<!-- HAPPRIME - examples.xml Introduction documentation section Paul Smith Copyright (C) 2008-2009 Paul Smith National University of Ireland Galway This file is part of HAPprime. HAPprime is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. HAPprime is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program. If not, see <http://www.gnu.org/licenses/>. $Id: examples.xml 356 2008-12-10 12:40:51Z pas $ --> <!-- ********************************************************** --> <Chapter Label="Examples"> <Heading>Examples</Heading> <Section> <Heading>Computing the mod &p; group cohomology</Heading> Let <M>G</M> be a group and &FF; be a field, and let &FG; be the group ring of &G; over &FF;. A free &FG;-resolution of the ground ring is an exact sequence of module homomorphisms <Alt Only="LaTeX"> <Display> \ldots \rightarrow M_{n+1} \rightarrow M_n \rightarrow M_{n-1} \rightarrow \ldots \rightarrow M_1 \rightarrow \mathbb{F}G \twoheadrightarrow \mathbb{F} </Display> </Alt> <Alt Not="LaTeX"> <Display> ... ---> M_(n+1) ---> M_n ---> M_(n-1) ---> ... ---> M_1 ---> FG --->> F </Display> </Alt> Where each <M>M_n</M> is a free &FG;-module and the image of <M>d_{n+1}: M_{n+1} \rightarrow M_{n}</M> equals the kernel of <M>d_n: M_n \rightarrow M_{n-1}</M> for all <M>n > 0</M>. The maps <M>d_n</M> are called boundary homomorphisms. In &HAPprime; we consider the case where &G; is a &p;-group and &FF; is the prime field <M>GF(p)</M>, and this is assumed from now on. <P/> If we now define the Abelian group <M>D_n</M> to be <M>Hom(M_n, \mathbb{F})</M>, the set of all homomorphisms <M>M_n \to \mathbb{F}</M>, we can obtain the dual of this sequence, which will be a cochain complex of Abelian group homomorphisms <Alt Only="LaTeX"> <Display> \ldots \leftarrow D_{n+1} \leftarrow D_n \leftarrow D_{n-1} \leftarrow \ldots \leftarrow D_1 \leftarrow \mathbb{F} \leftarrow \mathbb{F} </Display> </Alt> <Alt Not="LaTeX"> <Display><![CDATA[ ... <--- D_(n+1) <--- D_n ---> D_(n-1) <--- ... <--- D_1 <--- F <--- F ]]></Display> </Alt> Each group <M>D_n</M> will be isomorphic to <M>\mathbb{F}^{|M_n|}</M> where <M>|M_n|</M> is the rank of the module <M>M_n</M>. Unlike the resolution, this sequence will generally not be exact, but one can define the mod-&p; cohomology group of &G; at degree <M>n</M> to be <Alt Only="LaTeX"> <Display> H^n(G, \mathbb{F}) = \frac{ker(D_n \to D_{n+1})}{im(D_{n-1} \to D_n)} </Display> </Alt> <Alt Not="LaTeX"> <Display> H^n(G, F) = ker(D_n ---> D_(n+1)) / im(D_(n-1) ---> D_n) </Display> </Alt> for all <M>n > 0 </M>. As with the <M>D_n</M>, the mod-&p; cohomology groups will also be isomorphic to vector spaces over &FF;. In the case where the resolution <M>R</M> is minimal (where each module <M>M_n</M> has the minimal number of generators), the dimensions of the (co)homology groups <M>H^n(G, \mathbb{F})</M> are identical to the dimensions of the resolution modules <M>M_n</M>. The group cohomology (and the similar group homology) is an invariant of &G;, and does not depend on a particular free &FG;-resolution. <P/> In the general case, there are thus two stages to computing the group cohomology of <M>G</M> up to the <M>n</M>th cohomology group: <Enum> <Item>Compute <M>R</M>, a free &FG;-resolution for &FG;, with at least <M>n+1</M> terms.</Item> <Item>Construct the cochain complex <M>C</M> from <M>R</M> and compute the <M>n</M> homology groups of <M>C</M></Item> </Enum> For example, to calculate the 9th mod-&p; cohomology group of the 134th order 64 in the &GAP; small groups library (which is the Sylow 2-subgroup of the Mathieu group <M>M_{12}</M>), we can use the &HAPprime; function <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/> to compute 10 terms of a free &FG;-resolution for &G; and then use &HAP; functions to find the rank <M>b_9</M> of the cohomology group, which will be isomorphic to <M>\mathbb{F}^{b_9}</M>. Alternatively, since <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/> always returns a minimal resolution, the cohomology group dimensions can be read directly from the resolution. <!