CSE 613: Parallel Programming. Lecture 10 ( Cache Performance of Divide-and-Conquer Algorithms )

Size: px
Start display at page:

Download "CSE 613: Parallel Programming. Lecture 10 ( Cache Performance of Divide-and-Conquer Algorithms )"

Transcription

1 CSE 613: Parallel Programmig Leture 10 ( Cahe Performae of Divide-ad-Coquer Algorithms ) Rezaul A. Chowdhury Deartmet of Comuter Siee SUN Stoy Brook Srig 2012

2 Moder Sigle Core Mahies Memory Hierarhy Cost of memory aess deeds o whether it s a (ahe)hitor miss Data i ahe may get evitedto make sae for ew data items Good erformae requires high loality i memory aesses

3 Multiores Itel Core Duo, Itel eo AMD Otero, AMD Athlo Itel Itaium 2 Itel Nehalem, AMD Bareloa

4 Assumtios Two Level Memory Hierarhy A sigle level of ahe (ahes) of size C(eah) oeted to a mai memory of ubouded size ad blok size B LRU (Least Reetly Used) Cahe Relaemet Poliy Whe a ew blok must be brought ito the ahe, but the ahe is full, the the ahe blok that was aessed least reetly is evited to make sae for the ew blok Automati Cahe Relaemet Doe automatially by the OS or the hardware Fully Assoiative Cahes A blok brought ito the ahe from memory a reside aywhere i the ahe

5 Parallel Cahig Model: Distributed Cahes 1 2 CPU CPU CPU rivate ahe (size = C) blok trasfer (size = B) rivate ahe (size = C) blok trasfer (size = B) mai memory rivate ahe (size = C) blok trasfer (size = B) Cofiguratio: roessig elemets a rivate aheof size Cfor eah roessig elemet a arbitrarily large global shared memory blok trasfer sizeb ( betwee ahes ad memory ) Cahe Performae Measure: umber of blok trasfers aross all ahes

6 Parallel Cahig Model: Shared Cahe 1 2 CPU CPU CPU shared ahe (size = C) blok trasfer (size = B) mai memory Cofiguratio: roessig elemets a shared aheof size C a arbitrarily large global shared memory C B, where Cis blok trasfer size Cahe Performae Measure: umber of blok trasfersbetwee the ahe ad the memory

7 Loality of Referee Satial Loality If a artiular memory loatio is aessed at a artiular time, the it is likely that earby memory loatios will also be aessed i the ear future. Take advatage of the blok size Bto load all memory loatios i the same blok ito the ahe whe a artiular memory loatio i that blok is aessed. Temoral Loality If a artiular memory loatio is aessed at a artiular time, the it is likely that the same memory loatio will be aessed agai i the ear future. Take advatage of the ahe size Cto retai memory loatios already loaded ito the ahe for future referees.

8 Iterative Matrix Multiliatio z = x y ij ik kj k= 1 z z z z z z z z z 1 2 x11 x12 x1 = x21 x22 x 2 x x x 1 2 y y y y y y y y y 1 2 Iter-MM ( Z,, ) {,, Z are matries, where is a ositive iteger } 1. for i 1 to do 2. for j 1 to do 3. Z[ i ][ j ] 0 4. for k 1 to do 5. Z[ i ][ j ] Z[ i ][ j ] + [ i ][ k ] [ k ][ j ]

9 Iterative Matrix Multiliatio Iter-MM ( Z,, ) {,, Z are matries, where is a ositive iteger } 1. for i 1 to do 2. for j 1 to do 3. Z[ i ][ j ] 0 4. for k 1 to do 5. Z[ i ][ j ] Z[ i ][ j ] + [ i ][ k ] [ k ][ j ] z z z z z z z z z 1 2 store i row-major order x11 x12 x1 = x21 x22 x 2 x x x 1 2 store i row-major order y y y y y y y y y 1 2 Eah iteratio of the forloo i lie 3 iurs Ο ahe misses. Cahe-omlexity of Iter-MM, Q Ο.

10 Iterative Matrix Multiliatio Iter-MM ( Z,, ) {,, Z are matries, where is a ositive iteger } 1. for i 1 to do 2. for j 1 to do 3. Z[ i ][ j ] 0 4. for k 1 to do 5. Z[ i ][ j ] Z[ i ][ j ] + [ i ][ k ] [ k ][ j ] z z z z z z z z z 1 2 store i row-major order x11 x12 x1 = x21 x22 x 2 x x x 1 2 store i olum-major order y y y y y y y y y 1 2 Eah iteratio of the forloo i lie 3 iurs Ο 1 Cahe-omlexity of Iter-MM, Q Ο 1 ahe misses. Ο.

11 Blok Matrix Multiliatio ahe ( size = C ) C/3 C/3 C/3 m m Z C/3 m C/3 m C/3 m m = Blok-MM(,, Z ) 1. for i 1 to / m do 2. for j 1 to / m do 3. for k 1 to / m do 4. Iter-MM( ik, kj, Z ij )

12 m m Blok Matrix Multiliatio Blok-MM(,, Z ) 1. for i 1 to / m do 2. for j 1 to / m do 3. for k 1 to / m do 4. Iter-MM( ik, kj, Z ij ) Choose 3, so that ik, kj adz ij just fit ito the ahe. The lie 4 iurs Θ 1 ahe misses. Cahe-omlexity of Blok-MM [assumig a tall ahe, i.e., ] Θ Θ Θ Θ ( Otimal: Hog & Kug, STOC 81 )

13 m m Blok Matrix Multiliatio Blok-MM(,, Z ) 1. for i 1 to / m do 2. for j 1 to / m do 3. for k 1 to / m do 4. Iter-MM( ik, kj, Z ij ) Choose 3, so that ik, kj adz ij just fit ito the ahe. Otimal for ay algorithm that erforms the oeratios give by the The lie 4 iurs Θ 1 followig defiitio of matrix multiliatio: ahe misses. Cahe-omlexity z of ij Blok-MM xiky kj[assumig a tall ahe, i.e., ] Θ = k = 1 Θ Θ Θ ( Otimal: Hog & Kug, STOC 81 )

14 s Multile Levels of Cahe s Blok-MM(,, Z ) 1. for i 1 to / s do 2. for j 1 to / s do 3. for k 1 to / s do 4. Iter-MM( ik, kj, Z ij )

15 t s Multile Levels of Cahe t s Blok-MM(,, Z ) 1. for i 1 1 to / s do 2. for j 1 1 to / s do 3. for k 1 1 to / s do 4. for i 2 1 to s / t do 5. for j 2 1 to s / t do 6. for k 2 1 to s / t do 7. Iter-MM( ( i1 k 1 ) i2 k 2, ( k1 j 1 ) k2 j 2, (Z i1 j 1 ) i2 j 2 )

16 t s Multile Levels of Cahe t s Oe Parameter Per Cahig Level! Blok-MM(,, Z ) 1. for i 1 1 to / s do 2. for j 1 1 to / s do 3. for k 1 1 to / s do 4. for i 2 1 to s / t do 5. for j 2 1 to s / t do 6. for k 2 1 to s / t do 7. Iter-MM( ( i1 k 1 ) i2 k 2, ( k1 j 1 ) k2 j 2, (Z i1 j 1 ) i2 j 2 )

