Modular Inverse Algorithms without Multiplications for Cryptographic Applications

Hardware and algorithmic optimization techniques are presented to the left-shift-, right- shift-and the traditional Euclidean-modular inverse algorithms. Theoretical arguments and extensive simulations determined the resulting expected running time. On many computational platforms these turn out to be the fastest known algorithms for moderate operand lengths. They are based on variants of Euclidean-type extended GCD algorithms. On the considered computational platforms for operand lengths used in cryptography, the fastest presented modular inverse algorithms need about twice the time of modular multiplications, or even less. Consequently, in elliptic curve cryptography delaying modular divisions is slower (affine coordinates are best) and the RSA and ElGamal cryptosystems can be accelerated.


Introduction
We present improved algorithms for computing the inverse of large integers modulo a given prime or composite number, without multiplications of any kind.In most computational platforms they are much faster than the commonly used algorithms employing multiplications, therefore, the multiplier engines should be used for other tasks in parallel.The considered algorithms are based on different variants of the Euclidean type Greatest Common Divisor algorithms.They are iterative, gradually decreasing the length of the operands and keeping some factors updated, maintaining a corresponding invariant.There are other algorithmic approaches, too.One can use system of equations or the Little Fermat Theorem (see [21]), but they are only competitive with the Euclidean type algorithms under rare, special circumstances.
Several variants of three extended GCD algorithms are modified for computing modular inverses for operand lengths used in public key cryptography (128…16K bits).We discuss algorithmic improvements and simple hardware enhancements for speedups in digit serial hardware architectures.The main point of the paper is investigating how much improvement can be expected from these optimizations.It helps implementers to choose the fastest or smallest algorithm; allows system designer to estimate accurately the response time of security systems; facilitates the selection of the proper point representation for elliptic curves, etc. 1/ 18/2006 The discussed algorithms run in quadratic time: O(n 2 ) for n-bit input.For very long operands more complex algorithms, such as Schönhage's half-GCD algorithm [8] get faster, running in O(n log 2 n) time, but for operand lengths used in cryptography they are far too slow (see [9]).

Extended Greatest Common Divisor Algorithms
Given 2 integers x and y the extended GCD algorithms compute their greatest common divisor g, and also two integer factors c and d: [g,c,d] = xCGD(x,y), such that g = c•x+d•y.For example, the greatest common divisor of 6 and 9 is 3; and 3 = (−1)•6 + 1•9.
In the sequel we will discuss several xGCD algorithms.(See also [2] or [10].)They are iterative, that is, their input parameters get gradually decreased, while keeping the GCD of the parameters unchanged (or keep track of its change).The following relations are used: • GCD(x,y) = GCD(x±y,y) • GCD(x,y) = 2•GCD(x/2,y/2) for even x and even y • GCD(x,y) = GCD(x/2,y) for even x and odd y.

Modular Inverse
The positive residues 1, 2, …, p −1 of integers modulo p (a prime number) form a multiplicative group G, that is, they obey the following 4 group laws: 1. Closure: If x and y are two elements in G, then the product x•y := x y mod p is also in G.
3. Identity: There is an identity element i ( = 1) such that i•x = x•i = x for every element x ∈ G. 4. Inverse: There is an inverse (or reciprocal) x −1 of each element x ∈ G, such that x•x −1 = i.The inverse mentioned in 4. above is called the Modular Inverse, if the group is formed by the positive residues modulo a prime number.For example the inverse of 2 is 3 mod 5, because 2•3 = 6 = _ 1 mod 5.
Positive residues modulo a composite number m do not form a group, as some elements don't have inverse.For example, 2 has no inverse mod 6, because every multiple of 2 is even, never 1 mod 6. Others, like 5 do have inverse, also called modular inverse.In this case the modular inverse of 5, 5 −1 mod 6, is also 5, because 5•5 = 25 = 24+1 = _ 1 mod 6.In general, if x is relative prime to m (they share no divisors) there is a modular inverse x −1 mod m. (See also in [2].) Modular inverses can be calculated with any of the numerous xGCD algorithms.If we set y = m, by knowing that GCD(x,m) = 1, we get 1 = c•x+d•m from the results of the xGCD algorithm.Taking this equation modulo m we get 1 = _ c•x.The modular inverse is the smallest positive such c, so either

