini- supercomputer computer 1995 Computer Food Chain
Mainframe
Work- PC Supercomputer station
Now who is eating whom? Massively Parallel Processors RHK.S96 4 Why Such Change in 10 years?
• Function – Rise of networking/local interconnection technology • Performance – Technology Advances » CMOS VLSI dominates TTL, ECL in cost & performance – Computer architecture advances improves low-end » RISC, superscalar, RAID, … • Price: Lower costs due to … – Simpler development » CMOS VLSI: smaller systems, fewer components – Higher volumes » CMOS VLSI : same dev. cost 10,000 vs. 100,000 units – Lower margins by class of computer, due to fewer services
RHK.S96 5 1985 Computer Food Chain Technologies ECL TTL MOS
Mainframe Work- PC Mini- station computer Supercomputer Mini- supercomputer
RHK.S96 6 Technology Trends: Microprocessor Capacity
100000000 “Graduation Window”
10000000 Pentium Pro: 5.5 million Pentium PowerPC 620: 6.9 million
1000000 i80486 Alpha 21164: 9.3 million Sparc Ultra: 5.2 million i80386 i80286
Transistors 100000 CMOS improvements: • Die size: 2X every 3 yrs i8086 • Line width: halve / 7 yrs 10000 i8080 i4004
1000
1970 1975 1980 1985 1990 1995 2000
Year RHK.S96 7 CMOS Improvements Die size 2X every 3 yrs Line widths halve every 7 yrs 25 Die size increase plus 20 transistor count increase
15 Transistors Transistorper Unit Count Area 10 Ratio to Area in 1980 5
LineDie Width Size Improvement 0
1980 1983 1986 1989 1992 RHK.S96 8 Memory Capacity (Single Chip DRAM)
size
1000000000
100000000 year size cyc time 10000000 1980 64 Kb 250 ns
1000000 1983 256 Kb 220 ns Bits 1986 1 Mb 190 ns 100000 1989 4 Mb 165 ns
10000 1992 16 Mb 145 ns 1996 64Mb?? ?? 1000 1970 1975 1980 1985 1990 1995 2000
Year
RHK.S96 9 Moore’s Law for Memory
Primary memory size of various sized systems 8G in number of chips and $1.6M 1 Gw memory system price at 1G $50/chip over time $200K 1G
128M 128M $25.6K Number of chips 32M $3.2K 8M 32K 8M $400 1M 4K 640K $50 DOS 128K 512 chips limit 64 8 chips $6 1 chip 1/8th chip 8K 1970 1980 1990 2000 1Kbit 4K 16K 64K 256K 1M 4M 16M 64M 256M RHK.S96 10 Memory Size of Various Systems Over Time
8G 4K chips
1G 512 chips
128M 64 chips workstation
8M 8 chips-PC
1M 640K
128K DOS limit 1/8 chip 1 chip 8K 1970 1980 1990 2000 1Kbit 4K 16K 64K 256K 1M 4M 16M 64M 256MRHK.S96 11 Technology Trends (Summary)
Capacity Speed Logic 2x in 3 years 2x in 3 years DRAM 4x in 3 years 1.4x in 10 years Disk 2x in 3 years 1.4x in 10 years
RHK.S96 12 Processor Performance Trends
1000
Supercomputers 100 Mainframes
10 Minicomputers
0.1 1965 1970 1975 1980 1985 1990 1995 2000
Year RHK.S96 13 Performance vs. Time 100 4K Mips (65 Mhz)
Mips o RISC 60%25 mhz / yr 10 9000 uVAX 6K ECL 15%/yr o (CMOS) 8600 | | MIPS (8 Mhz) TTL MV10K 1.0 780 5 Mhz uVAX CMOS Will RISC continue on a Performance (VAX 780s) 68K CMOS CISC 38%/yr 60%, (x4 / 3 years)? Moore's speed law?
0.1 1980 1985 1990 RHK.S96 14 Processor Performance
300 Sun UltraSparc 1.54X/yr 250 P e DEC 21064a r 200 f o r 150 m IBM Power 2/590 a n 100 DEC AXP 3000 c e HP 9000/750 50 MIPS M/120 Sun-4/260 MIPS M2000 1.35X/yr IBM RS6000/540 0 1987 1988 1989 1990 1991 1992 1993 1994 1995 Year RHK.S96 15 Performance Trends (Summary)
• Workstation performance (measured in Spec Marks) improves roughly 50% per year
• Improvement in cost performance estimated at 70% per year
RHK.S96 16 Instructions/Second/$ vs. Time 100000
10000 PC
1000 Workstation
100 Super-mini
10
Instructions/second/$ Mainframe
1 1983 1985 1987 1989 1991 1993 1995 1997 RHK.S961999 17 Processor Perspective
• Putting performance growth in perspective: IBM POWER2 Cray YMP Workstation Supercomputer Year 1993 1988 MIPS > 200 MIPS < 50 MIPS Linpack 140 MFLOPS 160 MFLOPS Cost $120,000 $1M ($1.6M in 1994$) Clock 71.5 MHz 167 MHz Cache 256 KB 0.25 KB Memory 512 MB 256 MB • 1988 supercomputer in 1993 server! RHK.S96 18 Mini- Mini- supercomputer computer 1995 Computer Food Chain Technologies ECL CMOS
Mainframe
Work- PC Supercomputer station
Endangered Species?? Massively Parallel Processors RHK.S96 19 Where Has This Performance Improvement Come From?