-- gap> SetInfoLevel(InfoHAPprime, 0); gap> G := SmallGroup(64, 134);; gap> # First construct a FG-resolution for the group G gap> R := ResolutionPrimePowerGroupRadical(G, 10); gap> # Convert this into a cochain complex (over the prime field with p=2) gap> C := HomToIntegersModP(R, 2); gap> # And get the rank of the 9th cohomology group gap> b9 := Cohomology(C, 9); gap> # Since R is a free resolution, the ranks of the cohomology groups gap> # are the same as the module ranks in R gap> ResolutionModuleRanks(R); --> <Example><![CDATA[ gap> G := SmallGroup(64, 134);; gap> # First construct a FG-resolution for the group G gap> R := ResolutionPrimePowerGroupRadical(G, 10); Resolution of length 10 in characteristic 2 for <pc group of size 64 with 6 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> # Convert this into a cochain complex (over the prime field with p=2) gap> C := HomToIntegersModP(R, 2); Cochain complex of length 10 in characteristic 2 . gap> # And get the rank of the 9th cohomology group gap> b9 := Cohomology(C, 9); 55 gap> gap> # Since R is a free resolution, the ranks of the cohomology groups gap> # are the same as the module ranks in R gap> ResolutionModuleRanks(R); [ 3, 6, 10, 15, 21, 28, 36, 45, 55, 66 ]]]></Example> </Section> <!-- ********************************************************** --> <Section Label="ExCohomology"> <Heading>Computing mod-&p; cohomology rings and their Poincaré series</Heading> The mod-&p; group cohomology of a &p;-group &G;, given a field <M>\mathbb{F} = GF(p)</M>, has a multiplicative structure, and so the group <Alt Only="LaTeX"> <Display> H^*(G, \mathbb{F}) = \bigoplus_{i=0}^\infty H^i(G, \mathbb{F}) </Display> </Alt> <Alt Not="LaTeX"> <Display> H^*(G, F) = H^0(G, F) + H^1(G, F) + H^2(G, F) + ... </Display> </Alt> is a ring. Since &Homring; is isomorphic to a vector space over &FF;, it is also an algebra, in fact a graded algebra: for elements <M>e \in H^n(G, \mathbb{F})</M> and <M>f \in H^m(G, \mathbb{F})</M>, the product <M>ef</M> is an element of <M>H^{m+n}(G, \mathbb{F})</M>. <P/> Some functions for investigating the ring structure of &Homring; using &GAP; are already provided by &HAP;, and also by Marcus Bishop's <Package>Crime</Package> package <URL>http://www.math.uic.edu/~marcus/Crime/</URL>. There have also been implementations using other systems, in particular, Jon Carlson has computed the cohomology rings for all 2-groups of order 64 and fewer using MAGMA (see <URL>http://www.math.uga.edu/~lvalero/cohointro.html</URL> for results) and David Green has calculated the same, and some of order 128, using C (see <URL>http://www.math.uni-wuppertal.de/~green/Coho_v2/index.html</URL> for results). <!-- ********************************************************** --> <Subsection Label="ExPresentation"> <Heading>A ring presentation for the mod &p; cohomology (up to degree <M>n</M>)</Heading> The cohomology ring &Homring; is an infinite vector space, but if all elements of degree higher than some degree <M>n</M> are ignored, a related finite algebra can be considered. Multiplication in this finite algebra can be represented by a multiplication table giving the product of each basis element in the vector space (up to products of degree <M>n</M>). The &HAP; function <Ref Func="ModPCohomologyRing" BookName="HAP"/> computes such a finite algebra representation for the mod-&p; cohomology ring of a &p;-group &G; for a given value of <M>n</M>. <P/> The full infinite-dimensional cohomology ring has a finite presentation as a quotient ring <Alt Only="LaTeX"> <Display> H^*(G, \mathbb{F}) = \frac{\mathbb{F}[x_1, x_2, \ldots, x_n]}{\langle