SPEC Benchmark Specifications - 052.alvinn GENERAL 1.1 Classification Alvinn is a single precision floating point benchmark. 1.2 Description This program trains a neural network called ALVINN (Autonomous Land Vehicle In a Neural Network) using backpropagation. ALVINN is designed to take as input sensory data from a video camera and a laser range finder and give as output the direction for a vehicle to travel in order to stay on the road. The 1220 input units comprise the two input retinas, one from a video camera and one from a laser range finder. The 35 output units are a linear representation of the direction in which the network thinks the vehicle should travel. The network is fully connected and has 30 hidden units. 1.3 Source/Author Pomerleau, D.A. With apreciation to Kingsley Morse. 1.4 Version/Date The parameter NUM_EPOCHS has been increased to 200 to give a more substantial run-time. 1.5 Other Information PERFORMANCE 2.1 Metrics Elapsed time in seconds. 2.2 Elapsed Time On most systems, the sum of system+user time should be very close to the elapsed time, as this benchmark does little I/O. The original intent of the benchmark was to measure only the time for the computation, omitting setup code. The SPEC reference time (to 3 sig. fig. ) is 7690 seconds. 2.3 Reports See sample output. 2.4 Additional Performance Considerations SOFTWARE 3.1 Language C 3.2 Operating System 3.3 Portability Should be portable to non-UNIX systems. 3.4 Vectorizability/Multi-Processor Issues 3.5 Miscellaneous Software None. 3.6 Known Bugs None. 3.7 Additional Software Considerations HARDWARE 4.1 Memory No special requirements. 4.2 Disks No special requirements. 4.3 Communication None. 4.4 Special Hardware No special requirements. 4.5 Additional Hardware Considerations OPERATIONAL 5.1 Disk Space 1.25 megabytes are sufficient. 5.2 Installation No special requirements. 5.3 Execution Execute from SPEC scripts or by running make in the benchmark directory. 5.4 Correctness Verification Checking is performed by spiff with a relative tolerance of 0.001. 5.5 Additional Operational Considerations 5.6 Sample Run initializing backprop parameters starting training EPOCH NUMBER 1: ERROR = 0.21528 EPOCH NUMBER 2: ERROR = 0.17416 EPOCH NUMBER 3: ERROR = 0.12103 EPOCH NUMBER 4: ERROR = 0.08769 EPOCH NUMBER 5: ERROR = 0.07759 EPOCH NUMBER 6: ERROR = 0.07797 EPOCH NUMBER 7: ERROR = 0.08065 EPOCH NUMBER 8: ERROR = 0.08282 EPOCH NUMBER 9: ERROR = 0.08386 EPOCH NUMBER 10: ERROR = 0.08387 EPOCH NUMBER 11: ERROR = 0.08323 EPOCH NUMBER 12: ERROR = 0.08246 EPOCH NUMBER 13: ERROR = 0.08199 EPOCH NUMBER 14: ERROR = 0.08197 EPOCH NUMBER 15: ERROR = 0.08214 EPOCH NUMBER 16: ERROR = 0.08206 EPOCH NUMBER 17: ERROR = 0.08145 EPOCH NUMBER 18: ERROR = 0.08044 EPOCH NUMBER 19: ERROR = 0.07940 EPOCH NUMBER 20: ERROR = 0.07864 EPOCH NUMBER 21: ERROR = 0.07823 EPOCH NUMBER 22: ERROR = 0.07802 EPOCH NUMBER 23: ERROR = 0.07782 EPOCH NUMBER 24: ERROR = 0.07753 