17 Parallel Reursive MM Par-Re-MM ( Z,, ) {,, Z are matries, where = 2 k for iteger k 0 } 1. if = 1 the 2. Z Z + 3. else 4. saw Par-Re-MM ( Z 11, 11, 11 ) 5. saw Par-Re-MM ( Z 12, 11, 12 ) 6. saw Par-Re-MM ( Z 21, 21, 11 ) 7. Par-Re-MM ( Z 21, 21, 11 ) 8. sy 9. saw Par-Re-MM ( Z 11, 12, 21 ) 10. saw Par-Re-MM ( Z 12, 12, 22 ) 11. saw Par-Re-MM ( Z 21, 22, 21 ) 12. Par-Re-MM ( Z 22, 22, 22 ) 13. sy 14. edif Work: Θ 1, 1, 8 2 Θ 1,. Sa: Θ [ MT Case 1 ] Θ 1, 1, 2 2 Θ 1,. Θ [ MT Case 1 ] Parallel Ruig Time: Ο Parallelism: Ο Θ Additioal Sae: Θ 1

18 Parallel Reursive MM o a Sigle Core Par-Re-MM ( Z,, ) {,, Z are matries, where = 2 k for iteger k 0 } 1. if = 1 the 2. Z Z + 3. else 4. saw Par-Re-MM ( Z 11, 11, 11 ) 5. saw Par-Re-MM ( Z 12, 11, 12 ) 6. saw Par-Re-MM ( Z 21, 21, 11 ) 7. Par-Re-MM ( Z 21, 21, 11 ) 8. sy 9. saw Par-Re-MM ( Z 11, 12, 21 ) 10. saw Par-Re-MM ( Z 12, 12, 22 ) 11. saw Par-Re-MM ( Z 21, 22, 21 ) 12. Par-Re-MM ( Z 22, 22, 22 ) 13. sy Cahe Comlexity: Ο,, 8 2,. Θ 1, whe Ω 14. edif

19 Parallel Reursive MM o Distributed Cahes Par-Re-MM ( Z,, ) {,, Z are matries, where = 2 k for iteger k 0 } 1. if = 1 the 2. Z Z + 3. else 4. saw Par-Re-MM ( Z 11, 11, 11 ) 5. saw Par-Re-MM ( Z 12, 11, 12 ) 6. saw Par-Re-MM ( Z 21, 21, 11 ) 7. Par-Re-MM ( Z 21, 21, 11 ) 8. sy 9. saw Par-Re-MM ( Z 11, 12, 21 ) 10. saw Par-Re-MM ( Z 12, 12, 22 ) 11. saw Par-Re-MM ( Z 21, 22, 21 ) 12. Par-Re-MM ( Z 22, 22, 22 ) 13. sy 14. edif Cahe Comlexity: 4 Assumtio:if 1, the is evely distributed amog the simultaeously sawed futios., 1, 8 2,. Θ, whe Ω

20 The Logest Commo Subsequee (LCS) Problem A subsequeeof a sequee is obtaied by deletig zero or more symbols from. Examle: = abba Z = ba obtaied by deletig the 1 st a ad the 2 d b from A Logest Commo Subsequee (LCS)of two sequee ad is a sequee Z that is a subsequee of both ad, ad is the logest amog all suh subsequees. Give ad, the LCS roblemasks for suh a Z.

21 The Logest Commo Subsequee (LCS) Problem Give: = x 1 x 2 x ad = y 1 y 2 y Fills u a array [0, 0 ] usig the followig reurree. 0 if i = 0 j = 0, [ i, j] = [ i 1, j 1] + 1 if i, j > 0 x i = y j, max { [ i, j 1], [ i 1, j] } otherwise. y 1 y 2 y 3 y 4 y 5 y 6 y 7 y 8 x 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 [ i, j ] Traebak Path (LCS) [, ] = legth of LCS Loal Deedey: value of eah ell deeds oly o values of adjaet ells.

22 The Logest Commo Subsequee (LCS) Problem The lassi ( iterative ) serial LCS DPrus i Θ time, uses Θ sae, ad iurs Θ ahe misses. Ay algorithm usig Θ sae must iur Ω ahe misses. Hee i order to redue the ahe omlexity of the LCS algorithm from Θ we must first redue its sae usage below Θ.

23 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q Q traebak ath

24 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q Q traebak ath

25 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q Q traebak ath

26 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q Q traebak ath

27 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q Q traebak ath

28 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. Q 11 Q 12 Q 21 Q 22 Q traebak ath

29 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

30 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

31 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

32 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

33 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

34 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

35 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) 3. Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

36 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) 3. Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

37 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) 3. Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

38 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) 3. Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

39 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) 3. Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) Q 11 Q 12 Q 21 Q 22 Q traebak ath

40 Sequetial Cahe-effiiet LCS Algorithm Q [1, 1 ] = 2 q 1. Deomose Q: Slit Q ito four quadrats. 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of the quadrats reursively. ( of at most 3 quadrats ) Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat fragmets of the traebak ath from the quadrats reursively. ( from at most 3 quadrats ) 4. Comose Traebak Path: Combie the ath fragmets. Q 21 Q 22 Q traebak ath

41 Cahe omlexity: Cahe Performae Ο 1,, Ο 1,. where is the ahe omlexity of reursive boudary geeratio (i the forward ass ): Ο 1,, 4 2 Ο 1,. Ο 1 Substitutig, Ο 1

42 Cahe omlexity: Cahe Performae Ο 1,, Ο 1,. where is the ahe omlexity of reursive boudary geeratio (i the forward ass ): Ο 1,, 4 2 Ο 1,. Ο 1 Substitutig, Ο 1 otimal

43 Parallel Cahe-effiiet LCS Algorithm for Distributed Cahes

44 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C

45 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C

46 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C

47 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah.

48 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

49 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

50 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

51 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

52 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

53 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

54 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

55 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

56 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

57 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

58 Parallel Cahe-effiiet Boudary Comutatio ( PAR-BOUNDAR ) Q [1, 1 ] > C 1. Deomose Q: Slit Q ito 2 submatries of size ( / ) ( / ) eah. 2. Geerate Boudaries: I iteratio i [1, 2-1], solve all submatries o the i-th forward diagoal i arallel usig the sequetial ahe-oblivious algorithm. For eah ell also omute: the ell o the iut boudary where the traebak ath through the give ell itersets.

