function merging
play

FUNCTION MERGING by SEQUENCE ALIGNMENT DOES SIZE MATTER? DOES - PowerPoint PPT Presentation

Nov 21, 2022 •111 likes •1.89k views

Rodrigo Rocha http://rcor.me University of Edinburgh FUNCTION MERGING by SEQUENCE ALIGNMENT DOES SIZE MATTER? DOES SIZE MATTER? S E Y DOES SIZE MATTER? S E Y ... EMBEDDED SYSTEMS SAVINGS IN CODE Smaller Memories 1 0 1 1 0 1

  1. WHAT WE HAVE Production compilers: LLVM, GCC, MSVC

  2. WHAT WE HAVE Production compilers: LLVM, GCC, MSVC Only Identical Functions

  3. WHAT WE HAVE The state of the art

  4. WHAT WE HAVE The state of the art significant improvement but...

  5. WHAT WE HAVE The state of the art significant improvement but... MUST HAVE IDENTICAL

  6. WHAT WE HAVE The state of the art significant improvement but... MUST HAVE IDENTICAL CFGs

  7. WHAT WE HAVE The state of the art significant improvement but... MUST HAVE IDENTICAL CFGs Return types

  8. WHAT WE HAVE The state of the art significant improvement but... MUST HAVE IDENTICAL CFGs Return types Lists of parameters

  9. WHAT WE HAVE The state of the art significant improvement but... MUST HAVE IDENTICAL CFGs Return types Lists of parameters Number of instructions

  10. WHAT WE WANT

  11. WHAT WE WANT Merge ANY two functions

  12. WHAT WE WANT Merge ANY two functions Merge ANY two functions and CHOOSE when to do it

  13. HOW WE DO IT int f1(int a, int b) { void f2(int a, int b) { int var = a*b + 10; int var = a*b + b; int result = foo(var); int result = foo(var); printf("result: %d\n", result); printf("result: %d\n", result); return result; if (result==0) } printf("result is zero\n"); }

  14. HOW WE DO IT label : entry label : entry IR Level %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) ret %r2 %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  15. HOW WE DO IT label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) ret %r2 %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  16. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) ret %r2 %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  17. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %x1 = add %x0, %b %r1 = add %r0, 10 %x2 = call foo(%x1) %r2 = call foo(%r1) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 %r3 = call printf(@str,%r2) br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  18. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %x1 = add %x0, %b %r1 = add %r0, 10 %x2 = call foo(%x1) %r2 = call foo(%r1) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 %r3 = call printf(@str,%r2) br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  19. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  20. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  21. LINEARIZATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  22. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  23. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  24. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b Match %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  25. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  26. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  27. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b Match %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  28. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b Match %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  29. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b Match %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  30. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  31. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  32. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) Match %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  33. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) Match %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  34. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) Match %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  35. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  36. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo(%r1) %x2 = call foo(%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  37. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  38. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) Match %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  39. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  40. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf(@str,%r2) %x3 = call printf(@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  41. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  42. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  43. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end Mismatch label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  44. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 ret %r2 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  45. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  46. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  47. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  48. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end ret %r2 label : if.then %x5 = call printf(@str) br %if.end label : if.end ret void

  49. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) ret %r2 br %if.end label : if.end ret void

  50. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) ret %r2 br %if.end label : if.end ret void

  51. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) ret %r2 br %if.end label : if.end ret void

  52. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) ret %r2 br %if.end label : if.end ret void

  53. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end ret %r2 label : if.end ret void

  54. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end ret %r2 label : if.end ret void

  55. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

  56. IDENTIFY WHAT TO MERGE label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

  57. SEQUENCE ALIGNMENT Bioinformatics G C A T - G C U G - A T T A C A

  58. call printf SEQUENCE ALIGNMENT call foo label ret add mul label mul add call foo call printf icmp eq br label call br label ret

  59. call printf Match: +2 SEQUENCE ALIGNMENT Mismatch: -1 call foo Gap: -1 label ret add mul label mul add call foo call printf icmp eq br label call br label ret

  60. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 mul -2 add -3 call foo -4 call printf -5 ret -6

  61. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 mul -2 add -3 call foo -4 call printf -5 ret -6

  62. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 mul -2 add -3 call foo -4 call printf -5 ret -6 -1 -1 = -2

  63. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 mul -2 add -3 call foo -4 call printf -5 ret -6 -1 -1 = -2 -1 -1 = -2

  64. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 mul -2 add -3 call foo -4 call printf -5 ret -6 -1 -1 = -2 -1 -1 = -2 0 + 2 = 2