I_1, I_2, \ldots, I_m \rangle} </Display> </Alt> <Alt Not="LaTeX"> <Display><![CDATA[ H^*(G, F) = F[x_1, x_2, ..., x_n] / (I_1, I_2, ..., I_m) ]]></Display> </Alt> where the polynomial ring indeterminates <M>x_i</M> each have an associated degree <M>d_i</M> and the <M>I_j</M> are relations which together generate an ideal in the ring. Given a finite algebra <M>A</M>, the &HAPprime; function <Ref Oper="ModPCohomologyRingPresentation" Label="for algebra"/> calculates a presentation for the ring, modulo any generating elements of degree higher than <M>n</M>. If &Homring; has no generators or relations in degrees higher than <M>n</M>, then a ring presentation for an algebra computed via <Ref Func="ModPCohomologyRing" BookName="HAP"/> followed by <Ref Oper="ModPCohomologyRingPresentation" Label="for algebra"/> will be the same as a ring presentation for &Homring;. <P/> We shall calculate a ring presentation for the group <C>G := SmallGroup(16, 3)</C>: <!-- gap> G := SmallGroup(16, 3);; gap> A := ModPCohomologyRing(G, 5); gap> ModPCohomologyRingPresentation(A); --> <Example><![CDATA[ gap> G := SmallGroup(16, 3);; gap> A := ModPCohomologyRing(G, 5); <algebra of dimension 34 over GF(2)> gap> ModPCohomologyRingPresentation(A); Graded algebra GF(2)[ x_1, x_2, x_3, x_4, x_5 ] / [ x_1*x_2, x_1^2, x_1*x_5, x_2^2*x_3+x_5^2 ] with indeterminate degrees [ 1, 1, 2, 2, 2 ]]]></Example> The object returned by <Ref Oper="ModPCohomologyRingPresentation" Label="for algebra"/> tells us that <Alt Only="LaTeX"> <Display> H^*(G, \mathbb{F}) = \frac{\mathbb{F}[x_1,x_2,x_3,x_4,x_5]} {\langle x_1^2, \; x_1x_2, \; x_1x_5, \; x_2^2x_3+x_5^2 \rangle} </Display> where the indeterminates <M>x_1</M> and <M>x_2</M> </Alt> <Alt Not="LaTeX"> <Display><![CDATA[ H^*(G, F) = F[v,w,x,y,z] / (v^2, vw, vz, w^2y+z^2) ]]></Display> where the indeterminates <M>v</M> and <M>w</M> </Alt> are in degree one and the rest are in degree two. This assumes, however, that there are no generating elements in any degree higher than five. See Section <Ref Sect="ExPresentationLHS"/> below for a method that also computes the maximum degree necessary to present the cohomology ring, and so avoids this limitation. <P/> As well as taking the algebra as an argument, the function <Ref Oper="ModPCohomologyRingPresentation" Label="for group"/> can also take a group or a free &FG;-resolution, and a value for <M>n</M>, in which case it performs the calculation of the algebra as well. </Subsection> <!-- ********************************************************** --> <Subsection Label="ExPresentationLHS"> <Heading>Calculating a provably-correct mod-&p; cohomology</Heading> Using the &HAP; function <Ref Func="ModPCohomologyRing" BookName="HAP"/>, &HAPprime; can calculate a presentation for the cohomology ring &Homring; as described in Section <Ref Sect="ExPresentation"/>) It is known (Evens 1961) that mod-&p; cohomology rings are finitely-generated, and hence given a sufficiently-large <M>n</M> the presentation calculated in the above manner will be the complete ring. &HAPprime; can compute a sufficient value for &n; (in fact, it will be near, and usually at, the minimum). &HAPprime; uses the original Evens proof constructively. For a group <M>G</M>, a Lyndon-Hochschild-Serre spectral sequence can be computed. This will converge, and the limiting sheet of the sequence will be a ring which is an associated graded ring of the mod-&p; cohomology ring &Homring;. This means that, while it will not necessarily be isomorphic to &Homring;, it will have the same additive structure, and the generators and relations will lie in the same degrees. Thus the maximum degree in the presentation for the last sheet of the spectral sequence is a sufficient value for <M>n</M>. The spectral sequence computation requires minimal resolutions