Computing the xGCD Factors from the Modular Inverse
In embedded applications the code size is often critical, so if an application requires both xGCD and modular inverse, usually xGCD is implemented alone, because it can provide the modular inverse, as well.We show here that from the modular inverse the two xGCD factors can be reconstructed, even faster than it would take to compute them directly.Therefore, it is always better to implement a modular inverse algorithm than xGCD.These apply to subroutine libraries, too, there is no need for a full xGCD implementation.
The modular inverse algorithms return a positive result, while the xGCD factors can be negative.c = x −1 and c = x −1 -y provide the two minimal values of one xGCD factor.The other To get d, calculating only the MS half of x•x −1 , plus a couple of guard digits, is sufficient.Division with y provides an approximate quotient, which rounded to the nearest integer gives d.This way there is no need for longer than ||y||-bit arithmetic (except two extra digits for the proper rounding).The division is essentially of the same complexity as multiplication (for operand lengths in cryptography it takes 0.65…1.2times as long, see e.g.[17]).
For the general case g > 1 we need a trivial modification of the modular inverse algorithms: return the last candidate for the inverse before one of the parameters becomes 0 (as noted in [22] for polynomials).It gives x * such that x•x * ≡ g mod y.Again c = x * or c = x * -y and The extended GCD algorithm needs storage room for the 2 factors in addition to its internal variables.They get constantly updated during the course of the algorithm.As described above, one can compute the factors from the modular inverse and save the memory for one (long integer) factor and all of the algorithmic steps updating it.The xGCD algorithms applied for operand lengths in cryptography perform a number of iterations proportional to the length of the input, and so the operations on the omitted factor would add up to at least as much work as a shift-add multiplication algorithm would take.With a better multiplication (or division) algorithm not only memory, but also some computational work can be saved.

Cryptographic Applications
The modular inverse of long integers is used extensively in cryptography, like for RSA and ElGamal public key cryptosystems, but most importantly in Elliptic Curve Cryptography.

RSA
RSA encryption (decryption) of a message (ciphertext) g is done by modular exponentiation: g e mod m, with different encryption (e) and decryption (d) exponent, such that (g e ) d mod m = g.The exponent e is the public key, together with the modulus m = p•q, the product of 2 large primes.d is the corresponding private key.The security lies in the difficulty of factoring m. (See [10].)Modular inverse is used in − Modulus selection: in primality tests (excluding small prime divisors).If a random number has no modular inverse with respect to the product of many small primes, it proves that the random number is not prime.(In this case a simplified modular inverse algorithm suffice, which only checks if the inverse exists.)− Private key generation: computing the inverse of the chosen public key (similar to the signing/verification keys: the computation of the private signing key from the chosen public signature verification key).d = e −1 mod (p -1)(q -1) − Preparation for CRT (Chinese Remainder Theorem based computational speedup): the precalculated half-size constant C 2 = p −1 mod q (where the public modulus m = p•q) helps accelerating the modular exponentiation about 4-fold.
[10] − Signed bit exponent recoding: expressing the exponent with positive and negative bits facilitates the reduction of the number of non-zero signed bits.This way many multiplications can be saved in the multiply-square binary exponentiation algorithm.At negative exponent bits the inverse of the message g −1 mod m -which almost always exists and pre-computed in less time than 2 modular multiplications -is multiplied to the partial result [15].(In embedded systems, like smart cards or security tokens RAM is expensive, so other exponentiations methods, like windowing, are often inapplicable.)

ElGamal Encryption
The public key is (p, α, α a ), fixed before the encrypted communication, with randomly chosen α, a and prime p. Encryption of the message m is done by choosing a random k ∈ [1, p−2] and computing γ = α k mod p and δ = m•(α a ) k mod p.
Decryption is done with the private key a, by computing first the modular inverse of γ, then (γ −1 ) a = (α −a ) k mod p, and multiplying it to δ: δ•(α −a ) k mod p = m.(See also in [10].)

