BOLT BERANEK AND NEWMAN i K c CONSUITING DEVEIOPMENT RESEARCH THE STRUCTURE OP A LISP SYSTEM USING TWO-LEVEL STORAGE

Size: px
Start display at page:

Download "BOLT BERANEK AND NEWMAN i K c CONSUITING DEVEIOPMENT RESEARCH THE STRUCTURE OP A LISP SYSTEM USING TWO-LEVEL STORAGE"

Transcription

1 BOLT BERANEK AND NEWMAN i K c CONSUITING DEVEIOPMENT RESEARCH ^ APCRL » 0 G o / o ö O THE STRUCTURE OP A LISP SYSTEM USING TWO-LEVEL STORAGE Daniel G. Bobow Daniel L. Muphy Bolt Beanek and Newman Inc 50 Moulton Steet Cambidge, Massachusetts Contact No. AP19(628)-5065 Poject No. S668 Scientific Repot No. 6 k Novembe 1966 (The wok epoted was suppoted by the Advanced Reseach Pojects Agency, ARPA Ode No. 627, Amendment No. 2, dated 9 Mach 1965) Pepaed Fo: AIR FORCE CAMBRIDGE RESEARCH LABORATORIES OFFICE OF AEROSPACE RESEARCH UNITEP STATES AIR FORCE BEDFORD, MASSACHUSETTS Distibution of this document is unlimited CAMBRIDGE NEWYOR< CHICAGO I. OJANCEIES

2 AFCRL THE STRUCTURE OF A LISP SYSTEM USING TWO-LEVEL STORAGE Daniel G. Bobow Daniel L. Muphy Bolt Beanek and Newman Inc 50 Moulton Steet Cambidge, Massachusetts Contact No. AF19(628)-5065 Poject No Scientific Repot No. 6 4 Novembe 1966 (The wok epoted was suppoted by the Advanced Reseach Pojects Agency, ARPA Ode No. 627, Amendment No. 2, dated 9 Mach 1965) Pepaed Fo: AIR FORCE CAMBRIDGE RESEARCH LABORATORIES OFFICE OF AEROSPACE RESEARCH UNITED STATES AIR FORCE BEDFORD, MASSACHUSETTS Distibution of this document is unlimited

3 TABLE OF CONTENTS ABSTRACT - 1 SECTION I 1 INTRODUCTION SECTION IT 3 ORGANIZATION OP CORE MEMORY Segmentation of System Code Compiled LISP Functions. T 5 SECTION III 7 ORGANIZATION OP THE DRUM MEMORY JType Detemination of Pointes 8 Lltegl Acons., 8 Numeical Atoms 10 Constuction of Lists 11 I SECTION IV 14 VARIABLE BINDINGS AND THE PUSHDOWN LIST SECTION V 16 PERFORMANCE I SECTION VI 18 SUMMARY I ACKNOWLEDGEMENTS 19 REFERENCES 20 [ FIGURES I Fig. 1. Oganization of the Vitual Memoy 9 f

4 ABSTRACT In an Ideal list-pocessing system thee would be enough coe memoy to contain,a11 the data and pogams. This pape descibes a numbe of techniques used to build a LISP system which utilizes a dum fo its pincipal stoage medium, with a supisingly low time-penalty fo use of this slow stoage device. The techniques include caeful segmentation of system pogams, allocation of vitual memoy to allow addess aithmetic fo type detemination, and a special algoithm fo building easonably lineaized lists, A scheme Is descibed fo binding vaiables which is good in this envionment and allows fo complete compatibility between compiled and Intepeted pogams with no special declaations. -1-

5 i * [ SECTION I INTRODUCTION i ^- LISP is a list-pocessing language which Is being used extensively In eseach In atificial Intelligence. In the Ideal list-pocessing system, thee would be enough coe memoy to contain all the data that wee to be efeenced ove a long time peiod. In this case, a data efeence would take on the octe of one o two micoseconds, a speed typical of pesent coe memoies. When lage systems ae constucted, equiing upwads of one hunded o two hunded thousand wods, the cost of coe memoy usually becomes pohibitive. Memoy size equiements of this ode of magnitude ae, howeve, not the exception buy athe the ule fo many pojects planning eseach on natual language, speech pocessing, and a host of othe aeas. j i L Thus it becomes necessay to conside the use of bulk stoage memoy devices such as magnetic dums and discs. The poblem in using such devices is that, while thei data tansfe ate is sufficiently high (5-10 M-sec/wd), one must wait fo the data on the otating medium to come aound to the ead position. The aveage time to access data is onehalf the otation time, o typically about 17 to 33 milliseconds. To use a dum, the, as if it wee coe memoy, i.e. to ead single wods as they ae equied, would incease data efeence time by a facto of aound 20,000 making any list-pocessing system uselessly slow fo pactical poblems. -1-

6 As a fist step towad utilizing dum stoage efficiently, we oganize the dum into blocks of wods, o pages, and bing into coe an entie block of wods wheneve one fom that block was equied. As may be seen fom the timings above, the exta time equied to tansfe 200 to 300 wods instead of one, is negligible compaed to the access time. If multiple efeences ae made to a block once it has been moved into coe, then the speed of data efeences is inceased by a facto equal to the numbe of efeences made to a given page befoe anothe must be bought in. In this pape we descibe a numbe of techniques we have used in the BBN LISP system to maximize the numbe of Incoe efeences pe dum efeence. These techniques should incease the speed of opeation of any list-pocessing system embedded in a time-shaing system which uses paging to nup a lage vitual memoy stoed on a dum into a smalle coe memoy, e.g. the MULTICS system (3). -2-

7 SECTION II ORGANIZATION OF CORE MEMORY t Ou LISP system has been Implemented on a Digital Equipment Copoation PDP-1. This machine has a coe memoy of 16K (K = 1024) and a dum memoy of 8&C. Access time to one 18 bit wod of coe Is 5 micoseconds, and aveage access time to a wod on the dum Is 1? milliseconds. The PDP-1 has no Index egistes, floating point instuctions, special push-down-11st instuctions, o paging hadwae. It has an unusually lage collection of I/O devices, includig pape tape eade and punch, teletypes with eade and punch, mag-tape dives, high speed display and light pen, and a \ Gaphicon (RAND) tablet. The LISP system can communicate with all these i/o devices. f I I The BEN LISP system contains both an Intepete and a compile. The opeation of the compiled code is completely compatible with the Intepetation by the intepete of S-expession definitions of functions. One may un mixed sets consisting of functions which ae intepeted and othes which ae un afte being compiled. The scheme fo binding vaiables which allows this complete compatibility is discussed in detail below. The 16K of coe memoy must be allocated among pemanent code, list stuctue stoage, and compiled code. To this end we have dedicated 4K to compiled code, 4K to the -3-

8 supeviso and pemanent code, and 8K to list stuctue and pushdown list stoage. Since a list element consists of two 18 bit wods;, this 8K of memoy is equivalent to at most 4K LISP wods. Segmentation of System Code The pemanent code fo the system is well ove 10K in length itself. In ode to stay within the allotted ^K of memoy we have segmented the system code into 6 ovelays which have minimal inteaction oetween them, minimizing the numbe of swaps necessay between ovelays. We keep pemanently in coe the inteupt outines fo sevicing use on-line inteaction, and an elementay time-shaing supeviso (ou system allows a small numbe of uses - usually 2 o 3 - to use the machine as a LISP dedicated time-shaing system), The emainde of the 4K is used as swapping aea fo the feixowing ovelays: 1. Intepete and compiled code unne 2. Some non-citical special machine coded suboutines fo manipulating list stuctue 3. Input-Output and fomatting 4. A special package fo manipulating the athe stupid magnetic tape dives on the machine 5. The gabage collecto 6. An initialization package Since all segmentation and paging is done by softwae on this machine, we felt that explicit segmentation and complete swapping of ovelays was pefeable to calling suboutines off the dum as needed. This is in contadistinction to ou philosophy on unning compiled LISP code. Fo the system code we know what easonable segments ae; in the latte case the system could not know (and we did not want -4-

