Michael H. Schulz, Erwin Trischler, and Thomas M. Sarfert. SOCRATES: A Highly Efficient Automatic Test Pattern Generation System. IEEE Transactions on Computer-Aided Design, 7(1):126--137, Januray 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....learning algorithm and how a free BDD is constructed to represent all the solutions. 3 Search State Equivalence In a high performance ATPG or SAT solver, learning plays a very important role: the knowledge learned is stored and used for pruning search space in the future, e.g. in SOCRATES [2] the knowledge is in the form of implications and in GRASP [3] and CHAFF [4] it is in the form of conflict clauses. In our approach, the knowledge is equivalent search state. As shown in Figure 2, we discovered that different complete solutions may share the same partial solution. We will explain ....

M. H. Schulz, E. Trischler and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System", IEEE Trans. CAD, vol.7, no.1, pp. 126-137, 1988.

....assignments. Note that there is a small difference between the procedure described in [14] and the recursive learning procedure described here for verification. When applying recursive learning to logic verification, indirect implications are stored as was done in the static learning procedure of [19]. Therefore, the above routine performs all direct implications along with indirect implications that have been identified and stored before. Enumeration and reasoning in recursive learning differ greatly from conventional searching schemes for solving design automation problems. Conventional ....

....as the argument, depending on the signal assignment. For the justification process (justify( we use test generation techniques based on FAN s [11] multiple backtrace procedure and implicit enumeration. The prestored indirect implications are used to speedup the process similar to that done in [19]. Our experimental results show that consistent satisfy( is generally reasonably efficient to solve the false negative problem. However, in many cases the process can be speeded up considerably by the following technique, which allows to decrease the size of the OBDD that has to be traversed by ....

Schulz M., Trischler E., Sarfert T.: "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System", IEEE Trans. on CAD, vol. 7, Jan. 88, pp. 126-137.

....built on top of a Propositional Satisfiability (SAT) algorithm, the same concepts can be integrated on application specific algorithms. Robust Search Algorithms for Test Pattern Generation 1 1 Introduction During the last decade a wealth of algorithms for deterministic ATPG have been proposed [3 6, 8 11, 12 14, 1617 ], many of which are extremely effective on most existing benchmarks, and permit very high fault coverages. Most of these ATPG algorithms are based on implicit enumeration [7] and incorporate different search pruning techniques to effectively reduce the amount of search in most practical cases. ....

....coverages. Most of these ATPG algorithms are based on implicit enumeration [7] and incorporate different search pruning techniques to effectively reduce the amount of search in most practical cases. Among the most well known search pruning concepts we have head lines [5] non local implications [12, 13], recursive learning [9, 10] E frontiers [6] transitive closure [3] non chronological backtracking [14] among several others. Despite this continued research effort on the effectiveness of ATPG algorithms, they still significantly rely on heuristic techniques. For example, preprocessing is ....

[Article contains additional citation context not shown here]

M. H. Schulz, E. Trischler and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System," IEEE Transactions on Computer-Aided Design, vol. 7, no. 1, pp. 126-137, January 1988.

....all PPIs have the don t care values. If a test cannot be found for the currently selected PO, the next PO on the ordered PO list is chosen and the new search process starts. 4.2. 3 ESSENTIAL ESSENTIAL [4, 31] sequential test generator is an extension of a combinational test generator, SOCRATES [32]. It works on two time frames simultaneously, namely current and previous time frames. ESSENTIAL uses static learning which involves both time frames in the learning process and the learning is performed from the POs. ESSENTIAL has two phases for generating tests for the CUT, namely preprocessing ....

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System," in Proc. Int. Test Conf., pp. 1016--1026, Sept. 1987.

....into the evaluation procedure. During functional evaluation, a check is made to determine whether the current vector is unjustifiable based on the structural properties and can be skipped. The relations obtained from the structural technique can also be identified using static learning techniques [15]. The advantage of the structural technique is that it processes circuit information locally, whereas the implications required in static learning [15] may involve larger areas of the circuit. 4.3 Dynamic Justification and Property Extraction Consider a vector at the dominator cone inputs that ....

....properties and can be skipped. The relations obtained from the structural technique can also be identified using static learning techniques [15] The advantage of the structural technique is that it processes circuit information locally, whereas the implications required in static learning [15] may involve larger areas of the circuit. 4.3 Dynamic Justification and Property Extraction Consider a vector at the dominator cone inputs that produces different faulty values at the dominator gate output. Suppose that the vector cannot be shown to be unjustifiable based on the information ....

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System, " IEEE Trans. Computer-Aided Design, vol. 7, no. 1, pp. 126--137, Jan. 1988.

....finally comparative experimental results with other approaches are given and future work is discussed. 5.1 Introduction Finite state machine can be regarded as sequential circuit. Test generation for sequential circuits has been investigated widely and have been recognized as a difficult problem [14,15]. Traditional gate level test generation algorithms use an iterative array model where each time frame is represented by a cell of combinational logic. A single fault in the circuit is treated as a fault in each cell of the iterative array. Many of these algorithms use techniques developed for ....