59 Performae Bouds ( PAR-BOUNDAR ) Q [1, 1 ] Parallel Time Comlexity: > C Ο Ο Cahe Comlexity: Ο Ο Ο 1 Ο

60 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] > C Q traebak ath

61 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C Q 11 Q 12 Q 21 Q 22 Q traebak ath

62 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

63 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

64 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

65 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

66 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

67 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

68 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q 12 Q 21 Q 22 Q traebak ath

69 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Q 21 Q 22 Q traebak ath

70 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Q 21 Q 22 Q traebak ath

71 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Q 21 Q 22 Q traebak ath

72 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Q 21 Q 22 Q traebak ath

73 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat ath fragmets from Q 22, Q 12 ad Q 11 i arallel by allig PAR-TRACEBACK with / 3 roessors eah. Q 21 Q 22 Q traebak ath

74 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat ath fragmets from Q 22, Q 12 ad Q 11 i arallel by allig PAR-TRACEBACK with / 3 roessors eah. Q 21 Q 22 Q traebak ath

75 Parallel Cahe-effiiet Traebak Path ( PAR-TRACEBACK ) Q [1, 1 ] 1. Deomose Q: Slit Q ito four quadrats. > C 2. Forward Pass ( Geerate Boudaries ): Geerate the right ad the bottom boudaries of all quadrats by allig PAR-BOUNDAR ( usig all roessors ). Q 11 Q Bakward Pass ( Extrat Traebak Path Fragmets ): Extrat ath fragmets from Q 22, Q 12 ad Q 11 i arallel by allig PAR-TRACEBACK with / 3 roessors eah. Q 21 Q Comose Traebak Path: Combie the ath fragmets. Q traebak ath

76 Performae Bouds ( PAR-TRACEBACK ) Q [1, 1 ] Parallel Time Comlexity: 4 Ο > C Ο Q 11 Q 12 Cahe Comlexity: Ο 1 Ο Q 21 Q 22 Ο Q traebak ath

77 DP with Loal Deedeies Geeralizatio of the LCS Result Problem Time Sae Cahe-omlexity Parallel Time Classi Logest Commo Subsequee Pairwise Sequee Aligmet ( affie ga osts ) Θ ( 2 ) Θ ( ) Ο 2 BC Ο 2 B Gotoh, 1982 Ο 2 B Ο 2 + Media of three Sequees ( affie ga osts ) RNA Seodary Struture Preditio with Simle Pseudokots Θ ( 3 ) Θ ( 4 ) Θ ( 2 ) Ο Ο 3 B C 4 B C Kudse, 2003 Ο 3 B Akutsu, 2000 Ο 4 B Ο Ο log = sequee legth, C = ahe size, B = blok trasfer size, = #roessors 77

78 Performae of Cahe-effiiet Serial LCS Algorithms omared: The ahe-effiiet LCS algorithm ( CO ) Hirshberg s liear-sae LCS algorithm ( Hi ) Comutig Eviromet: Arhiteture Proessor Seed L1 Cahe ( B ) L2 Cahe ( B ) RAM Itel eo 3 GHz 8 KB ( 64 B ) 512 KB ( 64 B ) 4 GB AMD Otero 2.4 GHz 64 KB ( 64 B ) 1 MB ( 64 B ) 4 GB SUN Blade 1 GHz 64 KB ( 32 B ) 8 MB ( 512 B ) 1 GB

79 Ratio of Ruig Times o Radom Sequees (Hirshberg vs the Cahe-effiiet Algorithm)

80 Ratio of L1 Misses o Radom Sequees (Hirshberg vs the Cahe-effiiet Algorithm)

81 Cahe Performae of Divide-ad-Coquer Algorithms uder the Work-Stealig Sheduler

82 Series-Parallel DAG base ase serial omositio arallel omositio

83 Assumtios s ( fork / saw ) G 1 G 2 t ( joi / sy ) Two-way Divisio ( Saw ): Eah divisio geerates oly two subtasks. Serial Exeutio: The left (first) subtask geerated by a fork ode is always exeuted first. Parallel Exeutio: Oly the right (seod) subtask geerated by a fork ode a be stole. Drifted Nodes: I a arallel exeutio we say that a ode is driftedwhe it is exeuted o a differet roessig elemet tha its redeessor i the serial exeutio.

84 Observatios Observatio 1: Cosider two exeutios of a sequee of istrutios. Eah exeutio takes lae omletely o a sigle roessig elemet oeted to a ahe of size Cad blok size B. The the umber of ahe misses iurred by the two exeutios a differ by at most C / B. ( As uder LRU ahe relaemet oliy oly the first aess to eah of the C/ Bbloks a ause a ahe miss i oe exeutio that is ot a miss i the other ) Observatio 2: Eah steal a ause at most two odes to drift: the stole ode ad ossibly the joi ode with its siblig.

85 Imliatios If there are Ssuessful steals durig arallel exeutio the there will be at most 2 sequetial exeutio. additioal ahe misses omared to the Now suose a divide-ad-oquer algorithm iurs ahe misses o a serial mahie. The o a arallel mahie with arallel roessig elemets eah oeted to a ahe of size ad blok size, the total umber of ahe misses iurred: Ο Ο...

Algorithms. Elementary Sorting. Dong Kyue Kim Hanyang University

Algorithms. Elementary Sorting. Dong Kyue Kim Hanyang University Algorithms Elemetary Sortig Dog Kyue Kim Hayag Uiversity dqkim@hayag.a.kr Cotets Sortig problem Elemetary sortig algorithms Isertio sort Merge sort Seletio sort Bubble sort Sortig problem Iput A sequee

More information

COMP26120: Introducing Complexity Analysis (2018/19) Lucas Cordeiro

COMP26120: Introducing Complexity Analysis (2018/19) Lucas Cordeiro COMP60: Itroduig Complexity Aalysis (08/9) Luas Cordeiro luas.ordeiro@mahester.a.uk Itroduig Complexity Aalysis Textbook: Algorithm Desig ad Appliatios, Goodrih, Mihael T. ad Roberto Tamassia (hapter )

More information

Recurrences: Methods and Examples

Recurrences: Methods and Examples Reurrees: Methods ad Examples CSE 30 Algorithms ad Data Strutures Alexadra Stefa Uiversity of exas at Arligto Updated: 308 Summatios Review Review slides o Summatios Reurrees Reursive algorithms It may

More information

2 Scalable algorithms, Scheduling, and a glance at All Prefix Sum

2 Scalable algorithms, Scheduling, and a glance at All Prefix Sum CME 323: Distributed Algorithms ad Otimizatio, Srig 2017 htt://staford.edu/~rezab/dao. Istructor: Reza Zadeh, Matroid ad Staford. Lecture 2, 4/5/2017. Scribed by Adreas Satucci. 2 Scalable algorithms,

More information

Recursive Algorithms. Recurrences. Recursive Algorithms Analysis

Recursive Algorithms. Recurrences. Recursive Algorithms Analysis Recursive Algorithms Recurreces Computer Sciece & Egieerig 35: Discrete Mathematics Christopher M Bourke cbourke@cseuledu A recursive algorithm is oe i which objects are defied i terms of other objects

More information

Chapter 2. Finite Fields (Chapter 3 in the text)

Chapter 2. Finite Fields (Chapter 3 in the text) Chater 2. Fiite Fields (Chater 3 i the tet 1. Grou Structures 2. Costructios of Fiite Fields GF(2 ad GF( 3. Basic Theory of Fiite Fields 4. The Miimal Polyomials 5. Trace Fuctios 6. Subfields 1. Grou Structures

More information

In number theory we will generally be working with integers, though occasionally fractions and irrationals will come into play.