EPOCH NUMBER 25: ERROR = 0.07718 EPOCH NUMBER 26: ERROR = 0.07682 EPOCH NUMBER 27: ERROR = 0.07645 EPOCH NUMBER 28: ERROR = 0.07605 EPOCH NUMBER 29: ERROR = 0.07561 EPOCH NUMBER 30: ERROR = 0.07515 EPOCH NUMBER 31: ERROR = 0.07469 EPOCH NUMBER 32: ERROR = 0.07427 EPOCH NUMBER 33: ERROR = 0.07388 EPOCH NUMBER 34: ERROR = 0.07353 EPOCH NUMBER 35: ERROR = 0.07320 EPOCH NUMBER 36: ERROR = 0.07285 EPOCH NUMBER 37: ERROR = 0.07247 EPOCH NUMBER 38: ERROR = 0.07204 EPOCH NUMBER 39: ERROR = 0.07155 EPOCH NUMBER 40: ERROR = 0.07102 EPOCH NUMBER 41: ERROR = 0.07046 EPOCH NUMBER 42: ERROR = 0.06991 EPOCH NUMBER 43: ERROR = 0.06936 EPOCH NUMBER 44: ERROR = 0.06882 EPOCH NUMBER 45: ERROR = 0.06827 EPOCH NUMBER 46: ERROR = 0.06770 EPOCH NUMBER 47: ERROR = 0.06713 EPOCH NUMBER 48: ERROR = 0.06654 EPOCH NUMBER 49: ERROR = 0.06594 EPOCH NUMBER 50: ERROR = 0.06534 EPOCH NUMBER 51: ERROR = 0.06474 EPOCH NUMBER 52: ERROR = 0.06414 EPOCH NUMBER 53: ERROR = 0.06352 EPOCH NUMBER 54: ERROR = 0.06290 EPOCH NUMBER 55: ERROR = 0.06227 EPOCH NUMBER 56: ERROR = 0.06164 EPOCH NUMBER 57: ERROR = 0.06103 EPOCH NUMBER 58: ERROR = 0.06042 EPOCH NUMBER 59: ERROR = 0.05981 EPOCH NUMBER 60: ERROR = 0.05921 EPOCH NUMBER 61: ERROR = 0.05860 EPOCH NUMBER 62: ERROR = 0.05801 EPOCH NUMBER 63: ERROR = 0.05741 EPOCH NUMBER 64: ERROR = 0.05683 EPOCH NUMBER 65: ERROR = 0.05625 EPOCH NUMBER 66: ERROR = 0.05568 EPOCH NUMBER 67: ERROR = 0.05512 EPOCH NUMBER 68: ERROR = 0.05457 EPOCH NUMBER 69: ERROR = 0.05403 EPOCH NUMBER 70: ERROR = 0.05351 EPOCH NUMBER 71: ERROR = 0.05300 EPOCH NUMBER 72: ERROR = 0.05250 EPOCH NUMBER 73: ERROR = 0.05201 EPOCH NUMBER 74: ERROR = 0.05153 EPOCH NUMBER 75: ERROR = 0.05106 EPOCH NUMBER 76: ERROR = 0.05060 EPOCH NUMBER 77: ERROR = 0.05015 EPOCH NUMBER 78: ERROR = 0.04971 EPOCH NUMBER 79: ERROR = 0.04928 EPOCH NUMBER 80: ERROR = 0.04885 EPOCH NUMBER 81: ERROR = 0.04844 EPOCH NUMBER 82: ERROR = 0.04803 EPOCH NUMBER 83: ERROR = 0.04763 EPOCH NUMBER 84: ERROR = 0.04725 EPOCH NUMBER 85: ERROR = 0.04686 EPOCH NUMBER 86: ERROR = 0.04649 EPOCH NUMBER 87: ERROR = 0.04612 EPOCH NUMBER 88: ERROR = 0.04575 EPOCH NUMBER 89: ERROR = 0.04539 EPOCH NUMBER 90: ERROR = 0.04504 EPOCH NUMBER 91: ERROR = 0.04469 EPOCH NUMBER 92: ERROR = 0.04435 EPOCH NUMBER 93: ERROR = 0.04401 EPOCH NUMBER 94: ERROR = 0.04368 EPOCH NUMBER 95: ERROR = 0.04335 EPOCH NUMBER 96: ERROR = 0.04303 EPOCH NUMBER 97: ERROR = 0.04271 EPOCH NUMBER 98: ERROR = 0.04240 EPOCH NUMBER 99: ERROR = 0.04209 EPOCH NUMBER 100: ERROR = 0.04179 EPOCH NUMBER 101: ERROR = 0.04149 EPOCH NUMBER 102: ERROR = 0.04119 EPOCH NUMBER 103: ERROR = 0.04090 EPOCH NUMBER 104: ERROR = 0.04061 EPOCH NUMBER 105: ERROR = 0.04033 EPOCH NUMBER 106: ERROR = 0.04005 EPOCH NUMBER 107: ERROR = 0.03977 EPOCH NUMBER 108: ERROR = 0.03949 EPOCH NUMBER 109: ERROR = 0.03922 EPOCH NUMBER 110: ERROR = 0.03894 EPOCH NUMBER 111: ERROR = 0.03867 EPOCH NUMBER 112: ERROR = 0.03840 EPOCH NUMBER 113: ERROR = 0.03814 