Elliptic Curve Cryptography
Prime field elliptic curve cryptosystems (ECC) are gaining popularity especially in embedded systems, because of their smaller need in processing power and memory than RSA or ElGamal.Modular inverses are used extensively during point addition, doubling and multiplication (See more details in [2]).20….30% overall speedup is possible, just with the use of a better algorithm.
An elliptic curve E over GF(p) (the field of residues modulo the prime p) is defined as the set of points (x,y) (together with the point at infinity O) satisfying the reduced Weierstraß equation: In elliptic curve cryptosystems the data to be encrypted is represented by a point P on a chosen curve.Encryption by the key k is performed by computing Q = P + P + … + P = k•P.Its security is based on the hardness of computing the discrete logarithm in groups.This operation, called scalar multiplication (the additive notation for exponentiation), is usually computed with the double-and-add method (the adaptation of the well-known square-and-multiply algorithm to elliptic curves, usually with signed digit recoding of the exponent [15]).When the resulting point is not the point at infinity O, the addition of points P = (x P ,y P ) and Q = (x Q ,y Q ) leads to the resulting point R = (x R ,y R ) through the following computation: Here the divisions in the equations for λ are shorthand notations for multiplications with the modular inverse of the denominator.P = (x P ,y P ) is called the affine representation of the elliptic curve point, but it is also possible to represent points in other coordinate systems, where the field divisions (multiplications with modular inverses) are traded to a larger number of field additions and multiplications.These other point representations are advantageous when computing the modular inverse is much slower than a modular multiplication.In [11] the reader can find discussions about point representations and the corresponding costs of elliptic curve operations.

Hardware Platforms 2.1 Multiplications
There are situations where the modular inverse has to be, or it is better calculated without any multiplication operations.These include 1/18/2006 − If the available multiplier hardware is slow.− If there is no multiplier circuit in the hardware at all.For example, on computational platforms where long parallel adders perform multiplications by repeated shift-add operations.(See [13] for fast adder architectures.)− For RSA key generation in cryptographic processors, where the multiplier circuit is used in the background for the exponentiations of the (Miller-Rabin) primality test.
[10] − In prime field elliptic or hyper elliptic curve cryptosystems, where the inversion can be performed parallel to other calculations involving multiplications.Of course, there are also computational platforms, where multiplications are better used for modular inverse calculations.These include workstations with very fast or multiple multiplier engines (could be three: ALU, Floating point multiplier and Multimedia Extension Module).
In digit serial arithmetic engines there is usually a digit-by-digit multiplier circuit (for 8…128 bit operands), which can be utilized for calculating modular inverses.This multiplier is the slowest circuit component; other parts of the circuit can operate at much higher clock frequency.Appropriate hardware designs, with faster non-multiplicative operations, can defeat the speed advantage of those modular inverse algorithms, which use multiplications.This way faster and less expensive hardware cores can be designed.
This kind of hardware architecture is present in many modern microprocessors, like the Intel Pentium processors.They have 1 clock cycle base time for a 32-bit integer add or subtract instruction (discounting operand fetch and other overhead), and they can sometimes be paired with other instructions for concurrent execution.A 32-bit multiply takes 10 cycles (a divide takes 41 cycles), and neither can be paired.

Shift and Memory Fetch
The algorithms considered in this paper process the bits or digits of their long operands sequentially, so in a single cycle fetching more neighboring digits (words) into fast registers allows the use of slower, cheaper RAM, or pipeline registers.
We will use only add/subtract, compare and shift operations.With trivial hardware enhancements the shift operations can be done "on the fly" when the operands are loaded for additions or subtractions.This kind of parallelism is customarily provided by DSP chips, and it results in a close to two-fold speedup of the shifting xGCD based modular inverse algorithms.
Shift operations could be implemented with manipulating pointers to the bits of a number.At a subsequent addition/subtraction the hardware can provide the parameter with the corresponding offset, so arbitrary long shifts take only a constant number of operations with this offset-load hardware support.(See [16].)Even in traditional computers these pointer manipulating shift operations save time, allowing multiple shift operations to be combined into a longer one.

Number Representation
For multi-digit integers signed magnitude number representation is beneficial.The binary length of the result is also calculated at each operation (without significant extra cost), and pointers show the position of the most and least significant bit in memory.− Addition is done from right to left (from the least-to the most significant bits), the usual way.− Subtraction needs a scan of the operand bits from left to right, to find the first different pair.
They tell the sign of the result.The leading equal bits need not be processed again, and the right-to-left subtraction from the larger number leaves no final borrow.This way subtraction is of the same speed as addition, like with 2's complement arithmetic.− Comparisons can be done by scanning the bits from left to right, too.For uniform random inputs the expected number of bit operations is constant, less than: 1•½+2•¼+3• 1 /8 …= 2. − Comparisons to 0, 1 or 2 k take constant time also in the worst case, if the head and tail pointers have been kept updated.1/18/2006

Modular Inverse Algorithms
We consider all three Euclidean-type algorithm families commonly used: the extended versions of the right-shift-, the left-shift-and the traditional Euclidean -algorithm.They all gradually reduce the length of their operands in an iteration, maintaining some invariants, which are closely related to the modular inverse.