In number theory we will generally be working with integers, though occasionally fractions and irrationals will come into play. Number Theory Math 5840 otes. Sectio 1: Axioms. I umber theory we will geerally be workig with itegers, though occasioally fractios ad irratioals will come ito play. Notatio: Z deotes the set of all itegers

More information

t distribution [34] : used to test a mean against an hypothesized value (H 0 : µ = µ 0 ) or the difference

t distribution [34] : used to test a mean against an hypothesized value (H 0 : µ = µ 0 ) or the difference EXST30 Backgroud material Page From the textbook The Statistical Sleuth Mea [0]: I your text the word mea deotes a populatio mea (µ) while the work average deotes a sample average ( ). Variace [0]: The

More information

IP Reference guide for integer programming formulations.

IP Reference guide for integer programming formulations. IP Referece guide for iteger programmig formulatios. by James B. Orli for 15.053 ad 15.058 This documet is iteded as a compact (or relatively compact) guide to the formulatio of iteger programs. For more

More information

The beta density, Bayes, Laplace, and Pólya

The beta density, Bayes, Laplace, and Pólya The beta desity, Bayes, Laplae, ad Pólya Saad Meimeh The beta desity as a ojugate form Suppose that is a biomial radom variable with idex ad parameter p, i.e. ( ) P ( p) p ( p) Applyig Bayes s rule, we

More information

THE KENNESAW STATE UNIVERSITY HIGH SCHOOL MATHEMATICS COMPETITION PART II Calculators are NOT permitted Time allowed: 2 hours

THE KENNESAW STATE UNIVERSITY HIGH SCHOOL MATHEMATICS COMPETITION PART II Calculators are NOT permitted Time allowed: 2 hours THE 06-07 KENNESAW STATE UNIVERSITY HIGH SCHOOL MATHEMATICS COMPETITION PART II Calculators are NOT permitted Time allowed: hours Let x, y, ad A all be positive itegers with x y a) Prove that there are

More information

Analysis of Algorithms. Introduction. Contents

Analysis of Algorithms. Introduction. Contents Itroductio The focus of this module is mathematical aspects of algorithms. Our mai focus is aalysis of algorithms, which meas evaluatig efficiecy of algorithms by aalytical ad mathematical methods. We

More information

Classification of DT signals

Classification of DT signals Comlex exoetial A discrete time sigal may be comlex valued I digital commuicatios comlex sigals arise aturally A comlex sigal may be rereseted i two forms: jarg { z( ) } { } z ( ) = Re { z ( )} + jim {

More information

Model of Computation and Runtime Analysis

Model of Computation and Runtime Analysis Model of Computatio ad Rutime Aalysis Model of Computatio Model of Computatio Specifies Set of operatios Cost of operatios (ot ecessarily time) Examples Turig Machie Radom Access Machie (RAM) PRAM Map

More information

Skip Lists. Presentation for use with the textbook, Algorithm Design and Applications, by M. T. Goodrich and R. Tamassia, Wiley, 2015 S 3 S S 1

Skip Lists. Presentation for use with the textbook, Algorithm Design and Applications, by M. T. Goodrich and R. Tamassia, Wiley, 2015 S 3 S S 1 Presetatio for use with the textbook, Algorithm Desig ad Applicatios, by M. T. Goodrich ad R. Tamassia, Wiley, 2015 Skip Lists S 3 15 15 23 10 15 23 36 Skip Lists 1 What is a Skip List A skip list for

More information

Definitions and Theorems. where x are the decision variables. c, b, and a are constant coefficients.

Definitions and Theorems. where x are the decision variables. c, b, and a are constant coefficients. Defiitios ad Theorems Remember the scalar form of the liear programmig problem, Miimize, Subject to, f(x) = c i x i a 1i x i = b 1 a mi x i = b m x i 0 i = 1,2,, where x are the decisio variables. c, b,

More information

CSE 4095/5095 Topics in Big Data Analytics Spring 2017; Homework 1 Solutions

CSE 4095/5095 Topics in Big Data Analytics Spring 2017; Homework 1 Solutions CSE 09/09 Topics i ig Data Aalytics Sprig 2017; Homework 1 Solutios Note: Solutios to problems,, ad 6 are due to Marius Nicolae. 1. Cosider the followig algorithm: for i := 1 to α log e do Pick a radom

More information

Bertrand s Postulate

Bertrand s Postulate Bertrad s Postulate Lola Thompso Ross Program July 3, 2009 Lola Thompso (Ross Program Bertrad s Postulate July 3, 2009 1 / 33 Bertrad s Postulate I ve said it oce ad I ll say it agai: There s always a

More information

WRITTEN ASSIGNMENT 1 ANSWER KEY

WRITTEN ASSIGNMENT 1 ANSWER KEY CISC 65 Itrodutio Desig ad Aalysis of Algorithms WRITTEN ASSIGNMENT ANSWER KEY. Problem -) I geeral, this problem requires f() = some time period be solve for a value. This a be doe for all ase expet lg

More information

Solutions to Problem Set 7

Solutions to Problem Set 7 8.78 Solutios to Problem Set 7. If the umber is i S, we re doe sice it s relatively rime to everythig. So suose S. Break u the remaiig elemets ito airs {, }, {4, 5},..., {, + }. By the Pigeohole Pricile,

More information

Recurrence Relations

Recurrence Relations Recurrece Relatios Aalysis of recursive algorithms, such as: it factorial (it ) { if (==0) retur ; else retur ( * factorial(-)); } Let t be the umber of multiplicatios eeded to calculate factorial(). The

More information

4.3 Growth Rates of Solutions to Recurrences

4.3 Growth Rates of Solutions to Recurrences 4.3. GROWTH RATES OF SOLUTIONS TO RECURRENCES 81 4.3 Growth Rates of Solutios to Recurreces 4.3.1 Divide ad Coquer Algorithms Oe of the most basic ad powerful algorithmic techiques is divide ad coquer.

More information

Model of Computation and Runtime Analysis

Model of Computation and Runtime Analysis Model of Computatio ad Rutime Aalysis Model of Computatio Model of Computatio Specifies Set of operatios Cost of operatios (ot ecessarily time) Examples Turig Machie Radom Access Machie (RAM) PRAM Map

More information

Cache-oblivious Dynamic Programming for Bioinformatics

Cache-oblivious Dynamic Programming for Bioinformatics 1 Cache-oblivious Dyamic Programmig for ioiformatics Rezaul Alam Chowdhury, Hai-So Le, ad Vijaya Ramachadra Abstract We preset efficiet cache-oblivious algorithms for some well-studied strig problems i

More information

Classification of problem & problem solving strategies. classification of time complexities (linear, logarithmic etc)

Classification of problem & problem solving strategies. classification of time complexities (linear, logarithmic etc) Classificatio of problem & problem solvig strategies classificatio of time complexities (liear, arithmic etc) Problem subdivisio Divide ad Coquer strategy. Asymptotic otatios, lower boud ad upper boud:

More information

CONSTRUCTING TRUNCATED IRRATIONAL NUMBERS AND DETERMINING THEIR NEIGHBORING PRIMES

CONSTRUCTING TRUNCATED IRRATIONAL NUMBERS AND DETERMINING THEIR NEIGHBORING PRIMES CONSTRUCTING TRUNCATED IRRATIONAL NUMBERS AND DETERMINING THEIR NEIGHBORING PRIMES It is well kow that there exist a ifiite set of irratioal umbers icludig, sqrt(), ad e. Such quatities are of ifiite legth

More information

OPTIMAL ALGORITHMS -- SUPPLEMENTAL NOTES

OPTIMAL ALGORITHMS -- SUPPLEMENTAL NOTES OPTIMAL ALGORITHMS -- SUPPLEMENTAL NOTES Peter M. Maurer Why Hashig is θ(). As i biary search, hashig assumes that keys are stored i a array which is idexed by a iteger. However, hashig attempts to bypass

More information

Parallel Programming: Speedups and Amdahl s law

Parallel Programming: Speedups and Amdahl s law Parallel Programmig: Seedus ad Amdahl s law Mike Bailey mjb@cs.oregostate.edu Orego State Uiversity Orego State Uiversity Comuter Grahics seedus.ad.amdahls.law.tx Defiitio of Seedu 2 If you are usig rocessors,

More information

NUMERICAL METHODS FOR SOLVING EQUATIONS

NUMERICAL METHODS FOR SOLVING EQUATIONS Mathematics Revisio Guides Numerical Methods for Solvig Equatios Page 1 of 11 M.K. HOME TUITION Mathematics Revisio Guides Level: GCSE Higher Tier NUMERICAL METHODS FOR SOLVING EQUATIONS Versio:. Date:

More information

The picture in figure 1.1 helps us to see that the area represents the distance traveled. Figure 1: Area represents distance travelled

The picture in figure 1.1 helps us to see that the area represents the distance traveled. Figure 1: Area represents distance travelled 1 Lecture : Area Area ad distace traveled Approximatig area by rectagles Summatio The area uder a parabola 1.1 Area ad distace Suppose we have the followig iformatio about the velocity of a particle, how

More information

ECEN 655: Advanced Channel Coding Spring Lecture 7 02/04/14. Belief propagation is exact on tree-structured factor graphs.

ECEN 655: Advanced Channel Coding Spring Lecture 7 02/04/14. Belief propagation is exact on tree-structured factor graphs. ECEN 655: Advaced Chael Codig Sprig 014 Prof. Hery Pfister Lecture 7 0/04/14 Scribe: Megke Lia 1 4-Cycles i Gallager s Esemble What we already kow: Belief propagatio is exact o tree-structured factor graphs.

More information

Algorithm Analysis. Algorithms that are equally correct can vary in their utilization of computational resources

Algorithm Analysis. Algorithms that are equally correct can vary in their utilization of computational resources Algorithm Aalysis Algorithms that are equally correct ca vary i their utilizatio of computatioal resources time ad memory a slow program it is likely ot to be used a program that demads too much memory

More information

Linear chord diagrams with long chords

Linear chord diagrams with long chords Liear chord diagrams with log chords Everett Sulliva Departmet of Mathematics Dartmouth College Haover New Hampshire, U.S.A. everett..sulliva@dartmouth.edu Submitted: Feb 7, 2017; Accepted: Oct 7, 2017;

More information

Algorithms Design & Analysis. Divide & Conquer

Algorithms Design & Analysis. Divide & Conquer Algorithms Desig & Aalysis Divide & Coquer Recap Direct-accessible table Hash tables Hash fuctios Uiversal hashig Perfect Hashig Ope addressig 2 Today s topics The divide-ad-coquer desig paradigm Revised

More information

Design and Analysis of Algorithms

Design and Analysis of Algorithms Desig ad Aalysis of Algorithms Probabilistic aalysis ad Radomized algorithms Referece: CLRS Chapter 5 Topics: Hirig problem Idicatio radom variables Radomized algorithms Huo Hogwei 1 The hirig problem

More information

k-equitable mean labeling

k-equitable mean labeling Joural of Algorithms ad Comutatio joural homeage: htt://jac.ut.ac.ir k-euitable mea labelig P.Jeyathi 1 1 Deartmet of Mathematics, Govidammal Aditaar College for Wome, Tiruchedur- 628 215,Idia ABSTRACT

More information

Problem Set 2 Solutions

Problem Set 2 Solutions CS271 Radomess & Computatio, Sprig 2018 Problem Set 2 Solutios Poit totals are i the margi; the maximum total umber of poits was 52. 1. Probabilistic method for domiatig sets 6pts Pick a radom subset S

More information

Cache-oblivious Dynamic Programming for Bioinformatics

Cache-oblivious Dynamic Programming for Bioinformatics IEEE/ACM TRANSACTINS N CMPUTATINAL ILGY AND IINFRMATICS, MANUSCRIPT ID?????????? 1 Cache-oblivious Dyamic Programmig for ioiformatics Rezaul Alam Chowdhury, Hai-So Le, ad Vijaya Ramachadra Abstract We

More information

Curve Sketching Handout #5 Topic Interpretation Rational Functions

Curve Sketching Handout #5 Topic Interpretation Rational Functions Curve Sketchig Hadout #5 Topic Iterpretatio Ratioal Fuctios A ratioal fuctio is a fuctio f that is a quotiet of two polyomials. I other words, p ( ) ( ) f is a ratioal fuctio if p ( ) ad q ( ) are polyomials

More information

YALE UNIVERSITY DEPARTMENT OF COMPUTER SCIENCE

YALE UNIVERSITY DEPARTMENT OF COMPUTER SCIENCE YALE UNIVERSITY DEPARTMENT OF COMPUTER SCIENCE CPSC 467a: Crytograhy ad Comuter Security Notes 16 (rev. 1 Professor M. J. Fischer November 3, 2008 68 Legedre Symbol Lecture Notes 16 ( Let be a odd rime,

More information

Zeros of Polynomials

Zeros of Polynomials Math 160 www.timetodare.com 4.5 4.6 Zeros of Polyomials I these sectios we will study polyomials algebraically. Most of our work will be cocered with fidig the solutios of polyomial equatios of ay degree

More information

B. Maddah ENMG 622 ENMG /20/09

B. Maddah ENMG 622 ENMG /20/09 B. Maddah ENMG 6 ENMG 5 5//9 Queueig Theory () Distributio of waitig time i M/M/ Let T q be the waitig time i queue of a ustomer. The it a be show that, ( ) t { q > } =. T t e Let T be the total time of

More information

Class #25 Wednesday, April 19, 2018

Class #25 Wednesday, April 19, 2018 Cla # Wedesday, April 9, 8 PDE: More Heat Equatio with Derivative Boudary Coditios Let s do aother heat equatio problem similar to the previous oe. For this oe, I ll use a square plate (N = ), but I m

More information

Resampling Methods. X (1/2), i.e., Pr (X i m) = 1/2. We order the data: X (1) X (2) X (n). Define the sample median: ( n.