9 to put the onus on the use fo keeping ack of) segments that would fit in 4K of stoage. All ovelays ae absolute code and can theefoe un only in fixed locations. Compiled LISP Functions Pogams compiled fom S-expession definitions of LISP ^ functions ae stoed on ehe dum in elocatable fom. When unning, these functions ae contained in a ing buffe in coe tj about 3^00 wods, popely elocated. Let us conside what happens when a compiled function is called. The call contains a pointe to the atom which names the function (essentially the symbolic name of the function). This name is used to seach an in-coe tansfe vecto containing stating locations --f all outines in coe. The seach is done by "hashing" the name, ad seaching the tansfe vecto until the name o an empty space is found. We have found that with the ing buffe full, the tansfe vecto of 128 wods (64 name-stating location pais) is only 1 about l/4 full. Theefoe the aveage numbe of checks to detemine the pesence of a function in coe is only slightly 1 lagei than 1. If the function called is in coe, a tansfe is made to the stating addess given. Thus, in this case, 1 we have made no dum efeences to link functions. This pocedue could be impoved if we wanted to modify the code of the calling function to contain a Jump indiect though the tansfe vecto. Howeve, we pefe that the only addess binding Jone on compiled code be simple elocation. 1 If the function called is not in coe, we obtain its dum addess fom the function cell of its atom. We then ead into the ing buffe the fist page of the compiled code fo thlg function (i.e. fom its initial position to the end of the 256 wod block which constitutes a page on the dum). [ -5-

10 The fist wod of the pogam contains the length of the pogam. Successive pages ae ead In until the entie pogam Is In coe. If at any time we ovelap the end of the ing buff-, we stat again fom the beginning of the ing buffe with the beginning of the pogam. The tansfe vecto is updated by emoving enties fo functions which have been wholly o even patially ovewitten, (taking c»e to mak them popely fo the hash lookup) and a new enty is added fo the pogam. Retuns fom calls ae also made though the tansfe vecto. Thus, pogams which have called suboutinea may be ovewitten, In tnis case, the pogam will be ecalled fom the dum befoe the etun Is effected. Sets of functions seem to stabilize unde this system, and if all the pogams to be used fo a easonable peiod can fit in coe, they soon eside thee. -6-

11 SECTION III ORGANIZATION OF THE DRUM MEMORY LISP assumes that it Is opeating in an envionment containing 128K wods, that is fom 0 to 400,000 octal. Only 88K actually exist on the dum. The emaining potion of the addess space is used fo epesentation jf small integes between -32,76? and 32,767 (offset by 300,000 octal), as descibed below. All data stoage is contained within this vitual memoy, including list stuctue, compiled code, atom value cells, popety cells and function cells, pint name stoage and pushdown list stoage. This vitual memoy is divided into pages of 256 wods. Refeence to the vitual stoage ae made via an in-coe map which supplies «/he addess of the equied page if it is in coe, o taps to a supevisoy outine if the page is not in coe. This dum supevisoy outine selects an in-coe page, wites it back on the dum if it has been changed, and eads the equied page fom the dum. Closed suboutine efeences to an in-coe wod though the map takes appoximately 170 micoseconds (be- cause of the poo set of opeation codes on the machine, and the lack of an index egiste). A efeence to a wod not in coe, which must be obtained fom the dum, takes between 17 and 33 milliseconds. It takes the longe time if a page must be witten out on th> dum befoe the efeenced page can be ead in. Thus, it ally pays to minimize dum efeences. -7-

12 Type Detemination of Pointes In standad wholly-in-coe LISP systems the type of element a pointe Is efeencing (fo example, an atom vesus an S-expesslon} can only be detemined by looking at the Item Itself, which might equie a dum efeence. We avoid this unnecessay dum efeence by dividing the vitual memoy space into a numbe of aeas as shown In Pig. 1. As can be seen fom this map of stoage, simple aithmetic on the addess of a pointe will detemine its type. We chose to allocate stoage athe than povide an in-coe map of stoage aeas, because the map would take up valuable in-coe space, and evey aldlvlonal page of stoage that we can squeeze into coe educes he nutbe of dum efeences. Liteal Atoms When a sting of chaactes epesenting a liteal atom is ead in, a seach is made to detemine if an atom with the same pintname has been seen befoe. If so, a pointe to that atom is used fo the cuent atom. If not, a new atom is ceated. Thus, as in all LISP systems, a liteal atom has a unfoue epesentation. Pou cells (wods) ae associated with each liteal atom. These cells contain pointes tc the pint-name of the atom, the function which it identifies, its top level o global value, and its popety list. A pointe to an atom points to its value cell. Since these value cells occu in only one pat of the addess space, one can tell fom a pointe (addess) whethe o not it is pointing to a liteal atom. Instead of having the othe cells on the sa^.e page with the value cell, each is put in a sepaate space in a nosition computable fom the addess of the value cell. Sepaating value -8-

13 t f VIRTUAL MEMORY 400,000Q i HASH TABLE 3oo,ooo 8 m PRINT-NAMES 270,000 8 C H 03 PRINT-NAME POINTERS 230,0008 I E CO E C I FUNCTION CELLS 220,00C 8 210,000g PUSHDOWN LIST FULL WORDS 200,000 8 JL 170,000 8 cm PROPERTY LIST CELLS VALUE CELLS I LIST 160, ,000 8 STRUCTURE 1 COMPILED CODE io,ooo 8 Pig. 1. Ocn'.zatlon of the Vitual Memoy.

14 cells and function cells, fo example. Is useful because most uses will not use the same name fo a global vaiable as they will fo a function, and, theefoe. If the fou cells ae bought In wheneve any one was asked fo. It Is likely that the othe thee cells would neve be efeenced. Yet, they use up oom In coe which could be used fo othe stoage. Similaly, the pint-name pointes associated with atoms ae needed duing input and output, but aely duing a computation. Theefoe, duing computation these cells ae neve in coe. Numeical Atoms In 709^ LISP, numeical atoms (numbes) do not have a unique epesentation; that is, a numbe of diffeent pointes may be efeencing numbes with the same value. This implies thzt fo compaison of numbes, o fo aithmetic opeations, the value of the numbes mut be obtained, and compaison of numeical atoms cannot be Just a compaison of pointes. The values of numbes ae stoed in "full wod" space. In 7094 LISP, pointes to numbes in list stuctue do not point diectly to the values of the numbes in full wod space. Fo each numeical atom, a special wod fom fee stoage is acquied with bits set in this wod to indicate type, and a pointe to the value of the numbe in full wod space. A pointe in list stuctue which efeences a numbe points to this "heade" wod. Because type infomation is implicit in pointes (addesses) in ou LISP system, we do not need this exta level of indiection, and pointes to numbe values diectly addess fee wod space. This obviously saves possible dum efeences in aithmetic opeations and compaisons. In addition, we utilize the fact that not all addesses in the 17 bit addess space of the dum can legitimately appea as -10-

15 -s pointes In list stuctue. Pointes between 200,000 (octal) and ^00,000 (octal) ae, theefoe, used In the context of list stuctue to epesent diectly "small" integes between and (offset by 300,000 octal). Thus, a pointe of 300,003 (octal) occuing In a list Is the numbe 3. Again, eliminating a level of indiectness educes the numbe of dum efeences equied. Anothe advantage of using this fom fo 1 small Integes is that numbes in this ange ae epesented uniquely; thus, aithmetic compaisons can be done diectly on the pointes. In addition, no additional stoage is equied, and this educes the numbe of gabage collections that must be invoked. Taditionally, almost all the numeical opeations done in LISP ae on these small integes. Constuction of Lists I Caeful allocation of the addess space of vitual memoy alleviates only some of the poblems of list pocessing in a paging envionment. List stuctue, unless specifically and puposefully oganized, tends to become andom and thus defeats the advantages of the paging scheme. If only one-tenth of the existing list stuctue can eside in coe, and it is efeenced andomly, then nine out of ten data efeences, on the aveage, will equie a dum opeation. An examination of list-handling pocesses indicates that lists ae usually pocessed sequetially; that is, pogams geneally m poceed down the elements of a list by taking the CDR of the I successive tails. A pocess may be handling seveal lists at onoe, but will typically make numeous efeences to each. One of the best means of speeding up a paged LISP appeaed to be lineaizing lists and concentating them on as few pages as -11-

16 possible. To do this, a special CONS, o list-wod constucting suboutine, was witten. This attempts to assign a new list-wod on a page aleady In use by the list of which this new wod Is a pat. The algoithm Is descibed below. In constucting a new list pointe, a fee wod pai must be obtained fom a fee stoage list. Instead of keeping one lage fee stoage list as Is done on the 7094 vesion of LISP, we have a sepaate fee stoage list on each page. Thus, the system can detemine If a new pointe pai can be placed on a paticula page. Using these fee stoage lists, we now constuct a new pointe pai (a LISP dotted pai consisting of two 18 bit PDP-1 wods) accoding to the following algoithm: To constuct Z CONS [XjY]: 1. If Y Is not an atom and thee Is oom on the page with Y then place Z on this page. 2. Othewise, If X Is not an atom and thee Is oom on the page with X put Z on that page. 3. Othewise, If thee Is a page In coe with oom place Z on this page. 4«Othewise, place Z on a page In vitual memoy with oom. This page Is found by seaching an In-coe table which gives the Initial location of the fee stoage list on each page. If stoage Is available. This algoithm tends to minimize coss linkages between pages and to limit any single stuctue to a vey few pages. This has been bon out by tacing though a numbe of faily lage -12-