Binary Right-shift: Algorithms RS
At the modular inverse algorithm based on the rightshift binary extended GCD (variants of the algorithm of M. Penk, see in [1], exercise 4.5.2.39 and [24]), the modulus m must be odd.The trailing 0-bits from two internal variables U and V (initialized to the input a, m) are removed by shifting them to the right, then their difference replaces the larger of them.It is even, so shifting right removes the new trailing 0-bits (Figure 1.).Repeat these until V = 0, when U = GCD(m,a).If U > 1, there is no inverse, so we return 0, which is not an inverse of anything.
In the course of the algorithm two auxiliary variables, R and S are kept updated.At termination R is the modular inverse.

Modification: Algorithm RS1
The 2 instructions marked with "/**/" in Figure 1.keep R and S nonnegative and so assure that they don't grow too large (the subsequent subtraction steps decrease the larger absolute value).These instructions are slow and not necessary, if we ensure otherwise, that the intermediate values of R and S do not get too large.
Handling negative values and fixing the final result is easy, so it is advantageous if instead of the marked instructions, we only check at the add-halving steps (R ←(R+m)/2 and S ←(S+m)/2), whether R or S was already larger (or longer) than m, and add or subtract m such that the result becomes smaller (shorter).These steps cost no additional work beyond choosing '+' or '−' and, if |R| ≤ 2m was beforehand, we get |R| ≤ m, the same as at the simple halving of R ← R/2 and S ← S/2.If |R| ≤ m and |S| ≤ m, |R−S| ≤ 2m (the length could increase by one bit) but these instructions are always followed by halving steps, which prevent R and S to grow larger than 2m during the calculations.(See code details at the plus-minus algorithm below.)

Even Modulus
This algorithm cannot be used for RSA key generation, because m must be odd (to ensure that either R or R±m is even for the subsequent halving step).We can go around the problem by swapping the role of m and a (a must be odd, if m is even, otherwise there is no inverse).The algorithm returns m −1 mod a, such that m•m −1 + k'•a = 1, for some negative integer k'.k' ≡ a −1 mod m, easily seen if we take both sides of the equation mod m.It is simple to compute the smallest positive k ≡ k' mod m: As we saw before, the division is fast with calculating only the MS half of m•m −1 , plus a couple of guard digits to get an approximate quotient, to be rounded to the nearest integer.
Unfortunately there is no trivial modification of the algorithm to handle even moduli directly, because at halving only an integer multiple of the modulus can be added without changing the result, and only adding an odd number can turn odd intermediate values to even.Fortunately, the only time we need to handle even moduli in cryptography is at RSA key generation, which is so slow anyway (requiring thousands of modular multiplications for the primality tests), that this black box workaround does not cause a noticeable difference in processing time.
An alternative was to perform the full extended GCD algorithm, calculating both factors c and d: [g,c,d] = xCGD(m,a), such that the greatest common divisor g = c•m+d•a [10].It would need extra storage for two factors, which are constantly updated during the course of the algorithm and it is also slower than applying the method above transforming the result of the modular inverse algorithm with swapped parameters.