and correct cohomology rings for two smaller groups (a central subgroup <M>N</M> and the quotient group <M>G/N</M>), but these rings can be in turn computed by the same process (and the cohomology rings for some small groups can simply be stated). The cohomology ring for &G; will thus be proved correct by induction. <P/> When called with just a group (i.e. no value for <M>n</M>), <Ref Oper="ModPCohomologyRingPresentation" Label="for group"/> performs this spectral sequence calculation to find <M>n</M> and then calculates the cohomology ring using this value for <M>n</M>. In the example below we use this function, and by setting the value of <Ref InfoClass="InfoHAPprime"/> to 1, we can see the details of the spectral sequence calculation. <!-- gap> SetInfoLevel(InfoHAPprime, 1); gap> G := SmallGroup(16, 3);; gap> A := ModPCohomologyRingPresentation(G); gap> # Now extract data about the presentation gap> BaseRing(A); gap> GeneratorsOfPresentationIdeal(A); gap> IndeterminateDegrees(A); gap> MaximumDegreeForPresentation(A); --> <!-- This should be an Example, but the autoloaded ResClasses package modifies Print for PolynomialRings, so the output differs if this package is available --> <Log><![CDATA[ gap> SetInfoLevel(InfoHAPprime, 1); gap> G := SmallGroup(16, 3);; gap> A := ModPCohomologyRingPresentation(G); #I E_2 = GF(2)[ x_1, x_2 ] x GF(2)[ x_3, x_4 ] #I with generator degrees [ 1, 1 ] and [ 1, 1 ] respectively #I d_2(x_1) = zero #I d_2(x_2) = zero #I d_2(x_3) = x_1*x_2 #I d_2(x_4) = x_1^2 #I E_3 = GF(2)[ x_5, x_6, x_7, x_8, x_9 ]/[ x_6*x_9, x_6^2, x_5*x_6, x_5^2*x\ _7+x_9^2 ] #I E_3 = GF(2)[ x_2, x_1, x_4^2, x_3^2, x_1*x_3+x_2*x_4 ]/[ x_1^2*x_3+x_1*x_\ 2*x_4, x_1^2, x_1*x_2, x_1^2*x_3^2 ] #I d_3(x_2) = zero #I d_3(x_1) = zero #I d_3(x_4^2) = zero #I d_3(x_3^2) = x_1^2*x_2+x_1*x_2^2 = 0*Z(2) mod I #I d_3(x_1*x_3+x_2*x_4) = zero #I E_4 = GF(2)[ x_10, x_11, x_12, x_13, x_14 ]/[ x_11*x_14, x_11^2, x_10*x_1\ 1, x_10^2*x_12+x_14^2 ] #I E_4 = GF(2)[ x_2, x_1, x_4^2, x_3^2, x_1*x_3+x_2*x_4 ]/[ x_1^2*x_3+x_1*x_\ 2*x_4, x_1^2, x_1*x_2, x_1^2*x_3^2 ] #I d_4(x_2) = zero #I d_4(x_1) = zero #I d_4(x_4^2) = zero #I d_4(x_3^2) = zero #I d_4(x_1*x_3+x_2*x_4) = zero #I E_inf = GF(2)[ x_10, x_11, x_12, x_13, x_14 ]/[ x_11*x_14, x_11^2, x_10*x\ _11, x_10^2*x_12+x_14^2 ] #I E_inf = GF(2)[ x_2, x_1, x_4^2, x_3^2, x_1*x_3+x_2*x_4 ]/[ x_1^2*x_3+x_1*\ x_2*x_4, x_1^2, x_1*x_2, x_1^2*x_3^2 ] #I Renaming indeterminates and sorting into increasing degree Graded algebra GF(2)[ x_1, x_2, x_3, x_4, x_5 ] / [ x_1*x_2, x_1^2, x_1*x_5, x_2^2*x_3+x_5^2 ] with indeterminate degrees [ 1, 1, 2, 2, 2 ] gap> # Now extract data about the presentation gap> BaseRing(A); GF(2)[x_1,x_2,x_3,x_4,x_5] gap> GeneratorsOfPresentationIdeal(A); [ x_1*x_2, x_1^2, x_1*x_5, x_2^2*x_3+x_5^2 ] gap> IndeterminateDegrees(A); [ 1, 1, 2, 2, 2 ] gap> MaximumDegreeForPresentation(A); 4 ]]></Log> This (the correct cohomology ring) is in fact the same as the presentation computed in Section <Ref Sect="ExPresentation"/> using <M>n=5</M> since there are no generators or relations of degree greater than four. </Subsection> <Subsection Label="ExPoincare"> <Heading>Computing Poincaré series</Heading> The Poincaré series for the mod-&p; cohomology ring &Homring; is the infinite series <Display> a_0 + a_1x + a_2x^2 + a_3x^3 + ... </Display> where <M>a_k</M> is the dimension of the vector space <M>H^k(G, \mathbb{F})</M>. The Poincaré function is a rational function <M>P(x)/Q(x)</M> which is equal to the Poincaré series. <P/> The ranks of the modules in a minimal resolution for a group &G; are identical to the dimensions <M>a_k</M>, so a minimal resolution can be used to calculate the Poincaré series without first calculating the cohomology ring. This is the method used by the &HAP; function <Ref Func="PoincareSeries" BookName="HAP"/>, but