  65. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 mul -2 add -3 call foo -4 call printf -5 ret -6 -1 -1 = -2 -1 -1 = -2 0 + 2 = 2

  66. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 mul -2 add -3 call foo -4 call printf -5 ret -6 2 -1 = 1 -2 -1 = -3 -1 -1 = -2

  67. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 1 mul -2 add -3 call foo -4 call printf -5 ret -6 2 -1 = 1 -2 -1 = -3 -1 -1 = -2

  68. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 mul -5 -2 1 4 3 2 1 0 -1 -2 -3 -4 -6 add -3 0 3 6 5 4 3 1 0 -1 -3 2 -2 call foo -4 -1 2 5 8 7 6 5 4 3 2 1 0 call printf -5 -2 1 4 7 10 9 8 7 6 5 4 3 ret -6 -3 0 3 6 9 9 8 7 6 5 4 6

  69. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 mul -5 -2 1 4 3 2 1 0 -1 -2 -3 -4 -6 add -3 0 3 6 5 4 3 1 0 -1 -3 2 -2 call foo -4 -1 2 5 8 7 6 5 4 3 2 1 0 call printf -5 -2 1 4 7 10 9 8 7 6 5 4 3 ret -6 -3 0 3 6 9 9 8 7 6 5 4 6

  70. SEQUENCE ALIGNMENT Match: +2 call printf Mismatch: -1 call foo icmp eq Gap: -1 label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 mul -5 -2 1 4 3 2 1 0 -1 -2 -3 -4 -6 add -3 0 3 6 5 4 3 1 0 -1 -3 2 -2 call foo -4 -1 2 5 8 7 6 5 4 3 2 1 0 call printf -5 -2 1 4 7 10 9 8 7 6 5 4 3 ret -6 -3 0 3 6 9 9 8 7 6 5 4 6

  71. SEQUENCE ALIGNMENT call printf call foo icmp eq label label label call mul add ret br br 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 label -1 2 1 0 -1 -2 -3 -4 -5 -6 -7 -8 -9 mul -5 -2 1 4 3 2 1 0 -1 -2 -3 -4 -6 add -3 0 3 6 5 4 3 1 0 -1 -3 2 -2 call foo -4 -1 2 5 8 7 6 5 4 3 2 1 0 call printf -5 -2 1 4 7 10 9 8 7 6 5 4 3 ret -6 -3 0 3 6 9 9 8 7 6 5 4 6

  72. CODE GENERATION label : entry label : entry %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

  73. CODE GENERATION Renaming Table label: b1 entry b1 %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

  74. CODE GENERATION Renaming Table label: b1 entry b1 %r0 = mul %a, %b %x0 = mul %a, %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

  75. CODE GENERATION Renaming Table label: b1 entry b1 %m0 = mul %r0 = mul %a , %b %x0 = mul %a , %b %r1 = add %r0, 10 %x1 = add %x0, %b %r2 = call foo (%r1) %x2 = call foo (%x1) %r3 = call printf (@str,%r2) %x3 = call printf (@str,%x2) %x4 = icmp eq %x2, 0 br %x4, %if.then, %if.end label : if.then %x5 = call printf(@str) br %if.end label : if.end ret %r2 ret void

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

Parton Showers and Matching/Merging Lecture 2 of 2: Matching/Merging & Non-Perturbative

Parton Showers and Matching/Merging Lecture 2 of 2: Matching/Merging & Non-Perturbative

Parton Showers and Matching/Merging Lecture 2 of 2: Matching/Merging & Non-Perturbative Corrections (Hadron Decays) Hadronisation Parton Shower Matching & Merging Hard process Peter Skands (Monash University) Feynrules/Madgraph

698 views • 27 slides
Optimal Merging in Quantum k -xor and k -sum Algorithms Mara Naya-Plasencia, Andr

Optimal Merging in Quantum k -xor and k -sum Algorithms Mara Naya-Plasencia, Andr

Optimal Merging in Quantum k -xor and k -sum Algorithms Mara Naya-Plasencia, Andr Schrottenloher Inria, France Merging with many Solutions Quantum Merging With a Single Solution Outline Merging with many Solutions 1 Quantum Merging 2

727 views • 34 slides
Track Filtering/Quality/Merging A proposal for data format of track quality and track merging in

Track Filtering/Quality/Merging A proposal for data format of track quality and track merging in

Track Filtering/Quality/Merging A proposal for data format of track quality and track merging in CMSSW K.Burkett, L.Lista, B.Mangano, S.Rahatlou, A.Rizzi, J.-R.Vlimant 2/4/08 Track Quality/Filtering/Merging Jean-Roch Vlimant 1 Requirements