Justification
The algorithm starts with U = m, V = a, R = 0, S = 1.In the course of the algorithm U and V are decreased, keeping GCD(U,V) = GCD(m,a) true.The algorithm reduces U and V until V = 0 and U = GCD(m,a): If one of U or V is even, it can be replaced by its half, since GCD(m,a) is odd.If both are odd, the larger one can be replaced by the even U−V, which then can be decreased by halving, leading eventually to 0. The binary length of the larger of U and V is reduced by at least one bit, guaranteeing that the procedure terminates in at most ||a||+||m|| iterations.
At termination of the algorithm V = 0 otherwise a length reduction was still possible.U = GCD(U,0) = GCD(m,a).Furthermore, the calculations maintain the following two congruencies: U ≡ Ra mod m, V ≡ Sa mod m. ( Having an odd modulus m, at the step halving U we have two cases.When R is even: U / 2 ≡ (R / 2)•a mod m, and when R is odd: U/2 ≡ ((R+m) / 2)•a mod m.The algorithm assigns these to U and R. Similarly for V and S, and with their new values (1) remains true.The difference of the two congruencies in (1) gives U−V ≡ (R−S)•a mod m, which ensures that at the subtraction steps (1) remains true after updating the corresponding variables: U or V ← U−V, R or S ← R−S.Choosing +m or −m, as discussed above, guarantees that R and S does not grow larger than 2m, so at the end we can just add or subtract m to make 0 < R < m.If U = 1 = GCD(m,a) we get 1 ≡ Ra mod m, and R is of the right magnitude, so R = a −1 mod m. []

Plus-Minus: Algorithm RS+−
There is a very simple modification often used for the right-shift algorithm [6]: for the odd U and V check, if U+V has 2 trailing zero bits, otherwise we know that U−V does.In the former case, if U+V is of the same length as the larger of them, the shift operation reduces the length by 2 bits from this larger length, otherwise by only one bit (as before with the rigid subtraction steps).It means that the length reduction is sometimes improved, so the number of iterations decreases.Unfortunately, this reduction is not large, only 15% (half of the time the reduction was by at least 2 bits, anyway, and longer shifts are not affected either), but it comes almost for free.Furthermore, R and S need more halving steps, and these get a little more expensive (at least one of the halving steps needs an addition of m), so the RS+− algorithm is not faster than RS1.

Double Plus-minus: Algorithm RS2+−
The plus-minus reduction can be applied also to R and S (Figure 2).In the course of the algorithm they get halved, too.If one of them happens to be odd, m is added or subtracted to make them even before the halving.The plus-minus trick on them ensures that the result has at least 2 trailing 0-bits.It provides a speedup, because most of the time we had exactly two divisions by 2 (shift right by two), and no more than one addition/subtraction of m is now necessary.

Delayed Halving: Algorithm RSDH
The variables R and S get almost immediately of the same length as m, because, when they are odd, m is added to them to allow halving without remainder.We can delay these add-halving steps, by doubling the other variable instead.When R should be halved we double S, and vice versa.Of course, a power-of-2 spurious factor is introduced to the computed GCD, but keeping track of the exponent a final correction step will fix R by the appropriate number of halving-or add-halving steps.(This technique is similar to the Montgomery inverse computation published in [19] and sped up for computers in [18], but the correction steps differ.)It provides an acceleration of the algorithm by 24…38% over RS1, due to the following: − R and S now increase gradually, so their average length is only half as it was in RS1.− The final halving steps are performed only with R. The variable S need not be fixed, being only an internal temporary variable.− At the final halving steps more short shifts can be combined to longer shifts, because they are not confined by the amount of shifts performed on U and V in the course of the algorithm.Note1: R and S are almost always of different lengths, and so their difference is not longer than the longer of R and S. Consequently, their lengths don't increase faster than what the shifts cause.Note2: it does not pay to check, if R or S is even, in the hope that some halving steps could be performed until the involved R or S becomes odd, and so speeding up the final correction, because they are already odd in the beginning (easily proved by induction).

Combined Speedups: Algorithm RSDH+−
The second variant of the plus-minus trick and the delayed halving trick can be combined, giving the fastest of the presented right shift modular algorithms.It is 43…60% faster than algorithm RS1 (which is 30% faster than the traditional implementation RS), but still slower on most computational platforms than the left shift and shifting Euclidean algorithms, discussed below.