17 stuc' es and computing the numbe of page cossings. When wokiüg with such a stuctue. It is unlikely that one will make efeences to moe than a few pages fo a elatively long peiod of time. Since these pages can eside ii- coe, no dum efeences ae needed. Fo example, in enteing the definition of a function, the entie definition tends to appea on a single page. Thus, duing the intepetation of a function multiple dum efeences ae usually unnecessay. When fee stoage is exhausted, gabage collection is necessay. A numbe of gabage collection schemes (4) have been Invented and Implemented in vaious vesions of LISP. In some of these schemes, fee stoage is compacted by a folding pocess in which empty cells in lowe stoage ae filled with the content of cells in highe stoage. Howeve, this is a vey bad type of scheme fo a paging envionment, because this tends to effectively shuffle pointes, and make lists extend ove moe pages than ae necessay. In ou system, we simply mak used cells and collect stoage on each page as it stands. In addition to this standad gabage collection algoithm, we also utilize anothe scheme fo dumping onto seconday stoage (magnetic tape) a compacted epesentation of tho list stuctue in use in the system. This scheme (descibed in (4)) based on an algoithm fist suggested by Mavin Minsky (5) has the desiable popety that lists ae, in geneal, lineaized in the CDR diection. i [ [ i i -13-

18 SECTION IV VARIABLE BINDINGS AND THE PUSHDOWN LIST A numbe o schemes have been used in diffeent vesions of LISP fo stoing the values of vaiables. These include: 1. Stoing values on an association list paied with the vaiable names. 2. Stoing values on the popety list of the atom which is the name of the vaiable. 3. Stoing values in a special value cell associated with the atom name, putting old values on the pushdown list, and estoing these values when exiting fom a function. 4. Stoing values on the pushdown list. The fist thee schemes all have the popety that values ae scatteed thoughout list stuctue space, and, in geneal, in a paging envionment would equie efeences to many pages to detemine the value of a vaiable. This would be vey undesiable in ou system. In ode to avoid this scatteing, and possible excessive dum efeences, we utilize a vaiation on the fouth standad scheme, usually only used fo tansmitting values of aguments to compiled functions; that is. -14-

19 we place these values on the pushdown list. But since we use an intepete as well as a compile, the vaiable names must be kept. The pushdown list thus contains pais each consisting of a vaiable name and its value. The Intepete need only seach down the pushdown list fo the binding (value) of a vaiable. One advantage of this scheme is that the cuent top of the pushdown stack is usually in coe, and thus, dum efeences ae aely equied. Fee vaiables wok automatically in a way simila to the association list scheme. In fact, the pushdown list is a guaanteed linea vesion of the association list.» An additional advantage of this scheme is that it is completely compatible with compiled functions which pick up thei aguments on the pushdown list fom known positions, instead of doing a seach. To keep complete compatibility, ou compiled functions put the names of thei aganents on the pushdown list, although they do not use them to efeence vaiables. Thus, fee vaiables can be used between compiled and intepeted functions with no special declaations necessay. The names on the pushdown list ae also vey useful in debugging, fo they povide a complt symbolic backtace in case of eo. Thus, this technique, fo a small exta ovehead, minimizes dum efeences, povides symbolic debugging infonnation, and allows completely fee mixing of compiled and intepeted outines. With appopiate passing down of pushdown list pointes we can even achieve the same effect as the standad PUNARQ device on the 7094 fo functional aguments. This PUNARQ device peseves local context well enough to handle the Knuth's "Man and Boy" compile poblem (Algol Bulletin 18.1). -15-

20 SECTION V PERFORMANCE The poof of pogamming techniques is in the unning. We have made a numbe of measuements on ou system to test the validity of ou assumptions. To test the hypothecis that ou CONS algoithm was helping to minimize dum efeences, we an a computation fo about 35 minutes with a numbe of countes in vaious functions. The computation made was the compilation of appoximately 10 pages of LISP functions which define an elementay pogamming language called GHOST. We wee utilizing about 12,000 wods of fee stoage; coe can only hold about 3,500. Typical gabage collections eclaimed about 5,000 wods. In this time we pefomed 31,000 CONS's, and 150,000 CAR's and CDR's (not counting those done in gabage collection). Fo these CONS's and CAR's-CDR's only 5,500 dum efeences wee needed, o only about 2.5 pe cent of the cases. If stoage wee distibuted andomly, then the expected numbe of efeences should have anged between 50 pe cent and 70 pe cent depending on whethe the space in use fo list stuctue was neae 7,000 o 12,000 wods. With the 2.5 pe cent figue, we computed that the system spent 10 pe cent of its time waiting fo the dum. If the system h&d to go to the dum fo 50 pe cent of its data :^= 3^"'TMMiain '.-^^ -...

21 efeences. It would be spending well ove 90 pe cent of Its time waiting fo the dum. Extacting the same Infomation fom a numbe of othe uns fo diffeent uses of the system indicates that the dum is efeenced between 2 pe cent and 10 pe cent of the time. The types of Jobs being done by the use wee not ecoded. The highe pecentage was ealized when the use was involved in many small online inteactions with the system. Howeve, it is in Just this case that waiting fo the dum is least costly to the use, since most of his time is spent thinking, Anothe facet of the cost of using the dum should be mentioned. Even in poblems in which all the stoage needed can fit into coe, you. ae paying a pice fo the dum facility. Instead of diect efeences to coe, the system must go though a softwae map to detemine coe addesses. We compaed the unning time of one pogam on an in-coe LISP system,and a paged LISP system,on an SDS-9^0. Despite the fact that all the data fit in coe, this expeiment indicated that we ae losing a facto of about 2 in speed when we use the softwae map. Of couse, if a hadwae paging map wee available this poblem would vanish. -17-

22 SECTION VI SUMMARY In auminay, ou LISP system supassed ou expectations and Is poving a useful eseach tool fo a numbe of atificial intelligence pojects. Caeful segmentation of system code, aangement of data spaces by type, oganization of list stuctue, and special attention to the binding of vaiables all contibute to the success of the system. 18- i «msi

23 , ^ ä mmmm^ ACKNOWLEDOEKENTS The authos would like to thank L, Pete Deutsch of Bekeley, Califonia fo KTltlng (in LISP) the fist vesion of the compile used on this system, and fo a numbe of stimulating discussions. The wok epoted hee was suppoted by the Advanced Reseach Pojects Agency, ARPA Ode No. 627* Amendment No. 2- dated 9 Mach

24 PEPERENCES 1. McCathy, J. et ai, LISP 1.3 Pogammes Manual, M. I. T. Pess, Cambidge, Massachusetts, Bekeley, E. G. and D. G. Bobow (eds.). The Pogamming Language LISP, Its Opeation and Applications, M. I. T. Pess, Cambidge, Massachusetts, Cobato, P., E. Glase et ai, "The MULTICS System," Poc. PJCC, Spatan Pess, Baltimoe, Mayland, Bobow, D. G., "Stoage Management in LISP," Poc. IPIP Conf. on Symbol Manipulation Languages, Noth Holland. (In pepaation) 5. Minsky, M. L., "A LISP Gabage Collecto Algoithm Using Seial Seconday Stoage (ev.)," Memo 38, Atificial Instelllgence Goup, M. I. T., Cambidge, Massachusetts, Cohen, J., "A Use of Past and Sic Memoies in List Poc< ssing Languages," Comm. ACM, Januay

25 Iln-lflHRi f1f>d Secuity Classification DOCUMENT CONTROL DATA - R&D (Secuity elatsification of title, body of ah*tact and indexing annotatici muil 6«enteed uihen the oveall epot it '.ossified) I. OftloiNATiNO ACTK.v (Copoal» auttjx) Bolt Beanek and Newman, Incopoated 50 Moulton Steet to. REPORT SECURITY. AJIIFIC ATION.. IlnJaasifled IK OR&UP l REPORT T The Stuctue of a LISP System U3lng Two-Level Stoage 4. DKJCR.fTive MOTES (Type of epot and inclusive dates) Inteim Scientific Repot %. AUTHORTS; (Laet none, fiisl nans, initial) Bobow, Daniel G. Muphy, Daniel L.». RtRORT DATE ^ Ko\imbe 1966 AP19(o28)-5065 «Si-i c SVxo T fl 0 C e cn T^c ARPA Ode No. 627 Amend> 2 k. PROJECT mtxmkwx * No C. OOO ELEMENT 615Ü501R d. OOO (UMELEMENT n/a AVAILASILITY/LIMITATION NOTICE» 7a TOTAL NO. OF PACES TS NO. OF RCFt 20 6 a ORlOIMATOR'* REPORT NUHiERTS; Scientific Repot No. 6»b. othin mamy HO(SI (Any olkt numbes Aot may be assigned this epot) APCRL-66-77^ DISTRIBUTION OP THIS DOCUMENT IS UNLIMTED II. SUPPLEMENTARY NOTES Hq. APCRL, OAR(ORB) United States Ai Poce L. fi. Hanscom Fid., Bed &ilj& > II AMTRAtT ' ' ti SPONSORING MILITARY ACTIVITY Advanced Reseach Pojects fta. Agency In an Ideal list-pocessing system thee would be enough coe memoy to contain all the data and pogams. This pape descibes a nunbe of techniques used to build a LISP system which utilizes a dum fo its pincipal stoage medium, with a supisingly low time-penalty fo use of this slow stoage device. The techniques include caeful segmentation of system pogams, allocation of vitual memoy to allow addess aithmetic fo type detemination, and a special algoithm fo building easonably lineaized lists. A scheme is descibed fo binding vaiables which is good in this envionment and allows fo complete compatibility between compiled and intepeted pogams with no special declaations. DD FO ', " 1473 '"' I JAN ««"" Unclassified Secuity Classiiicmtion