• Technology? • Organization? • Instruction Set Architecture? • Software? • Some combination of all of the above?
RHK.S96 20 Performance Trends Revisited (Architectural Innovation)
1000
Supercomputers 100 Mainframes
10 Minicomputers
Microprocessors 1 CISC/RISC
0.1 1965 1970 1975 1980 1985 1990 1995 2000
Year RHK.S96 21 Performance Trends Revisited (Technology Advances)
Logic Speed: 2x per 3 years Logic Capacity: 2x per 3 years
Leads to: Computing capacity:4x per 3 years – If can keep all the transistors busy all the time – Actual: 3.3x per 3 years
RHK.S96 22 Performance Trends Revisited (Microprocessor Organization)
100000000 • Bit Level Parallelism
10000000 • Pipelining
r4400 • Caches 1000000 r4000 • Instruction Level i80386 i80286 Parallelism
Transistors 100000 r3010
i8086 • Out-of-order Xeq 10000 i8080 • Speculation i4004 1000 • . . . 1970 1975 1980 1985 1990 1995 2000
Year
RHK.S96 23 What is Ahead?
• Greater instruction level parallelism? • Bigger caches? • Multiple processors per chip? • Complete systems on a chip? (Portable Systems)
• High performance LAN, Interface, and Interconnect
RHK.S96 24 Hardware Technology
1980 1990 2000 Memory chips 64 K 4 M 256 M-1 G Speed 1-2 20-40 400-1000 5-1/4 in. disks 40 M 1 G 20 G Floppies .256 M 1.5 M 500-2,000 G LAN (Switch) 2-10 Mbits 10 (100) 155-655 (ATM) Busses 2-20 Mbytes 40-400
RHK.S96 25 Software Technology
1980 1990 2000 • Languages C, FORTRAN C++, HPF object stuff?? • Op. System proprietary +DUM* +DUM+NT • User I/F glass Teletype WIMP* stylus, voice, audio,video, ?? • Comp. Styles T/S, PC Client/Server agents*mobile • New things PC & WS parallel proc. appliances • Capabilities WP, SS WP,SS, mail video, ??
• DUM = DOS, n-Unixes, MAC • WIMP = Windows, Icons, Mouse, Pull-down menus • Agents = robots that work on information
RHK.S96 26 Computing 2001
• Continue quadrupling memory every 3 years – 1K chip in 72 becomes 1 gigabit chip (128 Mbyte) in 2002
• On-line 12-25 Gigabytes; $10 1-Gbyte floppies & CDs
• Micros increase at 60% per year ... parallelism »100
• Radio links for untethered computing
RHK.S96 27 Computing 2001
• Telephone, fax, radio, television, camera, house, ... Real personal (watch, wallet,notepad) computers
• We should be able to simulate: – Nearly everything we make and their factories – Much of the universe from the nucleus to galaxies
• Performance implies: voice and visual Ease of use. Agents!
RHK.S96 28 Applications: Unlimited Opportunities
• Office agents: phone/FAX/comm; files/paper handling • Untethered computing: fully distributed offices ?? • Integration of video, communication, and computing: desktop video publishing, conferencing, & mail • Large, commercial transaction processing systems • Encapsulate knowledge in a computer: scientific & engineering simulation (e.g. planetarium, wind tunnel, ... )
RHK.S96 29 Applications: Unlimited Opportunities
• Generic mathematics; visualization & virtual reality • Computational chemistry e.g biochemistry and matterials • Mechanical engineering without prototypes • Image/signal processing: medicine, maps, surveillance. • Personal computers in 2001 are today's supercomputers • Integration of the PC & TV => TC
RHK.S96 30 Challenges for 1990s Platforms
• 64-bit computers & 16 Mbit (2Mbyte DRAM): video, voice, communication, any really new apps? • Increasingly large, complex systems and environments Usability? • Plethora of non-portable, distributed, incompatible, non-interoperable computers: Usability? • Scalable parallel computers can provide “commodity supercomputing”: Markets and trained users?
RHK.S96 31 Challenges for 1990s Platforms
• Apps to fuel and support a chubby industry: communications, paper/office, and digital video • Computer commoditization & eventual disappearance (true ubiquitous computing!) • The true portable, wireless communication computer • Truly personal card, pencil, pocket, wallet computer • Networks continue to limit: WAN, ISDN, and ATM?
RHK.S96 32