EPOCH NUMBER 114: ERROR = 0.03787 EPOCH NUMBER 115: ERROR = 0.03760 EPOCH NUMBER 116: ERROR = 0.03733 EPOCH NUMBER 117: ERROR = 0.03706 EPOCH NUMBER 118: ERROR = 0.03679 EPOCH NUMBER 119: ERROR = 0.03652 EPOCH NUMBER 120: ERROR = 0.03624 EPOCH NUMBER 121: ERROR = 0.03597 EPOCH NUMBER 122: ERROR = 0.03569 EPOCH NUMBER 123: ERROR = 0.03542 EPOCH NUMBER 124: ERROR = 0.03514 EPOCH NUMBER 125: ERROR = 0.03487 EPOCH NUMBER 126: ERROR = 0.03462 EPOCH NUMBER 127: ERROR = 0.03437 EPOCH NUMBER 128: ERROR = 0.03413 EPOCH NUMBER 129: ERROR = 0.03391 EPOCH NUMBER 130: ERROR = 0.03369 EPOCH NUMBER 131: ERROR = 0.03347 EPOCH NUMBER 132: ERROR = 0.03325 EPOCH NUMBER 133: ERROR = 0.03303 EPOCH NUMBER 134: ERROR = 0.03279 EPOCH NUMBER 135: ERROR = 0.03255 EPOCH NUMBER 136: ERROR = 0.03232 EPOCH NUMBER 137: ERROR = 0.03209 EPOCH NUMBER 138: ERROR = 0.03186 EPOCH NUMBER 139: ERROR = 0.03164 EPOCH NUMBER 140: ERROR = 0.03143 EPOCH NUMBER 141: ERROR = 0.03122 EPOCH NUMBER 142: ERROR = 0.03102 EPOCH NUMBER 143: ERROR = 0.03082 EPOCH NUMBER 144: ERROR = 0.03063 EPOCH NUMBER 145: ERROR = 0.03044 EPOCH NUMBER 146: ERROR = 0.03025 EPOCH NUMBER 147: ERROR = 0.03007 EPOCH NUMBER 148: ERROR = 0.02989 EPOCH NUMBER 149: ERROR = 0.02971 EPOCH NUMBER 150: ERROR = 0.02954 EPOCH NUMBER 151: ERROR = 0.02937 EPOCH NUMBER 152: ERROR = 0.02920 EPOCH NUMBER 153: ERROR = 0.02904 EPOCH NUMBER 154: ERROR = 0.02887 EPOCH NUMBER 155: ERROR = 0.02871 EPOCH NUMBER 156: ERROR = 0.02855 EPOCH NUMBER 157: ERROR = 0.02838 EPOCH NUMBER 158: ERROR = 0.02822 EPOCH NUMBER 159: ERROR = 0.02806 EPOCH NUMBER 160: ERROR = 0.02790 EPOCH NUMBER 161: ERROR = 0.02775 EPOCH NUMBER 162: ERROR = 0.02759 EPOCH NUMBER 163: ERROR = 0.02744 EPOCH NUMBER 164: ERROR = 0.02729 EPOCH NUMBER 165: ERROR = 0.02714 EPOCH NUMBER 166: ERROR = 0.02699 EPOCH NUMBER 167: ERROR = 0.02684 EPOCH NUMBER 168: ERROR = 0.02669 EPOCH NUMBER 169: ERROR = 0.02654 EPOCH NUMBER 170: ERROR = 0.02640 EPOCH NUMBER 171: ERROR = 0.02625 EPOCH NUMBER 172: ERROR = 0.02611 EPOCH NUMBER 173: ERROR = 0.02597 EPOCH NUMBER 174: ERROR = 0.02583 EPOCH NUMBER 175: ERROR = 0.02569 EPOCH NUMBER 176: ERROR = 0.02555 EPOCH NUMBER 177: ERROR = 0.02541 EPOCH NUMBER 178: ERROR = 0.02528 EPOCH NUMBER 179: ERROR = 0.02514 EPOCH NUMBER 180: ERROR = 0.02500 EPOCH NUMBER 181: ERROR = 0.02486 EPOCH NUMBER 182: ERROR = 0.02472 EPOCH NUMBER 183: ERROR = 0.02458 EPOCH NUMBER 184: ERROR = 0.02444 EPOCH NUMBER 185: ERROR = 0.02429 EPOCH NUMBER 186: ERROR = 0.02414 EPOCH NUMBER 187: ERROR = 0.02399 EPOCH NUMBER 188: ERROR = 0.02383 EPOCH NUMBER 189: ERROR = 0.02368 EPOCH NUMBER 190: ERROR = 0.02352 EPOCH NUMBER 191: ERROR = 0.02336 EPOCH NUMBER 192: ERROR = 0.02321 EPOCH NUMBER 193: ERROR = 0.02305 EPOCH NUMBER 194: ERROR = 0.02291 EPOCH NUMBER 195: ERROR = 0.02277 EPOCH NUMBER 196: ERROR = 0.02264 EPOCH NUMBER 197: ERROR = 0.02252 EPOCH NUMBER 198: ERROR = 0.02240 EPOCH NUMBER 199: ERROR = 0.02229 EPOCH NUMBER 200: ERROR = 0.02217