26 Secuity Classification KEY WORDS ROLE WT LINK C List Pocessing LISP Paging Systems Dynamic Stoage Allocation Two Level Stoage Allocation INi/RUCTIONS I. ORIGINATING ACTIVITY: Ente the name and adde^a of tne contacto, aubconuacto, gantee. Depatment of Defenae activity o othe oganization (copoate autho) iaau.ng the epot. 2a. REPORT SECURITY CLASSIFICATION; Ente the oveall secuity claaaification of the epot. Indicate whethe "Resticted Data" ia included. Making ia to be in accodance with appopiate secuity egulations. 26. GROUP: Automatic downcadiig ia specified in DoD Diective and Amed Foces Indvstiel Manual. Ente the goup numbe. Alao, when applicable, ahow that optional makings have been uaed fo Coup 3 and Goup 4 aa authoized. 3. REPORT TITLE: Ente the complete epot title in all capital lettea. Titlea in all caaea ahould be unclassified. If a meaningful title cannot be»elected without clasaificalion, show title claaaification in all capitala in paenlheaia immediately following the title. 4. DESCRIPTIVE NOTES: If appopiate, ente the type of epot, e.g.. inteim, pogess, summay, annual, o, final. Give the inclusive dales when s specific epoting peiod ia coveed. 5. AUTHOR(S): Ente th- named) of autboha) aa ahown on o in the epot. Ente laat name, tumi name, middle initial. If militay, ahow ank and banch of sevice. The name of the pincipal autho ia anabaolute minimum equiement. 6. REPORT DATE: Eme the date oi the epot aa day, month, yea, o month, yea. If moe than one date appeaa on the epot, use date of publication. 7a. TOTAL NUMBER OF PACES: The total page count ahould follow nomal pagination pocedue», :.«., ente the numbe of pages containing infomation. 76. NUMBER OF REFERENCES: Ente the total numbe of efeences cited in the epot. 8a. CONTRACT OR GRANT NUMBER: If sppopiate, ente the applicable numb.' of the contact o gant unde which the epot waa witten c, I id. PROJECT NUMBER: Ente the appopiate militay depatment identification, auch aa poject numbe, aubpojfcl numbe, system numbe», taak numbe, etc. 9a. ORIGINATOR'S REPORT NUMBER(S): F..e the official epot numbe by which the document will be identified and contolled by the oiginating activity. This numbe must be unique to this epot. 96. OTHER REPORT NUMBERS): If the epot hea been aaaigned any othe epot numbes (eithe by the oiginato o by the spomo), alao ente thia numbefs). 10. AVAILABILITY/LIMITATION NOTICES: Ente any limitations on futhe dissemination of the epot, othe than those impoaed by secuity claaaification, using standad atatementa such aa: (1) "Qualified equestes may obtain copiea of this epot fom DDC." (2) "Foeign announcement and diaaemination of thia epot by DDC is not authoized." (3) "U S. Govenment agencies may obtain copiea of this epot diectly fom DDC. Othe qualified DDC uses ahall equest though (4) "U. S. militay agencies may obtain copies of thia epot diectly fom DDC. Othe qualified uses shall equest though (5) "All distibution of this epot is contolled. Qualified DDC uses isalt equest though If the epot haa been funished to the Office of Technical Sevices, Depatment of Comaece, fo sale to the public, indicate thia fact and ente the pice, if known. 11. SUPPLEMENTARY NOTES: Uae fo additional explanatoy notes. 12. SPONSORING MILITARY ACTIV.: Ente the name of the depatmental poject office o laboatoy sponsoing (paying fojihe eseach and development. Include addess. 13. ABSTRACT: Ente an abstact giving a bief and factual summay of the document indicative of the epot, even though it may also appea elsewhee in the body of the technical epot. If additional apace ia equied, a continuation aheet ahall be attached. It ia highly desiable that the abatact of classified epots he unclassified. Each paagaph of the abatact ahall end with an indication of the militay aecuity claaaification of the infomation in the paagaph, epesented ss (TS), (S), (C), o (V). Thee is no limitation on the length of the abatact. Howeve, the suggested length ia fom ISO to 225 wod». 14. KEY WORDS: Key wods ae technically meaningful tema o shot phaaea that chaacteize a epot and may be used aa index entiea fo cataloging the epot. Key woda must be selected so that no aecuity claaaification ia equied. Identifies, such aa equipment model deaianation, tade name, militay poject code name, ceogaphic localion, may be uaed as key wods but will be followed by an indication of technical context. The assignment of links, ules, and weight» is optional. Unclassified Secuity Clasaificacion

Any modern computer system will incorporate (at least) two levels of storage:

Any modern computer system will incorporate (at least) two levels of storage: 1 Any moden compute system will incopoate (at least) two levels of stoage: pimay stoage: andom access memoy (RAM) typical capacity 32MB to 1GB cost pe MB $3. typical access time 5ns to 6ns bust tansfe

More information

dc - Linux Command Dc may be invoked with the following command-line options: -V --version Print out the version of dc

dc - Linux Command Dc may be invoked with the following command-line options: -V --version Print out the version of dc - CentOS 5.2 - Linux Uses Guide - Linux Command SYNOPSIS [-V] [--vesion] [-h] [--help] [-e sciptexpession] [--expession=sciptexpession] [-f sciptfile] [--file=sciptfile] [file...] DESCRIPTION is a evese-polish

More information

DYNAMIC STORAGE ALLOCATION. Hanan Samet

DYNAMIC STORAGE ALLOCATION. Hanan Samet ds0 DYNAMIC STORAGE ALLOCATION Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 07 e-mail: hjs@umiacs.umd.edu

More information

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012

Journal of World s Electrical Engineering and Technology J. World. Elect. Eng. Tech. 1(1): 12-16, 2012 2011, Scienceline Publication www.science-line.com Jounal of Wold s Electical Engineeing and Technology J. Wold. Elect. Eng. Tech. 1(1): 12-16, 2012 JWEET An Efficient Algoithm fo Lip Segmentation in Colo

More information

A Memory Efficient Array Architecture for Real-Time Motion Estimation

A Memory Efficient Array Architecture for Real-Time Motion Estimation A Memoy Efficient Aay Achitectue fo Real-Time Motion Estimation Vasily G. Moshnyaga and Keikichi Tamau Depatment of Electonics & Communication, Kyoto Univesity Sakyo-ku, Yoshida-Honmachi, Kyoto 66-1, JAPAN

More information

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives

a Not yet implemented in current version SPARK: Research Kit Pointer Analysis Parameters Soot Pointer analysis. Objectives SPARK: Soot Reseach Kit Ondřej Lhoták Objectives Spak is a modula toolkit fo flow-insensitive may points-to analyses fo Java, which enables expeimentation with: vaious paametes of pointe analyses which

More information

Lecture Topics ECE 341. Lecture # 12. Control Signals. Control Signals for Datapath. Basic Processing Unit. Pipelining

Lecture Topics ECE 341. Lecture # 12. Control Signals. Control Signals for Datapath. Basic Processing Unit. Pipelining EE 341 Lectue # 12 Instucto: Zeshan hishti zeshan@ece.pdx.edu Novembe 10, 2014 Potland State Univesity asic Pocessing Unit ontol Signals Hadwied ontol Datapath contol signals Dealing with memoy delay Pipelining

More information

DYNAMIC STORAGE ALLOCATION. Hanan Samet

DYNAMIC STORAGE ALLOCATION. Hanan Samet ds0 DYNAMIC STORAGE ALLOCATION Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 074 e-mail: hjs@umiacs.umd.edu