Resampling Methods. X (1/2), i.e., Pr (X i m) = 1/2. We order the data: X (1) X (2) X (n). Define the sample median: ( n. Jauary 1, 2019 Resamplig Methods Motivatio We have so may estimators with the property θ θ d N 0, σ 2 We ca also write θ a N θ, σ 2 /, where a meas approximately distributed as Oce we have a cosistet estimator

More information

Optimization Methods MIT 2.098/6.255/ Final exam

Optimization Methods MIT 2.098/6.255/ Final exam Optimizatio Methods MIT 2.098/6.255/15.093 Fial exam Date Give: December 19th, 2006 P1. [30 pts] Classify the followig statemets as true or false. All aswers must be well-justified, either through a short

More information

THE INTEGRAL TEST AND ESTIMATES OF SUMS

THE INTEGRAL TEST AND ESTIMATES OF SUMS THE INTEGRAL TEST AND ESTIMATES OF SUMS. Itroductio Determiig the exact sum of a series is i geeral ot a easy task. I the case of the geometric series ad the telescoig series it was ossible to fid a simle

More information

CS 332: Algorithms. Linear-Time Sorting. Order statistics. Slide credit: David Luebke (Virginia)

CS 332: Algorithms. Linear-Time Sorting. Order statistics. Slide credit: David Luebke (Virginia) 1 CS 332: Algorithms Liear-Time Sortig. Order statistics. Slide credit: David Luebke (Virgiia) Quicksort: Partitio I Words Partitio(A, p, r): Select a elemet to act as the pivot (which?) Grow two regios,

More information

HOMEWORK 2 SOLUTIONS

HOMEWORK 2 SOLUTIONS HOMEWORK SOLUTIONS CSE 55 RANDOMIZED AND APPROXIMATION ALGORITHMS 1. Questio 1. a) The larger the value of k is, the smaller the expected umber of days util we get all the coupos we eed. I fact if = k

More information

w (1) ˆx w (1) x (1) /ρ and w (2) ˆx w (2) x (2) /ρ.

w (1) ˆx w (1) x (1) /ρ and w (2) ˆx w (2) x (2) /ρ. 2 5. Weighted umber of late jobs 5.1. Release dates ad due dates: maximimizig the weight of o-time jobs Oce we add release dates, miimizig the umber of late jobs becomes a sigificatly harder problem. For

More information

[ 47 ] then T ( m ) is true for all n a. 2. The greatest integer function : [ ] is defined by selling [ x]

[ 47 ] then T ( m ) is true for all n a. 2. The greatest integer function : [ ] is defined by selling [ x] [ 47 ] Number System 1. Itroductio Pricile : Let { T ( ) : N} be a set of statemets, oe for each atural umber. If (i), T ( a ) is true for some a N ad (ii) T ( k ) is true imlies T ( k 1) is true for all

More information

CHAPTER I: Vector Spaces

CHAPTER I: Vector Spaces CHAPTER I: Vector Spaces Sectio 1: Itroductio ad Examples This first chapter is largely a review of topics you probably saw i your liear algebra course. So why cover it? (1) Not everyoe remembers everythig

More information

ENGI 4421 Confidence Intervals (Two Samples) Page 12-01

ENGI 4421 Confidence Intervals (Two Samples) Page 12-01 ENGI 44 Cofidece Itervals (Two Samples) Page -0 Two Sample Cofidece Iterval for a Differece i Populatio Meas [Navidi sectios 5.4-5.7; Devore chapter 9] From the cetral limit theorem, we kow that, for sufficietly

More information

Mathematical Induction

Mathematical Induction Mathematical Iductio Itroductio Mathematical iductio, or just iductio, is a proof techique. Suppose that for every atural umber, P() is a statemet. We wish to show that all statemets P() are true. I a

More information

A sequence of numbers is a function whose domain is the positive integers. We can see that the sequence

A sequence of numbers is a function whose domain is the positive integers. We can see that the sequence Sequeces A sequece of umbers is a fuctio whose domai is the positive itegers. We ca see that the sequece,, 2, 2, 3, 3,... is a fuctio from the positive itegers whe we write the first sequece elemet as

More information

THE SOLUTION OF NONLINEAR EQUATIONS f( x ) = 0.

THE SOLUTION OF NONLINEAR EQUATIONS f( x ) = 0. THE SOLUTION OF NONLINEAR EQUATIONS f( ) = 0. Noliear Equatio Solvers Bracketig. Graphical. Aalytical Ope Methods Bisectio False Positio (Regula-Falsi) Fied poit iteratio Newto Raphso Secat The root of

More information

After the completion of this section the student. V.4.2. Power Series Solution. V.4.3. The Method of Frobenius. V.4.4. Taylor Series Solution

After the completion of this section the student. V.4.2. Power Series Solution. V.4.3. The Method of Frobenius. V.4.4. Taylor Series Solution Chapter V ODE V.4 Power Series Solutio Otober, 8 385 V.4 Power Series Solutio Objetives: After the ompletio of this setio the studet - should reall the power series solutio of a liear ODE with variable

More information

Lovász Local Lemma a new tool to asymptotic enumeration?

Lovász Local Lemma a new tool to asymptotic enumeration? Lovász Local Lemma a ew tool to asymptotic eumeratio? Liyua Licol Lu László Székely Uiversity of South Carolia Supported by NSF DMS Overview LLL ad its geeralizatios LLL a istace of the Poisso paradigm

More information

Lecture 5: April 17, 2013

Lecture 5: April 17, 2013 TTIC/CMSC 350 Mathematical Toolkit Sprig 203 Madhur Tulsiai Lecture 5: April 7, 203 Scribe: Somaye Hashemifar Cheroff bouds recap We recall the Cheroff/Hoeffdig bouds we derived i the last lecture idepedet

More information

Introduction to Algorithms 6.046J/18.401J LECTURE 3 Divide and conquer Binary search Powering a number Fibonacci numbers Matrix multiplication

Introduction to Algorithms 6.046J/18.401J LECTURE 3 Divide and conquer Binary search Powering a number Fibonacci numbers Matrix multiplication Itroductio to Algorithms 6.046J/8.40J LECTURE 3 Divide ad coquer Biary search Powerig a umber Fiboacci umbers Matrix multiplicatio Strasse s algorithm VLSI tree layout Prof. Charles E. Leiserso The divide-ad-coquer

More information

Parallel Vector Algorithms David A. Padua

Parallel Vector Algorithms David A. Padua Parallel Vector Algorithms 1 of 32 Itroductio Next, we study several algorithms where parallelism ca be easily expressed i terms of array operatios. We will use Fortra 90 to represet these algorithms.