will only give the correct answer for a sufficiently-long resolution. &HAP; has a method for calculating a resolution that is likely to be long enough, but cannot prove that this is sufficient. <P/> &HAPprime; can instead calculate a provably-correct Poincaré series for a mod-&p; cohomology ring, and without first calculating the ring itself. The final sheet of a Lyndon-Hochschild-Serre spectral sequence for a group &G; will be a ring with the same additive structure as the cohomology ring for &G;. This will thus have the same Poincaré series, and can be used to provide the Poincaré series for the cohomology ring without having to also compute the cohomology ring. This is implemented in the &HAPprime; function <Ref Oper="PoincareSeriesLHS"/>. <P/> As well as being provably correct, <Ref Oper="PoincareSeriesLHS"/> is also often faster than the related &HAP; function, as demonstrated in this example: <!-- gap> G := SmallGroup(64, 73);; gap> # Compute the Poincare series using HAP gap> P1 := PoincareSeries(G);time; gap> # Compute the Poincare series using HAPprime gap> P2 := PoincareSeriesLHS(G);time; gap> P1 = P2; --> <Log><![CDATA[ gap> G := SmallGroup(64, 210);; gap> # Compute the Poincare series using HAP gap> P1 := PoincareSeries(G);time; (x_1^4+x_1^2+x_1+1)/(-x_1^7+3*x_1^6-5*x_1^5+7*x_1^4-7*x_1^3+5*x_1^2-3*x_1+1) 46434 gap> # Compute the Poincare series using HAPprime gap> P2 := PoincareSeriesLHS(G);time; (x_1^4+x_1^2+x_1+1)/(-x_1^7+3*x_1^6-5*x_1^5+7*x_1^4-7*x_1^3+5*x_1^2-3*x_1+1) 1889 gap> P1 = P2; true ]]></Log> In this case, &HAP; needs to compute 14 terms of a resolution for this group of order 64 before it is confident that it has a stable Poincaré series. By contrast &HAPprime;, in calculating the spectral sequence, needs to compute 5 terms of two resolutions of groups of order 8 and then construct a (non-minimal) resolution for &G; (also of length 5) from these two. Both methods give the same answer, but only &HAPprime;'s is guaranteed to be correct. </Subsection> </Section> <!-- ********************************************************** --> <Section Label="ExResolution"> <Heading>Comparing the memory usage and speed of &HAPprime; and &HAP;'s <K>ResolutionPrimePowerGroup</K> functions</Heading> For small &p;-groups, the group ring &FG; can be considered as a vector space of rank <M>|G|</M> with the elements of &G; as its basis elements. Each module <M>M_n</M> in a &FG;-resolution is also a vector space (of dimension <M>|M_n||G|</M>) and the boundary maps <M>d_n</M> can be represented as vector space homomorphisms. As a result, standard linear algebra techniques can be used to compute a minimal resolution by constructing a sequence of module homomorphisms where the kernel of one map is the image of the next, and where the modules have minimal generating sets. See Chapter <Ref Chap="Resolution" BookName="HAPprime Datatypes"/> in the datatypes manual for further details. <P/> As the groups get larger, this approach becomes less feasible due to the amount of time and memory needed to store and compute the null space of large matrices. The &HAP; function <Ref Func="ResolutionPrimePowerGroup" BookName="HAP"/> and the &HAPprime; functions <Ref Func="ResolutionPrimePowerGroupRadical" Label="for group"/> and <Ref Func="ResolutionPrimePowerGroupGF" Label="for group"/> all use this linear algebra approach, but the &HAPprime; functions are optimised to save memory, allowing the computation of resolutions which are longer, or are of larger groups, than are possible using &HAP; alone. <Subsection> <Heading>&HAPprime; takes less memory to store resolutions</Heading> Consider computing a resolution of a group of an arbitrary group of order 128, <C>G = SmallGroup(128, 844)</C> using &HAP;. Computation is performed on a dual-core Intel Core2Duo running at 2.66MHz, and the memory available to &GAP; is the standard initial allocation of 256Mb. <!