432 views • 7 slides
Merging DataFrames Merging DataFrames with pandas Population DataFrame In [1]: import pandas as

Merging DataFrames Merging DataFrames with pandas Population DataFrame In [1]: import pandas as

MERGING DATAFRAMES WITH PANDAS Merging DataFrames Merging DataFrames with pandas Population DataFrame In [1]: import pandas as pd In [2]: population = pd.read_csv('pa_zipcode_population.csv') In [3]: print(population) Zipcode 2010 Census

412 views • 40 slides
Code size reduction using Similar Function Merging Tobias Edler von Koch (University of

Code size reduction using Similar Function Merging Tobias Edler von Koch (University of

Open Source | Open Possibilities Code size reduction using Similar Function Merging Tobias Edler von Koch (University of Edinburgh, Qualcomm Innovation Center) Pranav Bhandarkar (Qualcomm Innovation Center) Open Source | Open Possibilities

500 views • 21 slides
Comparison Based Merging Upper and Lower bounds EMADS Fall 2003: Comparison Based Merging Page 1

Comparison Based Merging Upper and Lower bounds EMADS Fall 2003: Comparison Based Merging Page 1

Comparison Based Merging Upper and Lower bounds EMADS Fall 2003: Comparison Based Merging Page 1 Merging Input: Two sorted lists X and Y of length n and m . We may assume n m . n X : Y : m Theorem: In a comparison based model, the

435 views • 10 slides
Buying, Selling, Merging Buying, Selling, Merging and Valuation and Valuation Sponsored by: US

Buying, Selling, Merging Buying, Selling, Merging and Valuation and Valuation Sponsored by: US

Buying, Selling, Merging Buying, Selling, Merging and Valuation and Valuation Sponsored by: US Oncology Sponsored by: US Oncology Daryl Johnson Chris E. Rossman Principal, HealthCare Appraisers, Inc. Partner, Foley & Lardner LLP

718 views • 20 slides
Lecture 7 Rebasing Sign in on the attendance sheet! Today Review of merging Rebasing

Lecture 7 Rebasing Sign in on the attendance sheet! Today Review of merging Rebasing

Lecture 7 Rebasing Sign in on the attendance sheet! Today Review of merging Rebasing instead of merging Interactive rebases Review: Merging To combine two branches, with a merge, we do something like branchA, branchB HEAD

242 views • 22 slides
Quantum Merging Algorithms Mara Naya-Plasencia 2 , Andr Schrottenloher 2 Joint work with Andr

Quantum Merging Algorithms Mara Naya-Plasencia 2 , Andr Schrottenloher 2 Joint work with Andr

Quantum (Generalized) Collisions Quantum Merging Extended Quantum Merging Quantum Merging Algorithms Mara Naya-Plasencia 2 , Andr Schrottenloher 2 Joint work with Andr Chailloux 2 and Lorenzo Grassi 1 1 IAIK, Graz University of Technology,

727 views • 50 slides
Automatic Merging of Automatic Merging of Pedigree Information Pedigree Information Annual

Automatic Merging of Automatic Merging of Pedigree Information Pedigree Information Annual

Automatic Merging of Automatic Merging of Pedigree Information Pedigree Information Annual Workshop on Family History Technology April 3, 2003 Sue Dintelman and Tim Maness Pleiades Software Development, Inc. Pleiades Software Development,

340 views • 17 slides
NET ETCard Card Ci Citibank tibank Merging usernames under single log-in Presented by:

NET ETCard Card Ci Citibank tibank Merging usernames under single log-in Presented by:

Procu Procurem rement ent, , Travel, el, and nd NET ETCard Card Ci Citibank tibank Merging usernames under single log-in Presented by: Department of Finance Merging Cards, such as PCard and TCard, under a single username sign-on

342 views • 9 slides
Yet Another Approach To Model Merging merge and diff relations and very short version rules

Yet Another Approach To Model Merging merge and diff relations and very short version rules

model merging M. Famelis about model algebra Yet Another Approach To Model Merging merge and diff relations and very short version rules merging with rules Michail Famelis famelis@cs.toronto.edu ba3270:d17 September 16, 2008 model

411 views • 6 slides
Pin Merging in Planar Body Frameworks Rudi Penne rudi.penne@kdg.be Karel de Grote-Hogeschool

Pin Merging in Planar Body Frameworks Rudi Penne rudi.penne@kdg.be Karel de Grote-Hogeschool

Rigidity Workshop, Toronto 2011 Pin Merging in Planar Body Frameworks Rudi Penne rudi.penne@kdg.be Karel de Grote-Hogeschool University of Antwerp pin merging in planar body frameworks p. 1/25 Meet a 3 2 T -graph pin merging in planar

