Below is data for projecting Moore's law, adapted from Hans Moravec's data published at:
http://www.transhumanist.com/volume1/moravec.htm I began the analysis at 1985 and deleted computers not based upon single commodity CPU chips in order to focus the forecast. Without further debate, assume as does Moravec that 100 million MIPS equals one human's brain power. Also assume that MIPS can be measured by the popular SPECINT benchmark so that MIPS = 4 * SpecINT2000. http://www.spec.org/osg/cpu2000 You can modify the model to visualize results under different assumptions - if you disagree with those above. I define Human Equivalent Computing (HEC) as the common logarithmic ratio between a reference computer of 100 million MIPS (per Moravec) and a given computer. The use of the logarithmic ratio is important because of the exponential (possibly super-exponential) observed values over time. Computer power with regard to human equivalent brain power can be thus compared using decibels (db): http://www.phys.unsw.edu.au/~jw/dB.htm I fitted a linear regression to the logarithmic data and projected it forwards until 2047. In another post I will attach the MS Excel Spreadsheet and graph - and if our mail list does not accept attachments, then please send a request for the spreadsheet to me directly. HEC db Computer ====== ==================================== -50 500 MHz Pentium III (1999) -45 AMD Athlon 2000+ (2002) -40 (1/10000 human brain power est. 2004) -30 (1/1000 human brain power est. 2008) -20 (1/100 human brain power est. 2013) -10 (1/10 human brain power est 2017) 0 (1 human brain power est. 2021) +10 (10 human brain power est. 2026) +20 (100 human brain power est. 2030) +30 (1000 human brain power est. 2035) +40 (10000 human brain power est. 2039) Progressing from -50 db HEC to 0 db HEC in 22 years is equivalent to Moore's Law doubling every 16 months. [ 2^16.61 = 100025, 22/16.61*12 = 15.9 ] Forecasts beyond 0 HEC db do not include the contribution of AGI so are only an upper bound. I will keep this spreadsheet updated as new CPUs are benchmarked at SPEC. The data in tab delimited format: Computer year mips MB memory HEC (db) By_Hand 1892 1.19E-08 9.54E-05 -159.24 Ohdner 1891 3.33E-09 1.49E-07 -164.78 Steiger_Millionaire 1900 1.33E-08 3.58E-07 -158.76 Hollerith 1908 1.85E-08 0.000286 -157.33 Analytical_Engine 1910 3.77E-07 0.0244 -144.24 Monroe_Calculator 1911 2.18E-08 2.86E-06 -156.62 IBM_Tabulator 1919 4.12E-08 2.38E-05 -153.85 Torres_Arithmometer 1920 3.58E-08 4.77E-06 -154.46 National-Ellis_3000 1928 7.38E-08 4.29E-06 -151.32 Burroughs_Class_16 1929 7.38E-08 4.29E-06 -151.32 Zuse-1 1938 4.24E-08 3.05E-05 -153.73 Zuse-2 1939 4.24E-07 3.05E-05 -143.73 BTL_Model_1 1939 2.00E-06 3.81E-06 -136.99 Zuse-3 1941 2.04E-06 0.000244 -136.90 BTL_Model_2 1943 1.03E-06 1.19E-05 -139.87 Colossus 1943 0.000224 2.38E-06 -116.50 BTL_Model_3 1943 2.83E-06 4.29E-05 -135.48 ASCC_(Mark_1) 1944 2.33E-06 0.000601 -136.33 Zuse-4 1945 2.04E-06 0.000244 -136.90 BTL_Model_5 1946 3.29E-06 0.000147 -134.83 ENIAC 1946 0.00289 9.54E-05 -105.39 Harvard_Mark_2 1947 6.22E-06 0.000488 -132.06 IBM_SSEC 1948 0.000597 4.58E-05 -112.24 EDSAC 1949 0.00255 0.00214 -105.93 SEAC 1950 0.00416 0.00549 -103.81 UNIVAC_I 1951 0.00575 0.00537 -102.40 Zuse-5 1952 9.33E-06 0.000244 -130.30 IBM_CPC 1952 0.00176 0.000154 -107.54 IBM_650 1953 0.000966 0.00488 -110.15 EDVAC 1954 0.0017 0.00537 -107.70 Whirlwind 1955 0.0694 0.00391 -91.59 