Binary Left-Shift Modular Inverse: Algorithm LS1
The left-shift binary modular inverse algorithm (similar to the variant of R. Lórencz [12]) is described in Figure 3.It keeps the temporary variables U and V aligned to the left, such that a subtraction clears the leading bit(s).Shifting the result left until the most significant bit is again in the proper position restores the alignment.The number of known trailing 0-bits increases, until a single 1-bit remains, or the result is 0 (indicating that there is no inverse).As before, keeping 2 internal variables R and S updated, the modular inverse is calculated.1/18/2006 iterations and the number of shifts with the most frequent lengths.(Multiple shifts are combined together.)The computed curves fit to the data with less than 1% error at any operand length.
The right-shift algorithms are the slowest, because they halve two auxiliary variables (R, S) and if they happen to be odd, m is added or subtracted first, to make them even for the halving.Theoretical arguments and also our computational experiments showed that they are too slow at digit serial arithmetic.They were included in the discussions mainly, because there are surprisingly many systems deployed using some variant of the right-shift algorithm, although others are much better.
The addition steps are not needed in the left-shift-or in the shifting Euclidean algorithms.In all three groups of algorithms the length of U and V decreases bit-by-bit in each iteration, and in the left-shift-and shifting Euclidean algorithms the length of R and S increases steadily from 1.In the right shift case they get very soon as long as m, except in the delayed halving variant.In the average, the changing lengths roughly halve the work on those variables.Also, the necessary additions of m in the original right-shift algorithms prevent aggregation of the shift operations of R and S. On the other hand, in the other algorithms (including the delayed halving right shift algorithm) we can first determine by how many bits we have to shift all together in that phase.In the left-shift algorithms, dependent on the relative magnitude of u and v we need only one or two shifts by multiple bits, in the shifting Euclidean algorithm only one.This shift aggregation saves work at longer shifts than the most common lengths of 1 or 2.
On the other hand, the optimum shift lengths in the left-shift-and shifting Euclidean algorithms are only estimated from the MS bits.They are sometimes wrong, while in the rightshift algorithm only the LS bits play a role, so the optimum shift lengths can always be found exactly.Accordingly, the right-shift algorithms perform slightly fewer iterations (8.6…10%), but the large savings in additions in the other algorithms offset these savings.

Software Running Time Comparisons
We did not measure execution times of SW implementations, because of the following reasons: 1.The results are very much dependent on the characteristics of the hardware platforms (word length, instruction timings, available parallel instructions, length and function of the instruction pipeline, processor vs. memory speed, cache size and speed, number of levels of cache memory, page fault behavior, etc.).
2. The results also depend on the operating system (multitasking, background applications, virtual/paging memory handling, etc.).
3. The results are dependent on the code, the programming language and the compiler.For example, GMP [9] uses hand optimized assembler macros, and any other SW written in a higher level language is necessarily disadvantaged, like at handling carries.
In earlier publications running time measurements were reported, like in [5] Jebelean gave software execution time measurements of GCD algorithms on a DEC computer of RISC architecture.Our measurements on a 3 GHz Intel Pentium PC running Windows XP gave drastically different speed ratios.This large uncertainty was the reason why we decided to count the number of specific operations and sum up their time consumption dependent on the operand lengths, instead of the much easier running time measurements.This way the actual SW running time can be well estimated on many different computational platforms.

Notes on the Simulation Results
-The number of the different UV shifts, together, is the number of iterations, since there is one combined shift in each iteration.1/18/2006 parallel instructions takes about twice as much time as a modular multiplication.Still, in case of elliptic curve cryptography the most straightforward (affine) point representation and direct implementation of the point addition is the best (computation with the projective, Jacobian and Chudnovsky-Jacobian coordinates are slower, see [11]).
It is interesting to note, that modular division (p•q −1 ) can be performed faster than 3 modular multiplications.Similar results were presented in [22] and [23] for polynomials of practical lengths, showing that even in extension fields GF(p k ), elliptic curve points are best represented in affine coordinates.

Performance Relative to Parallel Adder-based Modular Multiplication
When very long adders are implemented in hardware, repeated shift-add steps can perform multiplications in linear time.To prevent the partial results from growing too large, interleaved modular reduction is performed.Scanning the bits of the second multiplicand from left to right, when a 1-bit is found, the first multiplicand is added in the appropriate position to the partial result r.If it gets too long: ||r|| > ||m||, it is reduced by subtracting the modulus m.These addsubtract steps are usually done in the same clock cycle, resulting in performing an n-bit modular multiplication in n clock cycles.
In these kinds of hardware architectures the speed of the different modular inverse algorithms becomes very close, because there is no advantage of having additions on diminishing length operands.An average iteration reduces the length of the longer operand by about 1.4 bits, so the left-and right shift algorithms don't differ much, in how many shift steps can be combined into one longer shift.The plus-minus right shift algorithm has the smallest number of iterations, its delayed halving variant can combine the largest number of shifts, so its running time becomes very close to that of the shifting Euclidean algorithm.
In each iteration the RSDH+− modular inverse algorithm needs to shift one of U or V, and double R or S the same many times, which give about the same amount of work as the modular multiplication performs, maybe even less.At the end we need to add-halve R, which makes the modular inverse slightly slower than one modular multiplication, but still faster than two.