More information

Design and Analysis of Algorithms

Design and Analysis of Algorithms Desig ad Aalysis of Algorithms CSE 53 Lecture 9 Media ad Order Statistics Juzhou Huag, Ph.D. Departmet of Computer Sciece ad Egieerig Dept. CSE, UT Arligto CSE53 Desig ad Aalysis of Algorithms Medias ad

More information

3.2 Properties of Division 3.3 Zeros of Polynomials 3.4 Complex and Rational Zeros of Polynomials

3.2 Properties of Division 3.3 Zeros of Polynomials 3.4 Complex and Rational Zeros of Polynomials Math 60 www.timetodare.com 3. Properties of Divisio 3.3 Zeros of Polyomials 3.4 Complex ad Ratioal Zeros of Polyomials I these sectios we will study polyomials algebraically. Most of our work will be cocered

More information

Recursive Algorithm for Generating Partitions of an Integer. 1 Preliminary

Recursive Algorithm for Generating Partitions of an Integer. 1 Preliminary Recursive Algorithm for Geeratig Partitios of a Iteger Sug-Hyuk Cha Computer Sciece Departmet, Pace Uiversity 1 Pace Plaza, New York, NY 10038 USA scha@pace.edu Abstract. This article first reviews the

More information

tests 17.1 Simple versus compound

tests 17.1 Simple versus compound PAS204: Lecture 17. tests UMP ad asymtotic I this lecture, we will idetify UMP tests, wherever they exist, for comarig a simle ull hyothesis with a comoud alterative. We also look at costructig tests based

More information

REVERSIBLE NON-FLOW PROCESS CONSTANT VOLUME PROCESS (ISOCHORIC PROCESS) In a constant volume process, he working substance is contained in a rigid

REVERSIBLE NON-FLOW PROCESS CONSTANT VOLUME PROCESS (ISOCHORIC PROCESS) In a constant volume process, he working substance is contained in a rigid REVERSIBLE NON-FLOW PROCESS CONSTANT VOLUME PROCESS (ISOCHORIC PROCESS) I a ostat olume roess, he workig substae is otaied i a rigid essel, hee the boudaries of the system are immoable, so work aot be

More information

CSE 1400 Applied Discrete Mathematics Number Theory and Proofs

CSE 1400 Applied Discrete Mathematics Number Theory and Proofs CSE 1400 Applied Discrete Mathematics Number Theory ad Proofs Departmet of Computer Scieces College of Egieerig Florida Tech Sprig 01 Problems for Number Theory Backgroud Number theory is the brach of

More information

1. By using truth tables prove that, for all statements P and Q, the statement

1. By using truth tables prove that, for all statements P and Q, the statement Author: Satiago Salazar Problems I: Mathematical Statemets ad Proofs. By usig truth tables prove that, for all statemets P ad Q, the statemet P Q ad its cotrapositive ot Q (ot P) are equivalet. I example.2.3

More information

Cache-Oblivious Computations: Algorithms and Experimental Evaluation

Cache-Oblivious Computations: Algorithms and Experimental Evaluation Cache-Oblivious Computations: Algorithms and Experimental Evaluation Vijaya Ramachandran Department of Computer Sciences University of Texas at Austin Dissertation work of former PhD student Dr. Rezaul

More information

Introduction to Algorithms 6.046J/18.401J

Introduction to Algorithms 6.046J/18.401J Itrodutio to Algorithms.04J/8.40J The divide-d-oquer desig prdigm. Divide the problem (iste) ito subproblems.. Coquer the subproblems by solvig them reursively. 3. Combie subproblem solutios. Leture 3

More information

Construction of Some New Three Associate Class Pbib Designs with Two Replicates

Construction of Some New Three Associate Class Pbib Designs with Two Replicates Iteratioal Joural of Academic Research i Progressive Educatio ad Develomet Costructio of Some New Three Associate Class Pbib Desigs with Two Relicates Silver J.Key Rambaei School of Sciece, Deartmet of

More information

Integer Linear Programming

Integer Linear Programming Iteger Liear Programmig Itroductio Iteger L P problem (P) Mi = s. t. a = b i =,, m = i i 0, iteger =,, c Eemple Mi z = 5 s. t. + 0 0, 0, iteger F(P) = feasible domai of P Itroductio Iteger L P problem

More information

Fortgeschrittene Datenstrukturen Vorlesung 11

Fortgeschrittene Datenstrukturen Vorlesung 11 Fortgeschrittee Datestruture Vorlesug 11 Schriftführer: Marti Weider 19.01.2012 1 Succict Data Structures (ctd.) 1.1 Select-Queries A slightly differet approach, compared to ra, is used for select. B represets

More information

An analog of the arithmetic triangle obtained by replacing the products by the least common multiples

An analog of the arithmetic triangle obtained by replacing the products by the least common multiples arxiv:10021383v2 [mathnt] 9 Feb 2010 A aalog of the arithmetic triagle obtaied by replacig the products by the least commo multiples Bair FARHI bairfarhi@gmailcom MSC: 11A05 Keywords: Al-Karaji s triagle;

More information

Math 299 Supplement: Real Analysis Nov 2013

Math 299 Supplement: Real Analysis Nov 2013 Math 299 Supplemet: Real Aalysis Nov 203 Algebra Axioms. I Real Aalysis, we work withi the axiomatic system of real umbers: the set R alog with the additio ad multiplicatio operatios +,, ad the iequality

More information

CO Algorithms: Brief History

CO Algorithms: Brief History Piyush Kumar Departmet of Computer Sciece Advisor: Joseph S.B. Mitchell CO Algorithms: Brief History Frigo, Leiserso, Prokop, Ramachadra (FOCS 99) Cache Oblivious Algorithms Thesis Harold Prokop s Thesis

More information

Optimally Sparse SVMs

Optimally Sparse SVMs A. Proof of Lemma 3. We here prove a lower boud o the umber of support vectors to achieve geeralizatio bouds of the form which we cosider. Importatly, this result holds ot oly for liear classifiers, but

More information

Lecture Overview. 2 Permutations and Combinations. n(n 1) (n (k 1)) = n(n 1) (n k + 1) =

Lecture Overview. 2 Permutations and Combinations. n(n 1) (n (k 1)) = n(n 1) (n k + 1) = COMPSCI 230: Discrete Mathematics for Computer Sciece April 8, 2019 Lecturer: Debmalya Paigrahi Lecture 22 Scribe: Kevi Su 1 Overview I this lecture, we begi studyig the fudametals of coutig discrete objects.

More information

Chimica Inorganica 3

Chimica Inorganica 3 himica Iorgaica Irreducible Represetatios ad haracter Tables Rather tha usig geometrical operatios, it is ofte much more coveiet to employ a ew set of group elemets which are matrices ad to make the rule

More information

The Boolean Ring of Intervals

The Boolean Ring of Intervals MATH 532 Lebesgue Measure Dr. Neal, WKU We ow shall apply the results obtaied about outer measure to the legth measure o the real lie. Throughout, our space X will be the set of real umbers R. Whe ecessary,

More information

Advanced Course of Algorithm Design and Analysis

Advanced Course of Algorithm Design and Analysis Differet complexity measures Advaced Course of Algorithm Desig ad Aalysis Asymptotic complexity Big-Oh otatio Properties of O otatio Aalysis of simple algorithms A algorithm may may have differet executio

More information

A Probabilistic Analysis of Quicksort

A Probabilistic Analysis of Quicksort A Probabilistic Aalysis of Quicsort You are assumed to be familiar with Quicsort. I each iteratio this sortig algorithm chooses a pivot ad the, by performig comparisios with the pivot, splits the remaider

More information

Langford s Problem. Moti Ben-Ari. Department of Science Teaching. Weizmann Institute of Science.