More information

Controlled Information Maximization for SOM Knowledge Induced Learning

Controlled Information Maximization for SOM Knowledge Induced Learning 3 Int'l Conf. Atificial Intelligence ICAI'5 Contolled Infomation Maximization fo SOM Knowledge Induced Leaning Ryotao Kamimua IT Education Cente and Gaduate School of Science and Technology, Tokai Univeisity

More information

Shortest Paths for a Two-Robot Rendez-Vous

Shortest Paths for a Two-Robot Rendez-Vous Shotest Paths fo a Two-Robot Rendez-Vous Eik L Wyntes Joseph S B Mitchell y Abstact In this pape, we conside an optimal motion planning poblem fo a pai of point obots in a plana envionment with polygonal

More information

Getting Started PMW-EX1/PMW-EX3. 1 Rotate the grip with the RELEASE button pressed. Overview. Connecting the Computer and PMW-EX1/EX3

Getting Started PMW-EX1/PMW-EX3. 1 Rotate the grip with the RELEASE button pressed. Overview. Connecting the Computer and PMW-EX1/EX3 A PMW-EX1/PMW-EX3 Getting Stated Oveview This document descibes how to use the XDCAM EX Vesion Up Tool (heeafte Vesion Up Tool ) to upgade the PMW-EX1 and PMW-EX3 to vesion 1.20 (PMW-EX1) o vesion 1.10

More information

Detection and Recognition of Alert Traffic Signs

Detection and Recognition of Alert Traffic Signs Detection and Recognition of Alet Taffic Signs Chia-Hsiung Chen, Macus Chen, and Tianshi Gao 1 Stanfod Univesity Stanfod, CA 9305 {echchen, macuscc, tianshig}@stanfod.edu Abstact Taffic signs povide dives

More information

Communication vs Distributed Computation: an alternative trade-off curve

Communication vs Distributed Computation: an alternative trade-off curve Communication vs Distibuted Computation: an altenative tade-off cuve Yahya H. Ezzeldin, Mohammed amoose, Chistina Fagouli Univesity of Califonia, Los Angeles, CA 90095, USA, Email: {yahya.ezzeldin, mkamoose,

More information

Point-Biserial Correlation Analysis of Fuzzy Attributes

Point-Biserial Correlation Analysis of Fuzzy Attributes Appl Math Inf Sci 6 No S pp 439S-444S (0 Applied Mathematics & Infomation Sciences An Intenational Jounal @ 0 NSP Natual Sciences Publishing o Point-iseial oelation Analysis of Fuzzy Attibutes Hao-En hueh

More information

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES

RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES RANDOM IRREGULAR BLOCK-HIERARCHICAL NETWORKS: ALGORITHMS FOR COMPUTATION OF MAIN PROPERTIES Svetlana Avetisyan Mikayel Samvelyan* Matun Kaapetyan Yeevan State Univesity Abstact In this pape, the class

More information

IP Network Design by Modified Branch Exchange Method

IP Network Design by Modified Branch Exchange Method Received: June 7, 207 98 IP Netwok Design by Modified Banch Method Kaiat Jaoenat Natchamol Sichumoenattana 2* Faculty of Engineeing at Kamphaeng Saen, Kasetsat Univesity, Thailand 2 Faculty of Management

More information

MapReduce Optimizations and Algorithms 2015 Professor Sasu Tarkoma

MapReduce Optimizations and Algorithms 2015 Professor Sasu Tarkoma apreduce Optimizations and Algoithms 2015 Pofesso Sasu Takoma www.cs.helsinki.fi Optimizations Reduce tasks cannot stat befoe the whole map phase is complete Thus single slow machine can slow down the

More information

GARBAGE COLLECTION METHODS. Hanan Samet

GARBAGE COLLECTION METHODS. Hanan Samet gc0 GARBAGE COLLECTION METHODS Hanan Samet Compute Science Depatment and Cente fo Automation Reseach and Institute fo Advanced Compute Studies Univesity of Mayland College Pak, Mayland 07 e-mail: hjs@umiacs.umd.edu

More information

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes

Information Retrieval. CS630 Representing and Accessing Digital Information. IR Basics. User Task. Basic IR Processes CS630 Repesenting and Accessing Digital Infomation Infomation Retieval: Basics Thosten Joachims Conell Univesity Infomation Retieval Basics Retieval Models Indexing and Pepocessing Data Stuctues ~ 4 lectues

More information

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number.

(a, b) x y r. For this problem, is a point in the - coordinate plane and is a positive number. Illustative G-C Simila cicles Alignments to Content Standads: G-C.A. Task (a, b) x y Fo this poblem, is a point in the - coodinate plane and is a positive numbe. a. Using a tanslation and a dilation, show

More information

Towards Adaptive Information Merging Using Selected XML Fragments

Towards Adaptive Information Merging Using Selected XML Fragments Towads Adaptive Infomation Meging Using Selected XML Fagments Ho-Lam Lau and Wilfed Ng Depatment of Compute Science and Engineeing, The Hong Kong Univesity of Science and Technology, Hong Kong {lauhl,

More information

Lecture 8 Introduction to Pipelines Adapated from slides by David Patterson

Lecture 8 Introduction to Pipelines Adapated from slides by David Patterson Lectue 8 Intoduction to Pipelines Adapated fom slides by David Patteson http://www-inst.eecs.bekeley.edu/~cs61c/ * 1 Review (1/3) Datapath is the hadwae that pefoms opeations necessay to execute pogams.

More information

Lecture #22 Pipelining II, Cache I

Lecture #22 Pipelining II, Cache I inst.eecs.bekeley.edu/~cs61c CS61C : Machine Stuctues Lectue #22 Pipelining II, Cache I Wiewold cicuits 2008-7-29 http://www.maa.og/editoial/mathgames/mathgames_05_24_04.html http://www.quinapalus.com/wi-index.html

More information

UCB CS61C : Machine Structures

UCB CS61C : Machine Structures inst.eecs.bekeley.edu/~cs61c UCB CS61C : Machine Stuctues Lectue SOE Dan Gacia Lectue 28 CPU Design : Pipelining to Impove Pefomance 2010-04-05 Stanfod Reseaches have invented a monitoing technique called

More information

COEN-4730 Computer Architecture Lecture 2 Review of Instruction Sets and Pipelines

COEN-4730 Computer Architecture Lecture 2 Review of Instruction Sets and Pipelines 1 COEN-4730 Compute Achitectue Lectue 2 Review of nstuction Sets and Pipelines Cistinel Ababei Dept. of Electical and Compute Engineeing Maquette Univesity Cedits: Slides adapted fom pesentations of Sudeep

More information

User Specified non-bonded potentials in gromacs

User Specified non-bonded potentials in gromacs Use Specified non-bonded potentials in gomacs Apil 8, 2010 1 Intoduction On fist appeaances gomacs, unlike MD codes like LAMMPS o DL POLY, appeas to have vey little flexibility with egads to the fom of

More information

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set

Modelling, simulation, and performance analysis of a CAN FD system with SAE benchmark based message set Modelling, simulation, and pefomance analysis of a CAN FD system with SAE benchmak based message set Mahmut Tenuh, Panagiotis Oikonomidis, Peiklis Chachalakis, Elias Stipidis Mugla S. K. Univesity, TR;

More information

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes

On the Conversion between Binary Code and Binary-Reflected Gray Code on Boolean Cubes On the Convesion between Binay Code and BinayReflected Gay Code on Boolean Cubes The Havad community has made this aticle openly available. Please shae how this access benefits you. You stoy mattes Citation

More information

IP Multicast Simulation in OPNET

IP Multicast Simulation in OPNET IP Multicast Simulation in OPNET Xin Wang, Chien-Ming Yu, Henning Schulzinne Paul A. Stipe Columbia Univesity Reutes Depatment of Compute Science 88 Pakway Dive South New Yok, New Yok Hauppuage, New Yok

More information

Design considerations for an educational time-sharing system

Design considerations for an educational time-sharing system Design consideations fo an educational time-shaing system by ROBERT F. HARGRAVES, JR. Datmouth College Hanove, New Hampshie and ANDREWG. STEPHENSON Time-Shae Copoation Hanove, New Hampshie INTRODUCTION

More information

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE

A New and Efficient 2D Collision Detection Method Based on Contact Theory Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai MIAO, Jian XUE 5th Intenational Confeence on Advanced Mateials and Compute Science (ICAMCS 2016) A New and Efficient 2D Collision Detection Method Based on Contact Theoy Xiaolong CHENG, Jun XIAO a, Ying WANG, Qinghai

More information

Configuring RSVP-ATM QoS Interworking

Configuring RSVP-ATM QoS Interworking Configuing RSVP-ATM QoS Intewoking Last Updated: Januay 15, 2013 This chapte descibes the tasks fo configuing the RSVP-ATM QoS Intewoking featue, which povides suppot fo Contolled Load Sevice using RSVP