-- gap> SetInfoLevel(InfoHAPprime, 0); gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroup(G, 9); gap> time; gap> # Can we construct a resolution of length ten? gap> R := ResolutionPrimePowerGroup(G, 10); --> <Log><![CDATA[ gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroup(G, 9); Resolution of length 9 in characteristic 2 for <pc group of size 128 with 7 generators> . gap> time; 27685 gap> # Can we construct a resolution of length ten? gap> R := ResolutionPrimePowerGroup(G, 10); exceeded the permitted memory (`-o' command line option) at res := SemiEchelonMatDestructive( List( mat, ShallowCopy ) ); called from SemiEchelonMat( NullspaceMat( BndMat ) ) called from ZGbasisOfKernel( i - 1 ) called from <function>( <arguments> ) called from read-eval-loop Entering break read-eval-print loop ... you can 'quit;' to quit to outer loop, or you can 'return;' to continue ]]></Log> The &HAPprime; function <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/> uses an almost identical algorithm, but stores its boundary maps more efficiently. As a result, with the same memory allowance: <!-- gap> SetInfoLevel(InfoHAPprime, 0); gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroupRadical(G, 9); gap> time; gap> # Can we construct a resolution of length ten? gap> R := ExtendResolutionPrimePowerGroupRadical(R);; gap> # Yes! How about eleven? gap> R := ExtendResolutionPrimePowerGroupRadical(R); gap> ResolutionModuleRanks(R); gap> gap> # But it will run out of memory if we try to go further gap> R := ExtendResolutionPrimePowerGroupRadical(R); --> <Log><![CDATA[ gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroupRadical(G, 9); Resolution of length 9 in characteristic 2 for <pc group of size 128 with 7 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> time; 25321 gap> # Can we construct a resolution of length ten? gap> R := ExtendResolutionPrimePowerGroupRadical(R);; gap> # Yes! How about eleven? gap> R := ExtendResolutionPrimePowerGroupRadical(R); Resolution of length 11 in characteristic 2 for <pc group of size 128 with 7 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> ResolutionModuleRanks(R); [ 3, 6, 11, 19, 30, 44, 62, 85, 113, 146, 185 ] gap> gap> # But it will run out of memory if we try to go to twelve terms gap> R := ExtendResolutionPrimePowerGroupRadical(R); exceeded the permitted memory (`-o' command line option) at ... ]]></Log> The &HAPprime; version can compute two further terms of the resolution, which given the sizes of the additional modules represents a considerable improvement. Just representing the homomorphism <M>d_{10}: (\mathbb{F}G)^{146} \to (\mathbb{F}G)^{113}</M> as vectors requires nearly as much memory again as representing the first nine homomorphisms. To compute and store the same resolution of length 11 using <Ref Func="ResolutionPrimePowerGroup" BookName="HAP"/> would need a little over three times the memory used here by &HAPprime;. The time taken by both versions is very similar. <P/> In the example above, note also the use of the &HAPprime; function <Ref Oper="ExtendResolutionPrimePowerGroupRadical"/>, which makes it much easier to add terms to an existing resolution. In standard &HAP;, if one decides that a resolution is too short and that more terms are required, then the entire resolution must be computed again from scratch. </Subsection> <Subsection> <Heading>&HAPprime; takes less memory to compute resolutions</Heading> The function <Ref Oper="ResolutionPrimePowerGroupGF" Label="for group"/> uses a new algorithm to compute the kernel of &FG;-module homomorphisms when &FG;-modules are represented using a set of &G;-generating vectors (see <Ref Chap="FG-module homomorphisms" BookName="HAPprime Datatypes"/> in the datatypes reference manual). This provides a further memory saving over <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/>, although at the cost of a much slower computation time: <!