1.54k views • 121 slides
Fluent Merging for Classical Planning Problems Jendrik Seipp Malte Helmert

Fluent Merging for Classical Planning Problems Jendrik Seipp Malte Helmert

Background Fluent Merging Fluent Selection Same object method Mutex threshold Fluent Merging for Classical Planning Problems Jendrik Seipp Malte Helmert Albert-Ludwigs-Universit at Freiburg, Germany ICAPS 2011 Workshop on Knowledge

739 views • 38 slides
Comment on bitonic merging; more CUDA performance tuning CSE 6230: HPC Tools & Apps Tu Sep

Comment on bitonic merging; more CUDA performance tuning CSE 6230: HPC Tools & Apps Tu Sep

Comment on bitonic merging; more CUDA performance tuning CSE 6230: HPC Tools & Apps Tu Sep 18, 2012 Tuesday, September 18, 12 Comment on bitonic merging , including ideas & hints for Lab 3 Note: Some figures taken from Grama et al.

1.88k views • 88 slides
Subaru Weak Lensing of Seven Merging Clusters:

Subaru Weak Lensing of Seven Merging Clusters:

Removed ALL Figures Subaru Weak Lensing of Seven Merging Clusters: Distributions of Mass and Baryons Nobuhiro Okabe

231 views • 8 slides
Week 7 - Friday What did we talk about last time? Array examples We can represent a

Week 7 - Friday What did we talk about last time? Array examples We can represent a

Week 7 - Friday What did we talk about last time? Array examples We can represent a deck of cards as an array of 52 items One easy way is to make each item a String giving the name of the card We can extend the lab with cards

880 views • 29 slides
PThreads Thanks to LLNL for their tutorial from which these slides are derived

PThreads Thanks to LLNL for their tutorial from which these slides are derived

PThreads Thanks to LLNL for their tutorial from which these slides are derived http://www.llnl.gov/computing/tutorials/worksh ops/workshop/pthreads/MAIN.html Processes and threads Understanding what a thread means knowing the relationship

437 views • 33 slides
Efficient Defenses Against Adversarial Examples for Deep Neural Networks Valentina Zantedeschi

Efficient Defenses Against Adversarial Examples for Deep Neural Networks Valentina Zantedeschi

Efficient Defenses Against Adversarial Examples for Deep Neural Networks Valentina Zantedeschi Irina Nicolae Ambrish Rawat @vzantedesc @ririnicolae @ambrishrawat Jean Monnet University IBM Research AI GreHack #5 November 17, 2017

483 views • 37 slides
Bandwidth costs Henning Schulzrinne Dept. of Computer Science Columbia University New York, NY

Bandwidth costs Henning Schulzrinne Dept. of Computer Science Columbia University New York, NY

Bandwidth costs Henning Schulzrinne Dept. of Computer Science Columbia University New York, NY IETF 72 (Dublin) July 29, 2008 Overview Video bandwidth consumption Cost of providing video content Economics Mechanisms

429 views • 11 slides
Objects/Inheritance Wrapup Class : descrip5on of a data type

Objects/Inheritance Wrapup Class : descrip5on of a data type

Objects/Inheritance Wrapup Class : descrip5on of a data type that can contain fields (variables) and methods (func5ons) Think of a class as a

669 views • 12 slides
Efficient Protection of Path-Sensitive Control Security Ren Ding , Chenxiong Qian, Chengyu Song*,

Efficient Protection of Path-Sensitive Control Security Ren Ding , Chenxiong Qian, Chengyu Song*,

Efficient Protection of Path-Sensitive Control Security Ren Ding , Chenxiong Qian, Chengyu Song*, Bill Harris, Taesoo Kim, Wenke Lee Georgia Tech, UC Riverside* What is Control Flow? The order of instruction execution Only limited sets of

480 views • 29 slides
CS 162 Intro to Programming II Polymorphism II 1

CS 162 Intro to Programming II Polymorphism II 1

CS 162 Intro to Programming II Polymorphism II 1 Virtual Func:ons Which is which, for changing the behavior of inherited func:ons?

504 views • 18 slides
NaturalNumber 7 January 2019 OSU CSE 1 NaturalNumber The NaturalNumber component family

NaturalNumber 7 January 2019 OSU CSE 1 NaturalNumber The NaturalNumber component family

NaturalNumber 7 January 2019 OSU CSE 1 NaturalNumber The NaturalNumber component family allows you to manipulate natural numbers (i.e., non-negative integers) Unlike an int variable, a NaturalNumber variable has no upper bound on its

1.13k views • 100 slides

More recommend