IBM_704 1955 0.0536 0.0352 -92.71 Librascope_LGP-30 1956 0.000701 0.0159 -111.54 IBM_7090 1959 0.326 0.141 -84.87 IBM_1620 1960 0.00103 0.0122 -109.87 DEC_PDP-1 1960 0.124 0.0176 -89.07 Atlas 1961 1.4 0.0234 -78.54 Burroughs_5000 1962 0.0989 0.0254 -90.05 IBM_7040 1963 0.063 0.07 -92.01 Honeywell_1800 1963 0.15 0.093 -88.24 DEC_PDP-6 1964 0.169 0.0703 -87.72 CDC_6600 1964 8.76 4 -70.57 IBM_1130 1965 0.15 0.0156 -88.24 IBM_360/75 1966 2.54 8 -75.95 IBM_360/65 1967 1.24 4 -79.07 DEC_PDP-10 1968 0.655 0.562 -81.84 CDC_7600 1969 25.7 8 -65.90 DG_Nova 1969 0.1175 0.008 -89.30 GE-635 1970 0.649 0.5 -81.88 SDS_920 1971 0.105 0.25 -89.79 IBM_360/195 1972 17.3 0.5 -67.62 Honeywell_700 1972 0.075 0.031 -91.25 Prime_Computer_100 1973 0.36 0.031 -84.44 IBM-370/168 1974 8.88 1 -70.52 MITS_Altair 1974 0.01 0.00024 -100.00 DG_Eclipse 1975 0.47 0.25 -83.28 DEC-KL-10 1975 2.3 4.5 -76.38 DEC_PDP-11/70 1976 0.4 0.125 -83.98 Cray-1 1976 150 32 -58.24 Apple_II 1977 0.02 0.0039 -96.99 DEC_VAX_11/780 1977 1 8 -80.00 TRS-80 1977 0.04 0.015 -93.98 Commodore_PET 1977 0.06 0.008 -92.22 CDC_IPL 1978 7.5 1 -71.25 Nanodata_VMX200 1979 2.1 2 -76.78 TRS-80_M3 1980 0.04 0.015 -93.98 Sun-1 1980 0.484 1 -83.15 CDC_Cyber-205 1981 73.2 16 -61.35 Vic_20 1981 0.04 0.005 -93.98 IBM_PC 1982 0.238 0.0635 -86.23 Sun-2 1982 0.741 2 -81.30 Commodore_64 1982 0.2 0.0825 -86.99 TRS-80_M4 1983 0.2 0.0635 -86.99 Vax_11/750 1983 0.799 4 -80.97 Macintosh-128K 1984 0.52 0.125 -82.84 Vax_11/785 1984 2.26 0.0156 -76.46 L.Edge_XT-7.16 1985 0.26 0.25 -85.85 Atari_800XL 1985 0.165 0.64 -87.83 MIT_XT-8 1986 0.534 0.25 -82.72 Mac_II 1987 2.5 2 -76.02 Mac-IIx 1988 3.9 4 -74.09 CompuAdd_386-16 1988 2.8 1 -75.53 PC_Brand_386-25 1988 4.3 1 -73.67 Macintosh_SE30 1989 3.9 5 -74.09 Dell_320LX 1990 12.5 1 -69.03 Mac_IIfx 1990 10 4 -70.00 Amiga_3000 1990 12.5 2 -69.03 Gateway-486DX2/66 1991 30.9 8 -65.10 ACT_468/33 1991 21.8 4 -66.62 Mac-Quadra-900 1991 22 8 -66.58 AST_Bravo 1992 12.9 2 -68.89 IBM_PS/2_55-041 1992 10.6 4 -69.75 NEC_Powermate 1992 21.8 4 -66.62 IBM_Valuepoint 1993 26.1 4 -65.83 Acer_Power 1993 44.5 4 -63.52 DECpc_LPv 1993 16.6 4 -67.80 IBM_433/DX/Si 1994 26.1 4 -65.83 Gateway_2000_486 1994 16.2 2 -67.90 PowerMac_7100/66 1994 100 8 -60.00 PowerMac_8500/120 1995 180 16 -57.45 PowerMac_9500/132 1995 200 16 -56.99 Intel_Xpress/60 1995 70 8 -61.55 Gateway_P5-75 1996 92 16 -60.36 Power_Tower_180e 1996 300 16 -55.23 PowerMac_7600/132 1996 160 16 -57.96 Gateway_G6-200 1997 350 64 -54.56 Power_Center_210 1997 375 32 -54.26 Mac_G3/266 1997 500 32 -53.01 iMac_G3/233 1998 450 32 -53.47 AcerPower8000/450B 1998 475 128 -53.23 Mac_G3/333 1998 650 64 -51.87 Pentium_II/455 1999 750 64 -51.25 Pentium_III/500 1999 820 128 -50.86 Mac_G4/450 1999 850 128 -50.71 Mac_G4/500_dual 2000 1500 256 -48.24 Mac_G4/867 2001 1500 128 -48.24 Dell_Workst_340/2G 2001 2500 256 -46.02 AMD Athlon 1 GHz 2001 1524 512 -48.17 AMD Athlon 2000+ 2002 2896 1536 -45.38 AMD Athlon 2800+ 2002 3732 1536 -44.28 Intel Pentium 4 3.06 GHz 2002 4428 1536 -43.54 Athlon-64 (est.) 2003 5971.2 4096 -42.24 -Steve -- =========================================================== Stephen L. Reed phone: 512.342.4036 Cycorp, Suite 100 fax: 512.342.4040 3721 Executive Center Drive email: [EMAIL PROTECTED] Austin, TX 78731 web: http://www.cyc.com download OpenCyc at http://www.opencyc.org =========================================================== ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
