### Table 1: Threshold Scheduling : Schedule Regeneration for Intel Paragon, Delta and Gamma

"... In PAGE 14: ... 4 Results and Conclusion The algorithm outlined in the previous section has been implemented in the Sisal com- piler backend for performing compile time scheduling for Intel i860 family of scalable distributed memory machines: Gamma, Delta, and Paragon. Table1 summarizes the results obtained from the compile time analysis on di erent packages and Liver- more loops. Initially the Threshold Scheduling Algorithm found the schedule for Intel Paragon which has the lowest communication cost amongst the three machines.... ..."

### Table 1. Characteristics of machines used in experiments. Alpha Beta Gamma Delta Epsilon

2000

"... In PAGE 3: ... The machines, which we call Alpha, Beta, Gamma, Delta, and Epsilon, were purchased in 1990, 1993, 1995, 1997, and (a better configured machine) 1999 respectively. Their clock speeds and model names are shown in Table1 . Other differences between the machines include, not just bus speed, but bus design, which has been considerably improved between Al- pha and Epsilon.... In PAGE 3: ... Alpha runs the SunOS operating system; the others run recent versions of Solaris. The capacities of individual disks that came with these machines are shown in Table1 . The disk drives are all typical of their era, and vary in seek and latency times only a little; the newer disks are less than two times faster than the old.... ..."

Cited by 4

### Table 4: Conditions on a valuation ( ; ; ).

"... In PAGE 43: ...Table4 are now in order. Condition (Delta Gamma 1) enforces the consistency between and the cardinality function C of the con guration.... In PAGE 49: ... The other distances between x and the entities within Emcs can be derived from those distances imposed by condition 2. Note that since 0 and 0 are in the context of a q0-valuation, they satisfy, by de nition, the conditions of Table4 . The choice of constraints imposed in the previous de nition derives from the compatibility we want to have between the graph GH and the allocation sequences in the language of H.... In PAGE 50: ... We start by computing the values for y. Since 0(e 3 ; y) = 0 and by condition (Delta Gamma 1) of Table4 it follows 0(e 3 ; e+ 3 ) 1 = , then we have: (y; e+ 3 ) 1 = which has as solution (y; e+ 3 ) 2 f3; g. Moreover, from these values we can deduce (y; e 4 ) = (y; e+ 3 ) 1 = .... ..."

### Table 4, = (e 3 ; e+ 3 ) 1 = 0(e 3 ; x) 0(x; e+ 3 ) 1

"... In PAGE 42: ... We denote by V the set of all -valuations ranged over by v and we write Vq for V q. (Delta Gamma 1) 8e 2 E : (e ; e+) 1 = C (e) (Delta Gamma 2) 8e1; e2 2 E : (e1 e2 , (e+ 1 ; e 2 ) = 1) (Delta Theta 1) 8x;y 2 fv( ) : (x; y) 2 dom( ) ) x; y 2 dom( ) ^ (x) (y) (Delta Theta 2) (x) = e , (e ; x) a5 0 ^ (x;e+) a5 0 (Delta Met 1) (x; x) 2 dom( ) ) (x; x) = 0 (Delta Met 2) (x; y);(y; z) 2 dom( ) ) (x;z) 2 dom( ) (Delta Met 3) (x; y);(x; z) 2 dom( ) ) ( (x; y) (y;z) = (x; z))_ ( (x;z) (z;y) = (x;y)) Table4 : Conditions on a valuation ( ; ; ). The function is the notion of distance that is used in a valuation together with the inter- pretation to decide the equality and leads-to propositions.... In PAGE 43: ...Table4 are now in order. Condition (Delta Gamma 1) enforces the consistency between and the cardinality function C of the con guration.... In PAGE 49: ... The other distances between x and the entities within Emcs can be derived from those distances imposed by condition 2. Note that since 0 and 0 are in the context of a q0-valuation, they satisfy, by de nition, the conditions of Table4 . The choice of constraints imposed in the previous de nition derives from the compatibility we want to have between the graph GH and the allocation sequences in the language of H.... In PAGE 50: ... We start by computing the values for y. Since 0(e 3 ; y) = 0 and by condition (Delta Gamma 1) of Table4 it follows 0(e 3 ; e+ 3 ) 1 = , then we have: (y; e+ 3 ) 1 = which has as solution (y; e+ 3 ) 2 f3; g. Moreover, from these values we can deduce (y; e 4 ) = (y; e+ 3 ) 1 = .... ..."