Langford s Problem. Moti Ben-Ari. Department of Science Teaching. Weizmann Institute of Science. Lagford s Problem Moti Be-Ari Departmet of Sciece Teachig Weizma Istitute of Sciece http://www.weizma.ac.il/sci-tea/beari/ c 017 by Moti Be-Ari. This work is licesed uder the Creative Commos Attributio-ShareAlike

More information

Problem 4: Evaluate ( k ) by negating (actually un-negating) its upper index. Binomial coefficient

Problem 4: Evaluate ( k ) by negating (actually un-negating) its upper index. Binomial coefficient Problem 4: Evaluate by egatig actually u-egatig its upper idex We ow that Biomial coefficiet r { where r is a real umber, is a iteger The above defiitio ca be recast i terms of factorials i the commo case

More information

Basics of Inference. Lecture 21: Bayesian Inference. Review - Example - Defective Parts, cont. Review - Example - Defective Parts

Basics of Inference. Lecture 21: Bayesian Inference. Review - Example - Defective Parts, cont. Review - Example - Defective Parts Basics of Iferece Lecture 21: Sta230 / Mth230 Coli Rudel Aril 16, 2014 U util this oit i the class you have almost exclusively bee reseted with roblems where we are usig a robability model where the model

More information

SINGLE-CHANNEL QUEUING PROBLEMS APPROACH

SINGLE-CHANNEL QUEUING PROBLEMS APPROACH SINGLE-CHANNEL QUEUING ROBLEMS AROACH Abdurrzzag TAMTAM, Doctoral Degree rogramme () Dept. of Telecommuicatios, FEEC, BUT E-mail: xtamta@stud.feec.vutbr.cz Supervised by: Dr. Karol Molár ABSTRACT The paper

More information

Chapter 9: Numerical Differentiation

Chapter 9: Numerical Differentiation 178 Chapter 9: Numerical Differetiatio Numerical Differetiatio Formulatio of equatios for physical problems ofte ivolve derivatives (rate-of-chage quatities, such as velocity ad acceleratio). Numerical

More information

Sequences, Mathematical Induction, and Recursion. CSE 2353 Discrete Computational Structures Spring 2018

Sequences, Mathematical Induction, and Recursion. CSE 2353 Discrete Computational Structures Spring 2018 CSE 353 Discrete Computatioal Structures Sprig 08 Sequeces, Mathematical Iductio, ad Recursio (Chapter 5, Epp) Note: some course slides adopted from publisher-provided material Overview May mathematical

More information

Linear Programming and the Simplex Method

Linear Programming and the Simplex Method Liear Programmig ad the Simplex ethod Abstract This article is a itroductio to Liear Programmig ad usig Simplex method for solvig LP problems i primal form. What is Liear Programmig? Liear Programmig is

More information

CHAPTER 5 SOME MINIMAX AND SADDLE POINT THEOREMS

CHAPTER 5 SOME MINIMAX AND SADDLE POINT THEOREMS CHAPTR 5 SOM MINIMA AND SADDL POINT THORMS 5. INTRODUCTION Fied poit theorems provide importat tools i game theory which are used to prove the equilibrium ad eistece theorems. For istace, the fied poit

More information

Nonequilibrium Excess Carriers in Semiconductors

Nonequilibrium Excess Carriers in Semiconductors Lecture 8 Semicoductor Physics VI Noequilibrium Excess Carriers i Semicoductors Noequilibrium coditios. Excess electros i the coductio bad ad excess holes i the valece bad Ambiolar trasort : Excess electros

More information

Chapter 2. Periodic points of toral. automorphisms. 2.1 General introduction

Chapter 2. Periodic points of toral. automorphisms. 2.1 General introduction Chapter 2 Periodic poits of toral automorphisms 2.1 Geeral itroductio The automorphisms of the two-dimesioal torus are rich mathematical objects possessig iterestig geometric, algebraic, topological ad

More information

Apply change-of-basis formula to rewrite x as a linear combination of eigenvectors v j.

Apply change-of-basis formula to rewrite x as a linear combination of eigenvectors v j. Eigevalue-Eigevector Istructor: Nam Su Wag eigemcd Ay vector i real Euclidea space of dimesio ca be uiquely epressed as a liear combiatio of liearly idepedet vectors (ie, basis) g j, j,,, α g α g α g α

More information

PAijpam.eu ON TENSOR PRODUCT DECOMPOSITION

PAijpam.eu ON TENSOR PRODUCT DECOMPOSITION Iteratioal Joural of Pure ad Applied Mathematics Volume 103 No 3 2015, 537-545 ISSN: 1311-8080 (prited versio); ISSN: 1314-3395 (o-lie versio) url: http://wwwijpameu doi: http://dxdoiorg/1012732/ijpamv103i314

More information

TEACHER CERTIFICATION STUDY GUIDE

TEACHER CERTIFICATION STUDY GUIDE COMPETENCY 1. ALGEBRA SKILL 1.1 1.1a. ALGEBRAIC STRUCTURES Kow why the real ad complex umbers are each a field, ad that particular rigs are ot fields (e.g., itegers, polyomial rigs, matrix rigs) Algebra

More information

CSE Introduction to Parallel Processing. Chapter 3. Parallel Algorithm Complexity

CSE Introduction to Parallel Processing. Chapter 3. Parallel Algorithm Complexity Dr. Izadi CSE-40533 Itroductio to Parallel Processig Chapter 3 Parallel Algorithm Complexity Review algorithm complexity ad various complexity classes Itroduce the otios of time ad time-cost optimality

More information

Square-Congruence Modulo n

Square-Congruence Modulo n Square-Cogruece Modulo Abstract This paper is a ivestigatio of a equivalece relatio o the itegers that was itroduced as a exercise i our Discrete Math class. Part I - Itro Defiitio Two itegers are Square-Cogruet

More information

1 of 7 7/16/2009 6:06 AM Virtual Laboratories > 6. Radom Samples > 1 2 3 4 5 6 7 6. Order Statistics Defiitios Suppose agai that we have a basic radom experimet, ad that X is a real-valued radom variable

More information

A widely used display of protein shapes is based on the coordinates of the alpha carbons - - C α

A widely used display of protein shapes is based on the coordinates of the alpha carbons - - C α Nice plottig of proteis: I A widely used display of protei shapes is based o the coordiates of the alpha carbos - - C α -s. The coordiates of the C α -s are coected by a cotiuous curve that roughly follows

More information

Round-off Errors and Computer Arithmetic - (1.2)

Round-off Errors and Computer Arithmetic - (1.2) Roud-off Errors ad Comuter Arithmetic - (1.) 1. Roud-off Errors: Roud-off errors is roduced whe a calculator or comuter is used to erform real umber calculatios. That is because the arithmetic erformed

More information

CALCULUS BASIC SUMMER REVIEW

CALCULUS BASIC SUMMER REVIEW CALCULUS BASIC SUMMER REVIEW NAME rise y y y Slope of a o vertical lie: m ru Poit Slope Equatio: y y m( ) The slope is m ad a poit o your lie is, ). ( y Slope-Itercept Equatio: y m b slope= m y-itercept=

More information