More information

User Visible Registers. CPU Structure and Function Ch 11. General CPU Organization (4) Control and Status Registers (5) Register Organisation (4)

User Visible Registers. CPU Structure and Function Ch 11. General CPU Organization (4) Control and Status Registers (5) Register Organisation (4) PU Stuctue and Function h Geneal Oganisation Registes Instuction ycle Pipelining anch Pediction Inteupts Use Visible Registes Vaies fom one achitectue to anothe Geneal pupose egiste (GPR) ata, addess,

More information

Lecture 27: Voronoi Diagrams

Lecture 27: Voronoi Diagrams We say that two points u, v Y ae in the same connected component of Y if thee is a path in R N fom u to v such that all the points along the path ae in the set Y. (Thee ae two connected components in the

More information

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM

A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Accepted fo publication Intenational Jounal of Flexible Automation and Integated Manufactuing. A VECTOR PERTURBATION APPROACH TO THE GENERALIZED AIRCRAFT SPARE PARTS GROUPING PROBLEM Nagiza F. Samatova,

More information

Module 6 STILL IMAGE COMPRESSION STANDARDS

Module 6 STILL IMAGE COMPRESSION STANDARDS Module 6 STILL IMAE COMPRESSION STANDARDS Lesson 17 JPE-2000 Achitectue and Featues Instuctional Objectives At the end of this lesson, the students should be able to: 1. State the shotcomings of JPE standad.

More information

Method of controlling access to intellectual switching nodes of telecommunication networks and systems

Method of controlling access to intellectual switching nodes of telecommunication networks and systems ISSN (e): 2250 3005 Volume 05 Issue 05 ay 2015 Intenational Jounal of Computational Engineeing eseach (IJCE) ethod of contolling access to intellectual switching nodes of telecommunication netwoks and

More information

All lengths in meters. E = = 7800 kg/m 3

All lengths in meters. E = = 7800 kg/m 3 Poblem desciption In this poblem, we apply the component mode synthesis (CMS) technique to a simple beam model. 2 0.02 0.02 All lengths in metes. E = 2.07 10 11 N/m 2 = 7800 kg/m 3 The beam is a fee-fee

More information

Optical Flow for Large Motion Using Gradient Technique

Optical Flow for Large Motion Using Gradient Technique SERBIAN JOURNAL OF ELECTRICAL ENGINEERING Vol. 3, No. 1, June 2006, 103-113 Optical Flow fo Lage Motion Using Gadient Technique Md. Moshaof Hossain Sake 1, Kamal Bechkoum 2, K.K. Islam 1 Abstact: In this

More information

Transmission Lines Modeling Based on Vector Fitting Algorithm and RLC Active/Passive Filter Design

Transmission Lines Modeling Based on Vector Fitting Algorithm and RLC Active/Passive Filter Design Tansmission Lines Modeling Based on Vecto Fitting Algoithm and RLC Active/Passive Filte Design Ahmed Qasim Tuki a,*, Nashien Fazilah Mailah b, Mohammad Lutfi Othman c, Ahmad H. Saby d Cente fo Advanced

More information

An Unsupervised Segmentation Framework For Texture Image Queries

An Unsupervised Segmentation Framework For Texture Image Queries An Unsupevised Segmentation Famewok Fo Textue Image Queies Shu-Ching Chen Distibuted Multimedia Infomation System Laboatoy School of Compute Science Floida Intenational Univesity Miami, FL 33199, USA chens@cs.fiu.edu

More information

Accelerating Storage with RDMA Max Gurtovoy Mellanox Technologies

Accelerating Storage with RDMA Max Gurtovoy Mellanox Technologies Acceleating Stoage with RDMA Max Gutovoy Mellanox Technologies 2018 Stoage Develope Confeence EMEA. Mellanox Technologies. All Rights Reseved. 1 What is RDMA? Remote Diect Memoy Access - povides the ability

More information

COSC 6385 Computer Architecture. - Pipelining

COSC 6385 Computer Architecture. - Pipelining COSC 6385 Compute Achitectue - Pipelining Sping 2012 Some of the slides ae based on a lectue by David Culle, Pipelining Pipelining is an implementation technique wheeby multiple instuctions ae ovelapped

More information

Illumination methods for optical wear detection

Illumination methods for optical wear detection Illumination methods fo optical wea detection 1 J. Zhang, 2 P.P.L.Regtien 1 VIMEC Applied Vision Technology, Coy 43, 5653 LC Eindhoven, The Nethelands Email: jianbo.zhang@gmail.com 2 Faculty Electical

More information

Multidimensional Testing

Multidimensional Testing Multidimensional Testing QA appoach fo Stoage netwoking Yohay Lasi Visuality Systems 1 Intoduction Who I am Yohay Lasi, QA Manage at Visuality Systems Visuality Systems the leading commecial povide of

More information

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS

ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS ANALYTIC PERFORMANCE MODELS FOR SINGLE CLASS AND MULTIPLE CLASS MULTITHREADED SOFTWARE SERVERS Daniel A Menascé Mohamed N Bennani Dept of Compute Science Oacle, Inc Geoge Mason Univesity 1211 SW Fifth

More information

Introduction To Pipelining. Chapter Pipelining1 1

Introduction To Pipelining. Chapter Pipelining1 1 Intoduction To Pipelining Chapte 6.1 - Pipelining1 1 Mooe s Law Mooe s Law says that the numbe of pocessos on a chip doubles about evey 18 months. Given the data on the following two slides, is this tue?

More information

Persistent Memory what developers need to know Mark Carlson Co-chair SNIA Technical Council Toshiba

Persistent Memory what developers need to know Mark Carlson Co-chair SNIA Technical Council Toshiba Pesistent Memoy what developes need to know Mak Calson Co-chai SNIA Technical Council Toshiba 2018 Stoage Develope Confeence EMEA. All Rights Reseved. 1 Contents Welcome Pesistent Memoy Oveview Non-Volatile

More information

The Java Virtual Machine. Compiler construction The structure of a frame. JVM stacks. Lecture 2

The Java Virtual Machine. Compiler construction The structure of a frame. JVM stacks. Lecture 2 Compile constuction 2009 Lectue 2 Code geneation 1: Geneating code The Java Vitual Machine Data types Pimitive types, including intege and floating-point types of vaious sizes and the boolean type. The

More information

User Group testing report

User Group testing report Use Goup testing epot Deliveable No: D6.10 Contact No: Integated Poject No. 506723: SafetyNet Aconym: SafetyNet Title: Building the Euopean Road Safety Obsevatoy Integated Poject, Thematic Pioity 6.2 Sustainable

More information

Obstacle Avoidance of Autonomous Mobile Robot using Stereo Vision Sensor

Obstacle Avoidance of Autonomous Mobile Robot using Stereo Vision Sensor Obstacle Avoidance of Autonomous Mobile Robot using Steeo Vision Senso Masako Kumano Akihisa Ohya Shin ichi Yuta Intelligent Robot Laboatoy Univesity of Tsukuba, Ibaaki, 35-8573 Japan E-mail: {masako,

More information

n If S is in convex position, then thee ae exactly k convex k-gons detemined by subsets of S. In geneal, howeve, S may detemine fa fewe convex k-gons.

n If S is in convex position, then thee ae exactly k convex k-gons detemined by subsets of S. In geneal, howeve, S may detemine fa fewe convex k-gons. Counting Convex Polygons in Plana Point Sets Joseph S. B. Mitchell a;1, Günte Rote b, Gopalakishnan Sundaam c, and Gehad Woeginge b a Applied Mathematics and Statistics, SUNY Stony Book, NY 11794-3600.

More information

Administrivia. CMSC 411 Computer Systems Architecture Lecture 5. Data Hazard Even with Forwarding Figure A.9, Page A-20

Administrivia. CMSC 411 Computer Systems Architecture Lecture 5. Data Hazard Even with Forwarding Figure A.9, Page A-20 Administivia CMSC 411 Compute Systems Achitectue Lectue 5 Basic Pipelining (cont.) Alan Sussman als@cs.umd.edu as@csu dedu Homewok poblems fo Unit 1 due today Homewok poblems fo Unit 3 posted soon CMSC

More information

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System

Slotted Random Access Protocol with Dynamic Transmission Probability Control in CDMA System Slotted Random Access Potocol with Dynamic Tansmission Pobability Contol in CDMA System Intaek Lim 1 1 Depatment of Embedded Softwae, Busan Univesity of Foeign Studies, itlim@bufs.ac.k Abstact In packet

More information

Lecture # 04. Image Enhancement in Spatial Domain

Lecture # 04. Image Enhancement in Spatial Domain Digital Image Pocessing CP-7008 Lectue # 04 Image Enhancement in Spatial Domain Fall 2011 2 domains Spatial Domain : (image plane) Techniques ae based on diect manipulation of pixels in an image Fequency

More information

A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Personification by Boulic, Thalmann and Thalmann.