Testing Relative Primality
We can simplify all of our shifting modular inverse algorithms if we only want to know whether the two arguments x, y are relative primes: leaving out all the calculations with R and S. In this case all the plus-minus right shift algorithms become the same, so the simplest one, RS+− is the best, with 0.3065 n 2 cost of bit-shifts and the same for subtractions.All together it is 0.6130 n 2 .SE3 is still slightly better, with a running time of 0.6088 n 2 .The modified left-shift algorithm LS3 takes 0.3967 n 2 clock cycles for the shift operations, and 0.3331 n 2 clock cycles for the subtract operations, which is only 9…19% more.When, in an application, not only relative primality has to be tested, but modular inverses have to be calculated as well, this little speed advantage might not justify the implementation of 2 different algorithms, so LS3 or SE3 should be used for both purposes (without computing R and S if not needed).

Further Optimizations Possibilities
There are countless possibilities to speed up the presented algorithms a little further.For example, when U and V become small (short), a table lookup could immediately finish the calculations.If only one of them becomes small, or there is a large difference of the lengths of U and V, we could perform a different algorithmic step, which is best tuned to this case on the particular computing 1/18/2006 platform.(Most of the time it is a traditional Euclidean reduction step.)We tried hundreds of ideas like these, but the little acceleration did not justify the increased complexity.Some of the presented speedup methods could have been applied already somewhere in the huge literature of algorithmic optimization, but we could not find the wining combinations of these optimization techniques for the modular inverse problem published anywhere.Many modifications accelerate one part of the algorithm while they slow down -or even invalidateother parts.We investigated hundreds of algorithmic changes, but only discussed here the original algorithms and those optimizations, which led to the largest speedups.

Working on the Ends
On some computational platforms speed increase can be achieved with delayed full update of the variables.See, for example [4], [5] or [7].It means working with MS and/or LS digits only, as long as we can determine the necessary reduction steps, and fix the rest of the digits only when more precision is needed.Speedup is achieved by the reduced number of data fetch-, and combined update operations on the middle digits.Unfortunately, the resulting algorithms are much more complex, less suitable for direct hardware implementations and the combined operations involve multiplications, what we wanted to avoid.In our computational model data load-store operations are free, so having fewer of them doesn't provide any speed advantage.

Hybrid Algorithms
The right shift-and the shifting Euclidean algorithms operate on the opposite ends of the numbers.At least when the modulus is odd, the reduction steps could be intermixed: check on a few bits on the appropriate end of the intermediate values which algorithm is expected to reduce the lengths the most, and perform one step of it.However, the right shift algorithms are so much slower, that the corresponding reduction is almost never expected to be faster than the shifting Euclidean reduction.This way we could not achieve any significant speedup, but the algorithms became complicated and convoluted.

Modular Division
If the initialization of S ← 1 is replaced by S ← d, we get da −1 as the result, as described in [22] for polynomials.In case the modular inverse is only needed once, and it is multiplied by another number, we could save that multiplication, like in elliptic curve cryptography.If the inverse is reused many times, like at signed digit exponentiation ( [15]), this trick does not improve performance.
We start with a full length S (=d) instead of length 1, so (S,R) do not gradually increase from length 1, but start at length n.Further steps are necessary to prevent them to grow too large.These more than double the total work updating (S,R) (but not (U,V)) at the left-shift and shifting Euclidean algorithms, all together 50…100% increase.The right shift algorithms do not change much, so modular divisions significantly reduce their performance lag.In general, it only pays doing divisions this way, when the underlying modular inverse algorithm is much faster than two modular multiplications (making a modular division faster than 3 modular multiplications).

Acknowledgment
The author thanks the anonymous reviewers for suggesting many corrections, improvements and the reorganization of the original manuscript.
/ y are the two minimal values.One of the c values is positive, the other is negative, likewise d.We pair the positive c with the negative d and vice versa to get the two sets of minimal factors.1/18/2006