Schulz, M. et al., "SOCRATES: A highly efficient automatic test pattern generation system." IEEE Trans. on CAD., (Jan.1988), pp.126-137.

....256Mb main memory operating under Linux (Specint95: 8.09 [12] The time limit of the total ATPG job was 4 hours. ATPG has been performed as follows: Pre processes: Identification of (structurally) uninitializable and unobservable signal lines, resulting in untestable faults [4] static learning [11] (time limit 15 minutes) and GIS learning [4] time limit 1 hour) If the GIS list is complete, then during the remaining ATPG process no further attempts are made to find or optimize GISes. ATPG: 4 stages, each using one or more STPG methods with different search strategies The first two ATPG ....

M. Schulz, E. Trischler and T.M. Sarfert. Socrates: A highly efficient automatic test pattern generation system. Proc. of International Test Conference, pages 1016--1026, 1987.

.... composed only of free lines are fanout free and may be ignored until consistent values have been given to every bound line, so FAN stops its backtrace operation at head lines (instead of continuing to a circuit input as PODEM does) FAN is faster than PODEM for most circuits [Fuj85a] SOCRATES [STS88] introduced in 1988 by Schulz et al.) is an improvement on FAN. SOCRATES improves on FAN s unique sensitization procedure, but more importantly, SOCRATES is the first ATPG system that uses information from non local implications. We will use Figure 1.12 to illustrate how non local implications ....

....fanout in a circuit. Figure 1.12 is repeated here (with tagged gates) as Figure 3.1. Once again, if line B has the value 1, line F has the value 1; conversely, if line F has the value 0, line B has the value 0. SOCRATES discovers this implication by performing a structural analysis of the circuit [STS88] we find it by analyzing the formula representing the circuit. Given the formula for an unfaulted circuit, we can list all the non local implications of a given variable assignment by binding the variable and then noting the direct implications that use a ternary clause. Any implication that ....

[Article contains additional citation context not shown here]

M. H. Schulz, E. Trischler, and T. M. Sarfert. Socrates: A highly efficient automatic test pattern generation system. IEEE Transactions on CAD, pages 126--137, January 1988.

....in both area overhead and performance degradation. Therefore, a test generator that can efficiently handle both combinational and sequential circuits is necessary for solving the testing problem for a wide range of VLSI circuits. Although there are effective ATPG algorithms for combinational [1, 15, 30, 35, 57, 63, 65, 113, 114, 116, 117, 118, 127] 1 as well as sequential circuits [21, 28, 58, 73, 79] their average case performance is decreasing because of the ever increasing complexity of today s VLSI circuits. Therefore, we propose new efficient and robust structure based techniques for speeding up the deterministic test pattern ....

....number of storage elements used in the scan chain. Small test sets also reduce the test storage requirements. Although the previously proposed compact test set generation algorithms are successful in producing small test sets, the resulting test sets are still larger than the known lower bounds [1, 6, 8, 19, 34, 53, 54, 55, 81, 90, 92, 113, 123]. Therefore, in order to close this gap further, we propose two new algorithms, redundant vector elimination and essential fault reduction, for generating compact test sets for combinational circuits under the single stuck at fault model, and a new heuristic for estimating the minimum single ....

[Article contains additional citation context not shown here]

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System," IEEE Trans. on Computer-Aided Design, vol. 7, no. 1, pp. 126-137, January 1988.

.... is proven to be NP hard [14] several test set compaction algorithms based on different heuristics are proposed in the literature, e.g. static compaction [6] dynamic compaction [6] independent and compatible fault sets based test generation [1] 13] 16] 19] reverse order fault simulation [18], maximal compaction [16] rotating backtrace [16] ROTCO [17] high level test generation [10] double detection [11] 13] Two by one [12] 13] Three by two [12] 13] forced pair merging [4] and essential fault pruning [4] Although these algorithms are successful in producing small test ....

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A highly efficient automatic test pattern generation system", IEEE Trans. on Computer-Aided Design, vol. 7, no. 1, pp. 126-137, January 1988.

....often fails to 9000 735 for 32 parallel patterns. The same simulation process a target fault within an acceptable time. with 256 parallel patterns was four times faster and Recursive learning is a generalization of the learning finished after 10 seconds. strategies originally introduced in [24]. Starting from a CURRENT uses a two valued logic which is sufficient for a leakage fault simulation. The term leakage fault simulation is somewhat misleading since no explicit fault simulation has to be performed. After a true value simulation with N parallel patterns it can be directly ....

Schulz, M.H., Trischler, E., Sarfert, T.M., "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System", IEEE Transactions on CAD, Vol.7, No.1, pp.126-137, January 1988

....after two days 2 unjust. gates 2 1536 37.3 75.1 4 unjust. gates 2 3175 53.9 83.9 8 unjust. gates 2 8274 63.2 86.8 16 unjust. gates 2 23848 74.8 90.1 case (c) gates 2 3020 69.8 84.5 8 case (c) gates 2 2989 69.8 84.5 only in region 2 50202 73.4 89.5 4. Static learning Static learning [11] is a learning technique that determines indirect implications using the law of contraposition. If the injection of value c u at a node u causes the assignment of value c v to node v, it uses a learning criterion to decide whether the implication v # c v # u # c u should be stored. Unlike ....