### Table 4. We denote by V the set of all -valuations ranged over by v and we write Vq for V q.

"... In PAGE 42: ...8x;y 2 fv( ) : (x; y) 2 dom( ) ) x; y 2 dom( ) ^ (x) (y) (Delta Theta 2) (x) = e , (e ; x) a5 0 ^ (x;e+) a5 0 (Delta Met 1) (x; x) 2 dom( ) ) (x; x) = 0 (Delta Met 2) (x; y);(y; z) 2 dom( ) ) (x;z) 2 dom( ) (Delta Met 3) (x; y);(x; z) 2 dom( ) ) ( (x; y) (y;z) = (x; z))_ ( (x;z) (z;y) = (x;y)) Table4 : Conditions on a valuation ( ; ; ). The function is the notion of distance that is used in a valuation together with the inter- pretation to decide the equality and leads-to propositions.... In PAGE 43: ...Table4 are now in order. Condition (Delta Gamma 1) enforces the consistency between and the cardinality function C of the con guration.... In PAGE 49: ... The other distances between x and the entities within Emcs can be derived from those distances imposed by condition 2. Note that since 0 and 0 are in the context of a q0-valuation, they satisfy, by de nition, the conditions of Table4 . The choice of constraints imposed in the previous de nition derives from the compatibility we want to have between the graph GH and the allocation sequences in the language of H.... In PAGE 50: ... We start by computing the values for y. Since 0(e 3 ; y) = 0 and by condition (Delta Gamma 1) of Table4 it follows 0(e 3 ; e+ 3 ) 1 = , then we have: (y; e+ 3 ) 1 = which has as solution (y; e+ 3 ) 2 f3; g. Moreover, from these values we can deduce (y; e 4 ) = (y; e+ 3 ) 1 = .... ..."

### Table 2: Delta approximation of VaR for the ROE warrant

1997

"... In PAGE 38: ... v +;#0B B:A: and v ,;#0B B:A: are obtained by applying the barycentric approximations after J = 200 re#0Cnements, v +;#0B #01,, and v ,;#0B #01,, correspond to the linear-quadratic approximation of the risk-pro#0Cles. We recognize that the barycentric and the linear-quadratic approximation yield comparable results #28 Table2 0#29. The accuracy of both approximations is insensitive with respect to the level #0B.... In PAGE 39: ...97#25 18.68#25 Table2 0: Error of the barycentric and the Delta-Gamma approximations for risk pro#0Cles g V #01... In PAGE 40: ... v +;#0B B:A: and v ,;#0B B:A: are obtained by applying the barycentric approximations for J = 100 re#0Cnements, v +;#0B #01,, and v ,;#0B #01,, correspond to the linear-quadratic approximation of the risk-pro#0Cles. The accuracy of the Delta-Gamma approximation is within the range #5B,4:52#25; 8:47#25#5D for levels #0B =1#25; 3#25; 5#25, that of the barycentric approximation is within #5B,2:95#25; 3:60#25#5D #28see Table2 3#29. Again, it is realized how the asymmetry of the pro#0Ct- and-loss distribution changes with respect to di#0Berent prices of the underlying ABB stock.... In PAGE 40: ...166 1.46#25 Table2 1: Barycentric approximation for the ROE warrant On the whole, it is recognized that the barycentric is competitive with the Delta-Gamma ap- proximation. The accuracy of both approximations is insensitive with respect to the level #0B.... In PAGE 41: ...166 -1.52#25 Table2 2: Delta-Gamma approximation for the ROE warrant Error #0B =1#25 #0B =3#25 #0B =5#25 Underlying v +;#0B B:A: v +;#0B #01,, v +;#0B B:A: v +;#0B #01,, v +;#0B B:A: v +;#0B #01,, 1800 -1.17#25 2.... In PAGE 41: ...46#25 -1.52#25 Table2 3: Error of the barycentric and the Delta-Gamma approximations for the ROE warrant... ..."

### Table 1.3 Comparison of variance reduction methods based on delta-gamma approximations. Variance ratios are estimated from 120,000 replications; the stratified estimator uses 40 strata and 3000 samples per strata. Variance ratios are estimates of the computational speed-up relative to standard Monte Carlo.

### Table 3-1: Example of Xetal Parameter file gamma 1

in Unclassified

"... In PAGE 21: ....cmd extension. The generation of the command file needs also the Xetal configuration pa- rameter values. They are stored in a text file with a pre-defined order and format (see Table3 -1). The parameter values can be configured as desired.... In PAGE 22: ...2 g227 Koninklijke Philips Electronics N.V. 2003 Note that a line memory writing must to be preceded by two program cycles before reading. The Table3 -2 shows an example of that situation. Table 3-2 MAC blue(0),blue(-1),1_2; Writing Blue(0) NOP; PASS accu , green(0); ADD accu , blue(0); Reading Blue(0) after 2 cycles 3.... ..."

### Table 1: Examples of deltas

2001

Cited by 54