A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Personification by Boulic, Thalmann and Thalmann. A Mathematical Implementation of a Global Human Walking Model with Real-Time Kinematic Pesonification by Boulic, Thalmann and Thalmann. Mashall Badley National Cente fo Physical Acoustics Univesity of

More information

You Are Here! Review: Hazards. Agenda. Agenda. Review: Load / Branch Delay Slots 7/28/2011

You Are Here! Review: Hazards. Agenda. Agenda. Review: Load / Branch Delay Slots 7/28/2011 CS 61C: Geat Ideas in Compute Achitectue (Machine Stuctues) Instuction Level Paallelism: Multiple Instuction Issue Guest Lectue: Justin Hsia Softwae Paallel Requests Assigned to compute e.g., Seach Katz

More information

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks

Spiral Recognition Methodology and Its Application for Recognition of Chinese Bank Checks Spial Recognition Methodology and Its Application fo Recognition of Chinese Bank Checks Hanshen Tang 1, Emmanuel Augustin 2, Ching Y. Suen 1, Olivie Baet 2, Mohamed Cheiet 3 1 Cente fo Patten Recognition

More information

HISTOGRAMS are an important statistic reflecting the

HISTOGRAMS are an important statistic reflecting the JOURNAL OF L A T E X CLASS FILES, VOL. 14, NO. 8, AUGUST 2015 1 D 2 HistoSketch: Disciminative and Dynamic Similaity-Peseving Sketching of Steaming Histogams Dingqi Yang, Bin Li, Laua Rettig, and Philippe

More information

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen

Accurate Diffraction Efficiency Control for Multiplexed Volume Holographic Gratings. Xuliang Han, Gicherl Kim, and Ray T. Chen Accuate Diffaction Efficiency Contol fo Multiplexed Volume Hologaphic Gatings Xuliang Han, Gichel Kim, and Ray T. Chen Micoelectonic Reseach Cente Depatment of Electical and Compute Engineeing Univesity

More information

Performance Optimization in Structured Wireless Sensor Networks

Performance Optimization in Structured Wireless Sensor Networks 5 The Intenational Aab Jounal of Infomation Technology, Vol. 6, o. 5, ovembe 9 Pefomance Optimization in Stuctued Wieless Senso etwoks Amine Moussa and Hoda Maalouf Compute Science Depatment, ote Dame

More information

Fifth Wheel Modelling and Testing

Fifth Wheel Modelling and Testing Fifth heel Modelling and Testing en Masoy Mechanical Engineeing Depatment Floida Atlantic Univesity Boca aton, FL 4 Lois Malaptias IFMA Institut Fancais De Mechanique Advancee ampus De lemont Feand Les

More information

The Dual Round Robin Matching Switch with Exhaustive Service

The Dual Round Robin Matching Switch with Exhaustive Service The Dual Round Robin Matching Switch with Exhaustive Sevice Yihan Li, Shivenda S. Panwa, H. Jonathan Chao Abstact Vitual Output Queuing is widely used by fixed-length highspeed switches to ovecome head-of-line

More information

Cardiac C-Arm CT. SNR Enhancement by Combining Multiple Retrospectively Motion Corrected FDK-Like Reconstructions

Cardiac C-Arm CT. SNR Enhancement by Combining Multiple Retrospectively Motion Corrected FDK-Like Reconstructions Cadiac C-Am CT SNR Enhancement by Combining Multiple Retospectively Motion Coected FDK-Like Reconstuctions M. Pümme 1, L. Wigstöm 2,3, R. Fahig 2, G. Lauitsch 4, J. Honegge 1 1 Institute of Patten Recognition,

More information

An Optimised Density Based Clustering Algorithm

An Optimised Density Based Clustering Algorithm Intenational Jounal of Compute Applications (0975 8887) Volume 6 No.9, Septembe 010 An Optimised Density Based Clusteing Algoithm J. Hencil Pete Depatment of Compute Science St. Xavie s College, Palayamkottai,

More information

Effects of Model Complexity on Generalization Performance of Convolutional Neural Networks

Effects of Model Complexity on Generalization Performance of Convolutional Neural Networks Effects of Model Complexity on Genealization Pefomance of Convolutional Neual Netwoks Tae-Jun Kim 1, Dongsu Zhang 2, and Joon Shik Kim 3 1 Seoul National Univesity, Seoul 151-742, Koea, E-mail: tjkim@bi.snu.ac.k

More information

Haptic Glove. Chan-Su Lee. Abstract. This is a final report for the DIMACS grant of student-initiated project. I implemented Boundary

Haptic Glove. Chan-Su Lee. Abstract. This is a final report for the DIMACS grant of student-initiated project. I implemented Boundary Physically Accuate Haptic Rendeing of Elastic Object fo a Haptic Glove Chan-Su Lee Abstact This is a final epot fo the DIMACS gant of student-initiated poject. I implemented Bounday Element Method(BEM)

More information

A Consistent, User Friendly Interface for Running a Variety of Underwater Acoustic Propagation Codes

A Consistent, User Friendly Interface for Running a Variety of Underwater Acoustic Propagation Codes Poceedings of ACOUSTICS 6 - Novembe 6, Chistchuch, New Zealand A Consistent, Use Fiendly Inteface fo Running a Vaiety of Undewate Acoustic Popagation Codes Alec J Duncan, Amos L Maggi Cente fo Maine Science

More information

FACE VECTORS OF FLAG COMPLEXES

FACE VECTORS OF FLAG COMPLEXES FACE VECTORS OF FLAG COMPLEXES ANDY FROHMADER Abstact. A conjectue of Kalai and Eckhoff that the face vecto of an abitay flag complex is also the face vecto of some paticula balanced complex is veified.

More information

Reachable State Spaces of Distributed Deadlock Avoidance Protocols

Reachable State Spaces of Distributed Deadlock Avoidance Protocols Reachable State Spaces of Distibuted Deadlock Avoidance Potocols CÉSAR SÁNCHEZ and HENNY B. SIPMA Stanfod Univesity We pesent a family of efficient distibuted deadlock avoidance algoithms with applications

More information

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation

A Minutiae-based Fingerprint Matching Algorithm Using Phase Correlation A Minutiae-based Fingepint Matching Algoithm Using Phase Coelation Autho Chen, Weiping, Gao, Yongsheng Published 2007 Confeence Title Digital Image Computing: Techniques and Applications DOI https://doi.og/10.1109/dicta.2007.4426801

More information

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH

SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH I J C A 7(), 202 pp. 49-53 SYSTEM LEVEL REUSE METRICS FOR OBJECT ORIENTED SOFTWARE : AN ALTERNATIVE APPROACH Sushil Goel and 2 Rajesh Vema Associate Pofesso, Depatment of Compute Science, Dyal Singh College,

More information

Reader & ReaderT Monad (11A) Young Won Lim 8/20/18

Reader & ReaderT Monad (11A) Young Won Lim 8/20/18 Copyight (c) 2016-2018 Young W. Lim. Pemission is ganted to copy, distibute and/o modify this document unde the tems of the GNU Fee Documentation License, Vesion 1.2 o any late vesion published by the

More information

Scaling Location-based Services with Dynamically Composed Location Index

Scaling Location-based Services with Dynamically Composed Location Index Scaling Location-based Sevices with Dynamically Composed Location Index Bhuvan Bamba, Sangeetha Seshadi and Ling Liu Distibuted Data Intensive Systems Laboatoy (DiSL) College of Computing, Geogia Institute

More information

GCC-AVR Inline Assembler Cookbook Version 1.2

GCC-AVR Inline Assembler Cookbook Version 1.2 GCC-AVR Inline Assemble Cookbook Vesion 1.2 About this Document The GNU C compile fo Atmel AVR isk pocessos offes, to embed assembly language code into C pogams. This cool featue may be used fo manually

More information

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer

Input Layer f = 2 f = 0 f = f = 3 1,16 1,1 1,2 1,3 2, ,2 3,3 3,16. f = 1. f = Output Layer Using the Gow-And-Pune Netwok to Solve Poblems of Lage Dimensionality B.J. Biedis and T.D. Gedeon School of Compute Science & Engineeing The Univesity of New South Wales Sydney NSW 2052 AUSTRALIA bbiedis@cse.unsw.edu.au

More information

ART GALLERIES WITH INTERIOR WALLS. March 1998

ART GALLERIES WITH INTERIOR WALLS. March 1998 ART GALLERIES WITH INTERIOR WALLS Andé Kündgen Mach 1998 Abstact. Conside an at galley fomed by a polygon on n vetices with m pais of vetices joined by inteio diagonals, the inteio walls. Each inteio wall