-- gap> SetInfoLevel(InfoHAPprime, 0); gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroupGF(G, 9); gap> time; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R); gap> ResolutionModuleRanks(R); gap> # But it will run out of (the inital 256Mb) of memory at sixteen terms --> <Log><![CDATA[ gap> G := SmallGroup(128, 844);; gap> R := ResolutionPrimePowerGroupGF(G, 9); Resolution of length 9 in characteristic 2 for <pc group of size 128 with 7 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> time; 422742 gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; gap> R := ExtendResolutionPrimePowerGroupGF(R);; Resolution of length 15 in characteristic 2 for <pc group of size 128 with 7 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> ResolutionModuleRanks(R); [ 3, 6, 11, 19, 30, 44, 62, 85, 113, 146, 185, 231, 284, 344, 412 ] gap> # But it will run out of (the inital 256Mb) of memory at sixteen terms ]]></Log> Using <Ref Oper="ResolutionPrimePowerGroupGF" Label="for group"/> we can get a further four terms of the resolution. For this resolution, this represents a memory saving of a factor of five over <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/> and fifteen over <Ref Oper="ResolutionPrimePowerGroup" BookName="HAP"/>, although it does take fifteen times as long as either of those just to compute the first nine terms, and scales less well with size. </Subsection> <Subsection> <Heading>Automatic selection of the best method</Heading> The two functions <Ref Oper="ResolutionPrimePowerGroupRadical" Label="for group"/> and <Ref Oper="ResolutionPrimePowerGroupGF" Label="for group"/> offer a trade-off between time and memory. The function <Ref Oper="ResolutionPrimePowerGroupAutoMem" Label="for group"/> automates the decision of which version to use, switching from the <C>Radical</C> to the <C>GF</C> version when it estimates that it is about to run out of available memory for the faster version. In this example, we have also increase the <Ref InfoClass="InfoHAPprime"/> info level to display progress information. At level two, the rank of each module in the resolution is displayed as it is calculated, giving an indication of progress. With this setting, the user is also notified when the <C>AutoMem</C> function switches, and the <C>GF</C> function displays a rolling estimate of its completion time (which is not shown since that output is overwritten when completed) <!-- gap> G := SmallGroup(128, 844);; gap> SetInfoLevel(InfoHAPprime, 2); gap> R := ResolutionPrimePowerGroupAutoMem(G, 15); gap> StringTime(time); --> <Log><![CDATA[ gap> G := SmallGroup(128, 844);; gap> SetInfoLevel(InfoHAPprime, 2); gap> R := ResolutionPrimePowerGroupAutoMem(G, 15); #I Dimension 2: rank 6 #I Dimension 3: rank 11 #I Dimension 4: rank 19 #I Dimension 5: rank 30 #I Dimension 6: rank 44 #I Dimension 7: rank 62 #I Dimension 8: rank 85 #I Dimension 9: rank 113 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 10: rank 146 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 11: rank 185 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 12: rank 231 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 13: rank 284 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 14: rank 344 #I Finding kernel of homomorphism by splitting: #I - Finding kernel of U #I - Finding kernel of V #I - Finding intersection of U and V #I - Finding intersection preimages #I Dimension 15: rank 412 Resolution of length 15 in characteristic 2 for <pc group of size 128 with 7 generators> . No contracting homotopy available. A partial contracting homotopy is available. gap> StringTime(time); " 5:45:53.613" ]]></Log> </Subsection> </Section> </Chapter>