### Table 3. Di erential analysis for an arbitrary number of rounds. (ylog2)

1998

"... In PAGE 11: ... This results in an improvement of the signal-to-noise-ratio by a factor of 232 pc, where pc represents the factor by which the probability of the characteristic is reduced when we perform a 1R-attack instead of a 2R-attack. Results for an arbitrary number of rounds Table3 lists for each number of rounds: the probability of the characteristic, the required number of chosen plaintexts (assuming 4 right pairs for the characteristic are su cient, and that we need both guesses for the permutation key bit in the rst round structure),... In PAGE 12: ...5 Key dependency of the attacks We described the previous attacks using the best conditional characteristic. In Table3 we have listed for how many keys this works. For Thin-ICE the attack works for a fraction 2?2 of the keys.... ..."

Cited by 1

### Table 5: Average performance data (rounded to nearest integer) in terms of number of iterations, number of nodes expanded and number of nodes generated for 6 different problem sizes of 75 random problem instances each on CATS arbitrary

### Table 2. Instruction sets (used abbreviations: s = signed, u = unsigned, sat = saturation, m = modulo 2n, r = rounded, h = only higher bits, l = only lower bits, a = arbitrary bits, quot; = stored in the upper bits, # = lled up with sign-extension)

"... In PAGE 3: ... It is suited for both integer and oating point calculations and o ers the most powerful in- struction set. Table2 lists all SIMD instructions that are required for the data-parallel implementation of the neural operations I to V on all ve SIMD units. Whereas a unique instruction is available for each SIMD-parallel op- eration on 32-bit oat data elements, a variety of arithmetic instructions (for di erent data sizes, for di erent types, with/without rounding, with/without saturation) exists in case of integer operands.... ..."

### Table 5. Bridging relations.

2004

"... In PAGE 26: ...1. The results are given in Table5 (the total adds up to slightly over 100% due to arbitrary approximations when rounding up decimals). Table 5.... ..."

Cited by 2

### Table 2. An example of AC work calculation. Each member of the AC can calculate this table indepen- dently. Here, the AC is executing an application with three functions. The choice of is somewhat arbitrary and can vary based on the context of a particular func- tion.

2006

"... In PAGE 6: ... Our algorithm works in two rounds. First, each member calculates a table similar to Table2 . Then, AC members enter into a distributed bidding phase to adjust their indi- vidual workload.... ..."

Cited by 14

### Table 2. An example of AC work calculation. Each member of the AC can calculate this table indepen- dently. Here, the AC is executing an application with three functions. The choice of is somewhat arbitrary and can vary based on the context of a particular func- tion.

2006

"... In PAGE 6: ... Our algorithm works in two rounds. First, each member calculates a table similar to Table2 . Then, AC members enter into a distributed bidding phase to adjust their indi- vidual workload.... ..."

Cited by 14