More information

On Error Estimation in Runge-Kutta Methods

On Error Estimation in Runge-Kutta Methods Leonado Jounal of Sciences ISSN 1583-0233 Issue 18, Januay-June 2011 p. 1-10 On Eo Estimation in Runge-Kutta Methods Ochoche ABRAHAM 1,*, Gbolahan BOLARIN 2 1 Depatment of Infomation Technology, 2 Depatment

More information

MIS to Prepress ICS. Version Date: File: ICS-MIS-Prepress-1.01.doc,.pdf. Origination & Prepress WG

MIS to Prepress ICS. Version Date: File: ICS-MIS-Prepress-1.01.doc,.pdf. Origination & Prepress WG MIS to Pepess ICS Vesion 1.01 Date: 2006-01-02 File: ICS-MIS-Pepess-1.01.doc,.pdf Oigination & Pepess WG Abstact This ICS defines the Inteface between the MIS and Pepess. It specifies the Pocesses fo a

More information

Query Language #1/3: Relational Algebra Pure, Procedural, and Set-oriented

Query Language #1/3: Relational Algebra Pure, Procedural, and Set-oriented Quey Language #1/3: Relational Algeba Pue, Pocedual, and Set-oiented To expess a quey, we use a set of opeations. Each opeation takes one o moe elations as input paamete (set-oiented). Since each opeation

More information

A Two-stage and Parameter-free Binarization Method for Degraded Document Images

A Two-stage and Parameter-free Binarization Method for Degraded Document Images A Two-stage and Paamete-fee Binaization Method fo Degaded Document Images Yung-Hsiang Chiu 1, Kuo-Liang Chung 1, Yong-Huai Huang 2, Wei-Ning Yang 3, Chi-Huang Liao 4 1 Depatment of Compute Science and

More information

= dv 3V (r + a 1) 3 r 3 f(r) = 1. = ( (r + r 2

= dv 3V (r + a 1) 3 r 3 f(r) = 1. = ( (r + r 2 Random Waypoint Model in n-dimensional Space Esa Hyytiä and Joma Vitamo Netwoking Laboatoy, Helsinki Univesity of Technology, Finland Abstact The andom waypoint model (RWP) is one of the most widely used

More information

A New Finite Word-length Optimization Method Design for LDPC Decoder

A New Finite Word-length Optimization Method Design for LDPC Decoder A New Finite Wod-length Optimization Method Design fo LDPC Decode Jinlei Chen, Yan Zhang and Xu Wang Key Laboatoy of Netwok Oiented Intelligent Computation Shenzhen Gaduate School, Habin Institute of Technology

More information

IBM Optim Query Tuning Offerings Optimize Performance and Cut Costs

IBM Optim Query Tuning Offerings Optimize Performance and Cut Costs IBM Optim Quey Tuning Offeings Optimize Pefomance and Cut Costs Saghi Amisoleymani Solution Achitect Integated Data Management amisole@us.ibm.com June 9, 2010 Disclaime Copyight IBM Copoation [cuent yea].

More information

Parallel processing model for XML parsing

Parallel processing model for XML parsing Recent Reseaches in Communications, Signals and nfomation Technology Paallel pocessing model fo XML pasing ADRANA GEORGEVA Fac. Applied Mathematics and nfomatics Technical Univesity of Sofia, TU-Sofia

More information

Using SPEC SFS with the SNIA Emerald Program for EPA Energy Star Data Center Storage Program Vernon Miller IBM Nick Principe Dell EMC

Using SPEC SFS with the SNIA Emerald Program for EPA Energy Star Data Center Storage Program Vernon Miller IBM Nick Principe Dell EMC Using SPEC SFS with the SNIA Emeald Pogam fo EPA Enegy Sta Data Cente Stoage Pogam Venon Mille IBM Nick Pincipe Dell EMC v6 Agenda Backgound on SNIA Emeald/Enegy Sta fo block Intoduce NAS/File test addition;

More information

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension

Segmentation of Casting Defects in X-Ray Images Based on Fractal Dimension 17th Wold Confeence on Nondestuctive Testing, 25-28 Oct 2008, Shanghai, China Segmentation of Casting Defects in X-Ray Images Based on Factal Dimension Jue WANG 1, Xiaoqin HOU 2, Yufang CAI 3 ICT Reseach

More information

Conversion Functions for Symmetric Key Ciphers

Conversion Functions for Symmetric Key Ciphers Jounal of Infomation Assuance and Secuity 2 (2006) 41 50 Convesion Functions fo Symmetic Key Ciphes Deba L. Cook and Angelos D. Keomytis Depatment of Compute Science Columbia Univesity, mail code 0401

More information

Modeling a shared medium access node with QoS distinction

Modeling a shared medium access node with QoS distinction Modeling a shaed medium access node with QoS distinction Matthias Gies, Jonas Geutet Compute Engineeing and Netwoks Laboatoy (TIK) Swiss Fedeal Institute of Technology Züich CH-8092 Züich, Switzeland email:

More information

How many times is the loop executed? middle = (left+right)/2; if (value == arr[middle]) return true;

How many times is the loop executed? middle = (left+right)/2; if (value == arr[middle]) return true; This lectue Complexity o binay seach Answes to inomal execise Abstact data types Stacks ueues ADTs, Stacks, ueues 1 binayseach(int[] a, int value) { while (ight >= let) { { i (value < a[middle]) ight =

More information

AUTOMATED LOCATION OF ICE REGIONS IN RADARSAT SAR IMAGERY

AUTOMATED LOCATION OF ICE REGIONS IN RADARSAT SAR IMAGERY AUTOMATED LOCATION OF ICE REGIONS IN RADARSAT SAR IMAGERY Chistophe Waceman (1), William G. Pichel (2), Pablo Clement-Colón (2) (1) Geneal Dynamics Advanced Infomation Systems, P.O. Box 134008 Ann Abo

More information

Automatically Testing Interacting Software Components

Automatically Testing Interacting Software Components Automatically Testing Inteacting Softwae Components Leonad Gallaghe Infomation Technology Laboatoy National Institute of Standads and Technology Gaithesbug, MD 20899, USA lgallaghe@nist.gov Jeff Offutt

More information

CS 2461: Computer Architecture 1 Program performance and High Performance Processors

CS 2461: Computer Architecture 1 Program performance and High Performance Processors Couse Objectives: Whee ae we. CS 2461: Pogam pefomance and High Pefomance Pocessos Instucto: Pof. Bhagi Naahai Bits&bytes: Logic devices HW building blocks Pocesso: ISA, datapath Using building blocks

More information

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM

ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM ADDING REALISM TO SOURCE CHARACTERIZATION USING A GENETIC ALGORITHM Luna M. Rodiguez*, Sue Ellen Haupt, and Geoge S. Young Depatment of Meteoology and Applied Reseach Laboatoy The Pennsylvania State Univesity,

More information

Hierarchically Clustered P2P Streaming System

Hierarchically Clustered P2P Streaming System Hieachically Clusteed P2P Steaming System Chao Liang, Yang Guo, and Yong Liu Polytechnic Univesity Thomson Lab Booklyn, NY 11201 Pinceton, NJ 08540 Abstact Pee-to-pee video steaming has been gaining populaity.

More information

4.2. Co-terminal and Related Angles. Investigate

4.2. Co-terminal and Related Angles. Investigate .2 Co-teminal and Related Angles Tigonometic atios can be used to model quantities such as

More information

Physical simulation for animation

Physical simulation for animation Physical simulation fo animation Case study: The jello cube The Jello Cube Mass-Sping System Collision Detection Integatos Septembe 17 2002 1 Announcements Pogamming assignment 3 is out. It is due Tuesday,

More information

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson

DEADLOCK AVOIDANCE IN BATCH PROCESSES. M. Tittus K. Åkesson DEADLOCK AVOIDANCE IN BATCH PROCESSES M. Tittus K. Åkesson Univesity College Boås, Sweden, e-mail: Michael.Tittus@hb.se Chalmes Univesity of Technology, Gothenbug, Sweden, e-mail: ka@s2.chalmes.se Abstact:

More information

Introduction to Engineering Seismology Lecture 18

Introduction to Engineering Seismology Lecture 18 Lectue 18: Intoduction to GIS, Integation of hazad paametes on GIS platfom; Final zonation map pepaation with case study of Bangaloe Topics Geogaphical Infomation System (GIS) Analytic Hieachy Pocess (AHP)

More information

Generalized Grey Target Decision Method Based on Decision Makers Indifference Attribute Value Preferences

Generalized Grey Target Decision Method Based on Decision Makers Indifference Attribute Value Preferences Ameican Jounal of ata ining and Knowledge iscovey 27; 2(4): 2-8 http://www.sciencepublishinggoup.com//admkd doi:.648/.admkd.2724.2 Genealized Gey Taget ecision ethod Based on ecision akes Indiffeence Attibute

More information