M.H. Schulz, E. Trischler, and T.M. Sarfert, "Socrates: A Highly Efficient Automatic Test Pattern Generation System ", IEEE Trans. on Computer-Aided Design of Integrated Circuits and Systems, vol. 7 no. 1, pp. 126--137, 1988.

....generation problem is proven to be NP complete [7] it is very important to find efficient techniques to speed up the test generation process. After the branch and bound search algorithm introduced in PODEM [9] a number of techniques have been proposed in the literature to improve its performance [1, 8, 11, 14, 19, 20, 21, 24]. The test generation problem can also be viewed as a boolean satisfiability (SAT) problem. Because it is possible to reduce it to an equivalent SAT problem using a polynomial time reduction. Recently a number of test generation systems are introduced for solving the test generation problem using ....

....uses the observability and controllability values based on the SCOAP [5, 10] controllability=observability measures to guide the search process. In addition to the new test generation techniques that we have developed, the test generator uses the local implications [8] static global implications [19], X path check [9] unique sensitization [8] and dynamic unique sensitization [20] techniques proposed in the literature. We did not use any compact test set generation techniques in the test generator. F l r m A B a b = 0 c d h n g D E j p = 0 k = 0 C e = D i f Figure 1: Improved Unique ....

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A highly efficient automatic test pattern generation system", IEEE Trans. on Computer-Aided Design, pp. 126-137, January 1988.

....faults. The global solution, the minimum length test vectors for the circuit under test (CUT) is generated after the compaction of the test vectors previously generated. Since the advent of D algorithm [1] there are many variations of test vector generation algorithms for a given fault model [2, 3, 4, 5, 6, 7]. However, there is no change in the fundamental procedure described above. The classical single stuck at fault model enables a relatively easy scheme for local test vector generation. To detect a stuck at 0 fault at a terminal, the local test vector is formed by assigning a logic 1 at the same ....

M. H. Schulz, E. Trischler, and T. M. Sarfert, "SOCRATES: A highly efficient automatic test pattern generation system," IEEE Trans. on CAD, pp. 126--137, Jan. 1988.

....test pattern generation (atpg) techniques are used to test each connection and return a status of either irredundant, redundant, or abort. Although deriving a test for a single fault is a hard problem, techniques have been developed which are efficient and still result in very few aborted tests [16]. More complex forms of iterative improvement are transduction [13, 12] rewiring[7, 6] and global flow[3] Each of these attempts to optimize a circuit by first adding redundant connections to some gates followed by deleting (presumably a different set of) redundant connections. While the basic ....

M. Schulz, E. Trischler, and T. Sarfert. Socrates: A highly efficient automatic test pattern generation system. ieee Trans. Comp. Aided. Design, CAD-7(1):126--137, January 1988.

No context found.

Michael H. Schulz, Erwin Trischler, and Thomas M. Sarfert. SOCRATES: A Highly Efficient Automatic Test Pattern Generation System. IEEE Transactions on Computer-Aided Design, 7(1):126--137, Januray 1988.

No context found.

Michael H. Schulz, Erwin Trischler, and Thomas M. Sarfert. SOCRATES: A Highly Efficient Automatic Test Pattern Generation System. In Proceedings of International Test Conference (ITC), pages 1016--1026, August 1987.

No context found.

M. Schulz, E. Trischler, and T. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System, " IEEE Trans. on CAD., pp. 126-137, Jan. 1988.

No context found.

M. H. Schulz, E. Trischler and T. M. Sarfert, "SOCRATES: A highly efficient automatic test pattern generation system", IEEE Trans. on Computer-Aided Design, pp. 126-137, January 1988.

No context found.

M. H. Schulz, E. Trischler and T. M. Sarfert, "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System", IEEE Trans. CAD, Vol.7, No.1, 1988, pp. 126-137.

No context found.

M. Schulz et al., "SOCRATES: A Highly Efficient Automatic Test Pattern Generation System", IEEE TCAD, Jan. 1988, pp. 126-137.

No context found.

M.H. Schulz, E. Trischler, T.M. Sarfert. SOCRATES: A highly efficient automatic test pattern generation system // IEEE Trans. CAD. - Vol. 7. -Jun.1988. -P.126-137.

No context found.

M. Schulz, E. Trishler, and T. Sarfert, "Socrates: A highly efficient automatic test pattern generation system," IEEE Transactions on Computer-Aided Design, pp. 126--137, Jan. 1988.

No context found.

M. Schulz, E. Trischler and T. Sarfert, "SOCRATES: A Highly Efficient Automatic Test pattern Generation System," IEEE Trans. on Computer Aided Design, Vol. 7, No. 1, January 1988.

No context found.

M. Schulz, E. Trishler, and T. Sarfert. SOCRATES: A Highly Efficient Automatic Test Pattern Generation System. IEEE Transactions on Computer-Aided Design, 7:126--137, January 1988.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC