| Yossi -
There is a difference between not parallelizing and not speeding up. With
the switches I used (given below) KAP was able to parallelize almost all of the
loops. There were 2 loops which KAP couldn't parallelize. These loops
had a CDIR IVDEP, which is a cray directive telling the compiler not to
worry about data dependencies. I added a similar KAP directive
(C*$* assert no recurrence(variable list)) to give our software the same chance
as cray/sgi had.
I ran the test with 5 steps and size 64. It took 79 seconds for 1 thread and
58 seconds for 2 threads on my 2 processor 185 Mhz Cobra (about a 1.4 speedup).
I have spent some time looking at which are the CPU intensive loops (pixie)
and checking to make sure that KAP has parallelized them (annotated source
listing and cmp file).
The process I use for doing the above is to collect pixie stats for a 1 thread
run. Then look for any program lines that are not in a routine that starts
with the letters "pk". These are lines that are not in a parallel routine
(however they may be called by a parallel routine so you have to check for
this). For any loops taking a significant amount of CPU time that KAP did
not parallelize, go back to the annotated source listing to determine
why KAP didn't parallelize that loop (typically this is due to a
data dependency).
We haven't parallelized all of the loops, but it is pretty close (pixie
stats attached).
% kf77 -fkapargs='-inl -inline_optimize=3 -conc -fuse -ur=1 -real=8' pipe.f fft.f
% setenv PARALLEL 1
% unlimit
% time a.out > x.x
72.95u 1.97s 1:19 94% 0+395k 4+2556io 0pf+0w
% setenv PARALLEL 2
% time a.out > x.x
108.13u 2.58s 0:58 190% 0+400k 0+2555io 0pf+0w
Code fragments from fft.f where I added the assertion:
-----------------------------------------------------------------------
C FOR NFAX EVEN, THE REARRANGEMENT MUST BE APPLIED DIRECTLY TO
C THE INPUT ARRAY. THIS CAN BE DONE BY SWAPPING ELEMENTS.
C
IBASE = 1
ILAST = (N-1)*INK
NH = N/2
DO 20 L=1,LOT
I1 = IBASE+INK
I2 = IBASE+ILAST
CDIR$ IVDEP
C*$* assert no recurrence (A)
DO 10 M=1,NH
------------------------------------------------
CDIR$ IVDEP
C*$* assert no recurrence (WORK,I,J)
DO 50 M=2,N
C MOVE REAL AND IMAGINARY PORTIONS OF OTHER ELEMENTS (POSSIBLY IN
C REVERSE ORDER, DEPENDING ON JST AND JNK)
WORK(J) = A(I)
WORK(J+1) = A(I+1)
I=I+INK
---------------------------------------------------------------------------
In the following profile, every routine beginning with a pk is a loop
which KAP has parallelized and moved into its own subroutine.
Profile listing generated Tue Feb 18 10:30:51 1997 with:
prof -pixie -heavy a.out a.out.Addrs a.out.Counts
----------------------------------------------------------------------------
* -h[eavy] using basic-block counts; *
* sorted in descending order by the number of cycles executed in each *
* line; unexecuted lines are excluded *
----------------------------------------------------------------------------
procedure (file) line bytes cycles % cum %
pkcfftmlt_14_ (./fft.cmp.f) 2436 632 184322400 7.37 7.37
pkcfftmlt_6_ (./fft.cmp.f) 3700 628 147457320 5.89 13.26
pkcfftmlt_13_ (./fft.cmp.f) 2592 340 108627700 4.34 17.60
cfftmlt_ (./fft.cmp.f) 179 148 25968640 1.04 18.64
pkcfftmlt_5_ (./fft.cmp.f) 3850 336 24904200 1.00 19.63
pklescyl_15_ (./pipe.cmp.f) 5961 72 22302720 0.89 20.52
pklescyl_15_ (./pipe.cmp.f) 5982 52 17039360 0.68 21.20
pklescyl_15_ (./pipe.cmp.f) 6014 48 15728640 0.63 21.83
pklescyl_15_ (./pipe.cmp.f) 6074 48 15728640 0.63 22.46
pklescyl_15_ (./pipe.cmp.f) 5965 48 15728640 0.63 23.09
pklescyl_15_ (./pipe.cmp.f) 6228 44 14417920 0.58 23.66
pklescyl_15_ (./pipe.cmp.f) 6321 44 14417920 0.58 24.24
pklescyl_15_ (./pipe.cmp.f) 5998 64 13149440 0.53 24.77
pklescyl_15_ (./pipe.cmp.f) 6036 40 13107200 0.52 25.29
pklescyl_15_ (./pipe.cmp.f) 6251 40 12492800 0.50 25.79
pklescyl_15_ (./pipe.cmp.f) 6348 40 12492800 0.50 26.29
pklescyl_15_ (./pipe.cmp.f) 6467 208 12022415 0.48 26.77
pklescyl_15_ (./pipe.cmp.f) 6326 44 11796490 0.47 27.24
pklescyl_15_ (./pipe.cmp.f) 6323 40 11796485 0.47 27.71
pklescyl_15_ (./pipe.cmp.f) 5988 36 11796480 0.47 28.18
pklescyl_15_ (./pipe.cmp.f) 6045 36 11796480 0.47 28.65
pklescyl_15_ (./pipe.cmp.f) 6128 36 11243520 0.45 29.10
pklescyl_23_ (./pipe.cmp.f) 3839 152 11202880 0.45 29.55
cfftmlt_ (./fft.cmp.f) 180 56 10649600 0.43 29.98
cfftmlt_ (./fft.cmp.f) 182 56 10649600 0.43 30.40
pkcfftmlt_17_ (./fft.cmp.f) 1809 48 10485840 0.42 30.82
pkcfftmlt__ (./fft.cmp.f) 4395 48 10485840 0.42 31.24
pkcfftmlt_17_ (./fft.cmp.f) 1808 48 10485840 0.42 31.66
pkcfftmlt__ (./fft.cmp.f) 4396 48 10485840 0.42 32.08
pklescyl_15_ (./pipe.cmp.f) 6231 44 10485775 0.42 32.50
pklescyl_15_ (./pipe.cmp.f) 6019 32 10485760 0.42 32.92
pklescyl_15_ (./pipe.cmp.f) 5971 32 10485760 0.42 33.33
pklescyl_15_ (./pipe.cmp.f) 6034 32 10485760 0.42 33.75
pklescyl_15_ (./pipe.cmp.f) 6063 32 10485760 0.42 34.17
pkcfftmlt_13_ (./fft.cmp.f) 2616 44 9749840 0.39 34.56
pklescyl_15_ (./pipe.cmp.f) 6110 44 9256960 0.37 34.93
pklescyl_23_ (./pipe.cmp.f) 3842 140 9236815 0.37 35.30
pkcfftmlt_17_ (./fft.cmp.f) 1807 72 9175260 0.37 35.67
pkcfftmlt__ (./fft.cmp.f) 4394 72 9175260 0.37 36.03
pklescyl_15_ (./pipe.cmp.f) 6026 28 9175040 0.37 36.40
pklescyl_15_ (./pipe.cmp.f) 6003 28 9175040 0.37 36.77
pklescyl_15_ (./pipe.cmp.f) 6041 28 9175040 0.37 37.13
pklescyl_15_ (./pipe.cmp.f) 6091 28 9175040 0.37 37.50
pklescyl_22_ (./pipe.cmp.f) 4267 112 9031680 0.36 37.86
pklescyl_15_ (./pipe.cmp.f) 6175 124 8947840 0.36 38.22
pklescyl_18_ (./pipe.cmp.f) 5116 148 8929280 0.36 38.58
pklescyl_15_ (./pipe.cmp.f) 6395 120 8928000 0.36 38.93
pklescyl_15_ (./pipe.cmp.f) 6298 120 8928000 0.36 39.29
pklescyl_15_ (./pipe.cmp.f) 6129 40 8803520 0.35 39.64
pklescyl_15_ (./pipe.cmp.f) 6252 32 8764480 0.35 39.99
pklescyl_15_ (./pipe.cmp.f) 6349 32 8764480 0.35 40.34
pkcfftmlt_14_ (./fft.cmp.f) 2442 116 8608200 0.34 40.69
pkcfftmlt_14_ (./fft.cmp.f) 2478 28 8601600 0.34 41.03
pkcfftmlt_14_ (./fft.cmp.f) 2468 28 8601600 0.34 41.37
pkcfftmlt_13_ (./fft.cmp.f) 2597 96 8361960 0.33 41.71
pkcfftmlt_13_ (./fft.cmp.f) 2593 80 8360600 0.33 42.04
pkcfftmlt_13_ (./fft.cmp.f) 2611 56 8358560 0.33 42.37
pkcfftmlt_13_ (./fft.cmp.f) 2626 48 8357880 0.33 42.71
pkcfftmlt_13_ (./fft.cmp.f) 2606 48 8357880 0.33 43.04
pkcfftmlt_13_ (./fft.cmp.f) 2621 44 8357540 0.33 43.38
pkcfftmlt_13_ (./fft.cmp.f) 2602 40 8357200 0.33 43.71
pklescyl_15_ (./pipe.cmp.f) 6332 52 7906880 0.32 44.03
pklescyl_24_ (./pipe.cmp.f) 3748 116 7905925 0.32 44.34
pklescyl_15_ (./pipe.cmp.f) 6238 36 7905285 0.32 44.66
pklescyl_15_ (./pipe.cmp.f) 6335 36 7905285 0.32 44.97
pklescyl_22_ (./pipe.cmp.f) 4301 116 7864345 0.31 45.29
pklescyl_15_ (./pipe.cmp.f) 5977 24 7864320 0.31 45.60
pklescyl_15_ (./pipe.cmp.f) 6051 24 7864320 0.31 45.92
pklescyl_15_ (./pipe.cmp.f) 5979 24 7864320 0.31 46.23
pklescyl_15_ (./pipe.cmp.f) 6016 24 7864320 0.31 46.55
pklescyl_22_ (./pipe.cmp.f) 4304 96 7864320 0.31 46.86
pklescyl_22_ (./pipe.cmp.f) 4307 96 7864320 0.31 47.17
pklescyl_15_ (./pipe.cmp.f) 6021 24 7864320 0.31 47.49
pklescyl_15_ (./pipe.cmp.f) 6059 24 7864320 0.31 47.80
pklescyl_15_ (./pipe.cmp.f) 6022 24 7864320 0.31 48.12
pklescyl_15_ (./pipe.cmp.f) 5962 24 7864320 0.31 48.43
pklescyl_15_ (./pipe.cmp.f) 6031 24 7864320 0.31 48.75
pklescyl_22_ (./pipe.cmp.f) 4233 120 7741440 0.31 49.05
pklescyl_22_ (./pipe.cmp.f) 4242 96 7741440 0.31 49.36
pkcfftmlt_14_ (./fft.cmp.f) 2437 56 7375200 0.29 49.66
pkcfftmlt_14_ (./fft.cmp.f) 2472 36 7373700 0.29 49.95
pkcfftmlt_14_ (./fft.cmp.f) 2452 32 7373400 0.29 50.25
pkcfftmlt_14_ (./fft.cmp.f) 2492 32 7373400 0.29 50.54
pkcfftmlt_14_ (./fft.cmp.f) 2482 24 7372800 0.29 50.84
pkcfftmlt_14_ (./fft.cmp.f) 2488 24 7372800 0.29 51.13
pkcfftmlt_14_ (./fft.cmp.f) 2498 24 7372800 0.29 51.43
pkcfftmlt_14_ (./fft.cmp.f) 2462 24 7372800 0.29 51.72
pkcfftmlt_14_ (./fft.cmp.f) 2458 24 7372800 0.29 52.02
pkcfftmlt_14_ (./fft.cmp.f) 2448 24 7372800 0.29 52.31
pkcfftmlt_6_ (./fft.cmp.f) 3706 108 6886080 0.28 52.59
pkcfftmlt_6_ (./fft.cmp.f) 3739 28 6881280 0.27 52.86
pkcfftmlt_6_ (./fft.cmp.f) 3748 28 6881280 0.27 53.14
pkcfftmlt_6_ (./fft.cmp.f) 3730 28 6881280 0.27 53.41
pklescyl_15_ (./pipe.cmp.f) 6241 96 6719385 0.27 53.68
pklescyl_15_ (./pipe.cmp.f) 6338 88 6698585 0.27 53.95
pklescyl_15_ (./pipe.cmp.f) 6109 32 6615040 0.26 54.21
pklescyl_15_ (./pipe.cmp.f) 6230 40 6595520 0.26 54.47
pklescyl_15_ (./pipe.cmp.f) 6469 104 6595210 0.26 54.74
pklescyl_15_ (./pipe.cmp.f) 6068 20 6553600 0.26 55.00
pklescyl_15_ (./pipe.cmp.f) 6005 20 6553600 0.26 55.26
pklescyl_15_ (./pipe.cmp.f) 6038 20 6553600 0.26 55.52
pklescyl_15_ (./pipe.cmp.f) 6056 20 6553600 0.26 55.78
pklescyl_15_ (./pipe.cmp.f) 6054 20 6553600 0.26 56.05
pklescyl_15_ (./pipe.cmp.f) 5993 20 6553600 0.26 56.31
pklescyl_15_ (./pipe.cmp.f) 6065 20 6553600 0.26 56.57
pklescyl_15_ (./pipe.cmp.f) 6067 20 6553600 0.26 56.83
pklescyl_15_ (./pipe.cmp.f) 6047 20 6553600 0.26 57.09
pklescyl_15_ (./pipe.cmp.f) 6008 20 6553600 0.26 57.36
pklescyl_15_ (./pipe.cmp.f) 6071 20 6553600 0.26 57.62
pklescyl_15_ (./pipe.cmp.f) 6085 20 6553600 0.26 57.88
pklescyl_15_ (./pipe.cmp.f) 6093 20 6553600 0.26 58.14
pklescyl_15_ (./pipe.cmp.f) 6135 20 6246400 0.25 58.39
pklescyl_15_ (./pipe.cmp.f) 6355 20 6246400 0.25 58.64
pklescyl_15_ (./pipe.cmp.f) 6258 20 6246400 0.25 58.89
pkcfftmlt_14_ (./fft.cmp.f) 2502 20 6144000 0.25 59.14
pkcfftmlt_6_ (./fft.cmp.f) 3701 56 5900160 0.24 59.37
pkcfftmlt_6_ (./fft.cmp.f) 3733 52 5899920 0.24 59.61
pkcfftmlt_6_ (./fft.cmp.f) 3715 32 5898720 0.24 59.84
pkcfftmlt_6_ (./fft.cmp.f) 3751 32 5898720 0.24 60.08
pkcfftmlt_6_ (./fft.cmp.f) 3712 24 5898240 0.24 60.31
pkcfftmlt_6_ (./fft.cmp.f) 3721 24 5898240 0.24 60.55
pkcfftmlt_6_ (./fft.cmp.f) 3742 24 5898240 0.24 60.79
pkcfftmlt_6_ (./fft.cmp.f) 3724 24 5898240 0.24 61.02
pklescyl_23_ (./pipe.cmp.f) 3840 100 5612180 0.22 61.25
pklescyl_18_ (./pipe.cmp.f) 5121 100 5468160 0.22 61.46
pklescyl_18_ (./pipe.cmp.f) 5096 96 5346240 0.21 61.68
pklescyl_15_ (./pipe.cmp.f) 6470 104 5304985 0.21 61.89
pklescyl_15_ (./pipe.cmp.f) 6112 28 5304320 0.21 62.10
pklescyl_15_ (./pipe.cmp.f) 6235 40 5285120 0.21 62.31
pklescyl_15_ (./pipe.cmp.f) 6107 36 5284800 0.21 62.52
pklescyl_15_ (./pipe.cmp.f) 6329 32 5284480 0.21 62.74
pklescyl_15_ (./pipe.cmp.f) 6240 36 5283855 0.21 62.95
pklescyl_15_ (./pipe.cmp.f) 6337 32 5283850 0.21 63.16
pklescyl_15_ (./pipe.cmp.f) 6237 28 5283845 0.21 63.37
pklescyl_15_ (./pipe.cmp.f) 6334 28 5283845 0.21 63.58
pklescyl_24_ (./pipe.cmp.f) 3750 84 5242905 0.21 63.79
pklescyl_24_ (./pipe.cmp.f) 3751 76 5242895 0.21 64.00
pklescyl_15_ (./pipe.cmp.f) 5984 16 5242880 0.21 64.21
pklescyl_15_ (./pipe.cmp.f) 6050 16 5242880 0.21 64.42
pklescyl_15_ (./pipe.cmp.f) 5989 16 5242880 0.21 64.63
pklescyl_15_ (./pipe.cmp.f) 5976 16 5242880 0.21 64.84
pklescyl_15_ (./pipe.cmp.f) 6024 16 5242880 0.21 65.05
pklescyl_15_ (./pipe.cmp.f) 6094 16 5242880 0.21 65.26
pklescyl_15_ (./pipe.cmp.f) 6098 16 5242880 0.21 65.47
pklescyl_15_ (./pipe.cmp.f) 6039 16 5242880 0.21 65.67
pklescyl_15_ (./pipe.cmp.f) 6082 16 5242880 0.21 65.88
pklescyl_15_ (./pipe.cmp.f) 6089 16 5242880 0.21 66.09
pklescyl_15_ (./pipe.cmp.f) 5973 16 5242880 0.21 66.30
pklescyl_15_ (./pipe.cmp.f) 6023 16 5242880 0.21 66.51
pklescyl_15_ (./pipe.cmp.f) 6101 16 5242880 0.21 66.72
pklescyl_15_ (./pipe.cmp.f) 5995 16 5242880 0.21 66.93
pklescyl_18_ (./pipe.cmp.f) 5118 112 5202560 0.21 67.14
pklescyl_15_ (./pipe.cmp.f) 6131 16 4997120 0.20 67.34
pkcfftmlt_14_ (./fft.cmp.f) 2467 72 4919400 0.20 67.54
pkcfftmlt_14_ (./fft.cmp.f) 2497 40 4917000 0.20 67.73
pkcfftmlt_14_ (./fft.cmp.f) 2477 32 4916400 0.20 67.93
pkcfftmlt_6_ (./fft.cmp.f) 3757 20 4915200 0.20 68.13
pkcfftmlt_6_ (./fft.cmp.f) 3760 20 4915200 0.20 68.32
pklescyl_24_ (./pipe.cmp.f) 3749 84 4301145 0.17 68.49
pklescyl_22_ (./pipe.cmp.f) 4243 72 4273925 0.17 68.66
pklescyl_15_ (./pipe.cmp.f) 6116 40 4055360 0.16 68.83
pklescyl_21_ (./pipe.cmp.f) 4579 84 4052160 0.16 68.99
pklescyl_15_ (./pipe.cmp.f) 6095 40 3994880 0.16 69.15
pklescyl_21_ (./pipe.cmp.f) 4578 72 3991680 0.16 69.31
pkcfftmlt_6_ (./fft.cmp.f) 3747 80 3936000 0.16 69.46
pkcfftmlt_6_ (./fft.cmp.f) 3729 60 3934800 0.16 69.62
pkcfftmlt_6_ (./fft.cmp.f) 3756 40 3933600 0.16 69.78
pkcfftmlt_6_ (./fft.cmp.f) 3738 36 3933360 0.16 69.94
pklescyl_15_ (./pipe.cmp.f) 5978 12 3932160 0.16 70.09
pklescyl_15_ (./pipe.cmp.f) 6006 12 3932160 0.16 70.25
pklescyl_15_ (./pipe.cmp.f) 6029 12 3932160 0.16 70.41
pklescyl_15_ (./pipe.cmp.f) 6079 12 3932160 0.16 70.56
pklescyl_15_ (./pipe.cmp.f) 6062 12 3932160 0.16 70.72
pklescyl_15_ (./pipe.cmp.f) 5975 12 3932160 0.16 70.88
pklescyl_15_ (./pipe.cmp.f) 6060 12 3932160 0.16 71.04
pklescyl_15_ (./pipe.cmp.f) 5972 12 3932160 0.16 71.19
pklescyl_15_ (./pipe.cmp.f) 6055 12 3932160 0.16 71.35
pklescyl_15_ (./pipe.cmp.f) 6324 12 3932160 0.16 71.51
pklescyl_15_ (./pipe.cmp.f) 5970 12 3932160 0.16 71.66
pklescyl_15_ (./pipe.cmp.f) 6035 12 3932160 0.16 71.82
pklescyl_15_ (./pipe.cmp.f) 6018 12 3932160 0.16 71.98
pklescyl_15_ (./pipe.cmp.f) 5987 12 3932160 0.16 72.14
pklescyl_15_ (./pipe.cmp.f) 5980 12 3932160 0.16 72.29
pklescyl_15_ (./pipe.cmp.f) 5967 12 3932160 0.16 72.45
pklescyl_15_ (./pipe.cmp.f) 6084 12 3932160 0.16 72.61
pklescyl_15_ (./pipe.cmp.f) 6234 12 3932160 0.16 72.76
pklescyl_15_ (./pipe.cmp.f) 5990 12 3932160 0.16 72.92
pklescyl_15_ (./pipe.cmp.f) 5996 12 3932160 0.16 73.08
pklescyl_21_ (./pipe.cmp.f) 4580 60 3931200 0.16 73.24
pklescyl_15_ (./pipe.cmp.f) 6396 64 3888640 0.16 73.39
pklescyl_15_ (./pipe.cmp.f) 6299 64 3888640 0.16 73.55
pklescyl_15_ (./pipe.cmp.f) 6176 60 3868800 0.15 73.70
pklescyl_15_ (./pipe.cmp.f) 6351 12 3747840 0.15 73.85
pklescyl_15_ (./pipe.cmp.f) 6137 12 3747840 0.15 74.00
pklescyl_15_ (./pipe.cmp.f) 6254 12 3747840 0.15 74.15
pklescyl_15_ (./pipe.cmp.f) 6260 12 3747840 0.15 74.30
pklescyl_15_ (./pipe.cmp.f) 6133 12 3747840 0.15 74.45
pklescyl_15_ (./pipe.cmp.f) 6357 12 3747840 0.15 74.60
pklescyl_15_ (./pipe.cmp.f) 6353 12 3747840 0.15 74.75
pklescyl_15_ (./pipe.cmp.f) 6256 12 3747840 0.15 74.90
pkcfftmlt_14_ (./fft.cmp.f) 2487 108 3693600 0.15 75.05
pkcfftmlt_14_ (./fft.cmp.f) 2447 80 3691500 0.15 75.19
pkcfftmlt_14_ (./fft.cmp.f) 2457 44 3688800 0.15 75.34
pklescyl_15_ (./pipe.cmp.f) 6468 44 3604480 0.14 75.49
pklescyl_22_ (./pipe.cmp.f) 4300 44 3604480 0.14 75.63
pklescyl_18_ (./pipe.cmp.f) 5115 88 3235840 0.13 75.76
pklescyl_22_ (./pipe.cmp.f) 4234 84 3031695 0.12 75.88
fkdat_ (./pipe.cmp.f) 14114 52 2957312 0.12 76.00
pkcfftmlt_6_ (./fft.cmp.f) 3720 104 2954640 0.12 76.12
pkcfftmlt_6_ (./fft.cmp.f) 3711 100 2954400 0.12 76.23
pklescyl_22_ (./pipe.cmp.f) 4216 124 2929310 0.12 76.35
pklescyl_15_ (./pipe.cmp.f) 6174 48 2916480 0.12 76.47
pklescyl_15_ (./pipe.cmp.f) 6297 48 2916480 0.12 76.58
pklescyl_15_ (./pipe.cmp.f) 6394 48 2916480 0.12 76.70
pklescyl_18_ (./pipe.cmp.f) 5117 48 2887680 0.12 76.82
pklescyl_15_ (./pipe.cmp.f) 6488 104 2847385 0.11 76.93
pklescyl_15_ (./pipe.cmp.f) 6183 76 2846720 0.11 77.04
pkcfftmlt_13_ (./fft.cmp.f) 2614 8 2785280 0.11 77.16
pkcfftmlt_13_ (./fft.cmp.f) 2622 8 2785280 0.11 77.27
pkcfftmlt_13_ (./fft.cmp.f) 2624 8 2785280 0.11 77.38
pkcfftmlt_13_ (./fft.cmp.f) 2613 8 2785280 0.11 77.49
pkcfftmlt_13_ (./fft.cmp.f) 2600 8 2785280 0.11 77.60
pkcfftmlt_13_ (./fft.cmp.f) 2623 8 2785280 0.11 77.71
pkcfftmlt_13_ (./fft.cmp.f) 2612 8 2785280 0.11 77.82
pklescyl_21_ (./pipe.cmp.f) 4577 68 2761920 0.11 77.93
pklescyl_15_ (./pipe.cmp.f) 6306 56 2744320 0.11 78.04
pklescyl_15_ (./pipe.cmp.f) 6403 56 2744320 0.11 78.15
pklescyl_21_ (./pipe.cmp.f) 4581 56 2701440 0.11 78.26
pklescyl_15_ (./pipe.cmp.f) 6139 48 2693760 0.11 78.37
pklescyl_15_ (./pipe.cmp.f) 6104 24 2663040 0.11 78.47
pklescyl_16_ (./pipe.cmp.f) 5419 44 2621455 0.10 78.58
pklescyl_15_ (./pipe.cmp.f) 6322 12 2621445 0.10 78.68
pklescyl_15_ (./pipe.cmp.f) 6229 12 2621445 0.10 78.79
pklescyl_15_ (./pipe.cmp.f) 6073 8 2621440 0.10 78.89
pklescyl_15_ (./pipe.cmp.f) 6000 8 2621440 0.10 79.00
pklescyl_15_ (./pipe.cmp.f) 5986 8 2621440 0.10 79.10
pklescyl_15_ (./pipe.cmp.f) 6072 8 2621440 0.10 79.21
pklescyl_15_ (./pipe.cmp.f) 6027 8 2621440 0.10 79.31
pklescyl_15_ (./pipe.cmp.f) 6044 8 2621440 0.10 79.42
pklescyl_15_ (./pipe.cmp.f) 6330 8 2621440 0.10 79.52
pklescyl_15_ (./pipe.cmp.f) 6030 8 2621440 0.10 79.63
pklescyl_15_ (./pipe.cmp.f) 6009 8 2621440 0.10 79.73
pklescyl_24_ (./pipe.cmp.f) 3753 32 2621440 0.10 79.84
pklescyl_15_ (./pipe.cmp.f) 6007 8 2621440 0.10 79.94
pklescyl_16_ (./pipe.cmp.f) 5420 32 2621440 0.10 80.05
pklescyl_15_ (./pipe.cmp.f) 5992 8 2621440 0.10 80.15
pklescyl_15_ (./pipe.cmp.f) 6001 8 2621440 0.10 80.26
pklescyl_15_ (./pipe.cmp.f) 6042 8 2621440 0.10 80.36
pklescyl_15_ (./pipe.cmp.f) 6080 8 2621440 0.10 80.46
pklescyl_15_ (./pipe.cmp.f) 5963 8 2621440 0.10 80.57
pklescyl_15_ (./pipe.cmp.f) 6070 8 2621440 0.10 80.67
pklescyl_15_ (./pipe.cmp.f) 6471 32 2621440 0.10 80.78
pklescyl_15_ (./pipe.cmp.f) 6002 8 2621440 0.10 80.88
pklescyl_15_ (./pipe.cmp.f) 6043 8 2621440 0.10 80.99
pklescyl_15_ (./pipe.cmp.f) 6083 8 2621440 0.10 81.09
pklescyl_15_ (./pipe.cmp.f) 5968 8 2621440 0.10 81.20
pklescyl_15_ (./pipe.cmp.f) 5997 8 2621440 0.10 81.30
pklescyl_15_ (./pipe.cmp.f) 5994 8 2621440 0.10 81.41
pklescyl_15_ (./pipe.cmp.f) 6017 8 2621440 0.10 81.51
pklescyl_15_ (./pipe.cmp.f) 6015 8 2621440 0.10 81.62
fkdat_ (./pipe.cmp.f) 14145 64 2621440 0.10 81.72
pklescyl_23_ (./pipe.cmp.f) 3846 32 2621440 0.10 81.83
pklescyl_23_ (./pipe.cmp.f) 3841 32 2621440 0.10 81.93
pklescyl_15_ (./pipe.cmp.f) 6262 28 2596160 0.10 82.03
pklescyl_15_ (./pipe.cmp.f) 6359 28 2596160 0.10 82.14
pklescyl_15_ (./pipe.cmp.f) 6283 20 2560000 0.10 82.24
pklescyl_15_ (./pipe.cmp.f) 6380 20 2560000 0.10 82.34
pklescyl_15_ (./pipe.cmp.f) 6168 12 2519040 0.10 82.44
pkcfftmlt_14_ (./fft.cmp.f) 2484 8 2457600 0.10 82.54
pkcfftmlt_14_ (./fft.cmp.f) 2489 8 2457600 0.10 82.64
pkcfftmlt_14_ (./fft.cmp.f) 2495 8 2457600 0.10 82.74
pkcfftmlt_14_ (./fft.cmp.f) 2474 8 2457600 0.10 82.84
pkcfftmlt_14_ (./fft.cmp.f) 2491 8 2457600 0.10 82.93
pkcfftmlt_14_ (./fft.cmp.f) 2454 8 2457600 0.10 83.03
pkcfftmlt_14_ (./fft.cmp.f) 2481 8 2457600 0.10 83.13
pkcfftmlt_14_ (./fft.cmp.f) 2465 8 2457600 0.10 83.23
pkcfftmlt_14_ (./fft.cmp.f) 2463 8 2457600 0.10 83.33
pkcfftmlt_14_ (./fft.cmp.f) 2475 8 2457600 0.10 83.43
pkcfftmlt_14_ (./fft.cmp.f) 2451 8 2457600 0.10 83.52
pkcfftmlt_14_ (./fft.cmp.f) 2485 8 2457600 0.10 83.62
pkcfftmlt_14_ (./fft.cmp.f) 2473 8 2457600 0.10 83.72
pkcfftmlt_14_ (./fft.cmp.f) 2469 8 2457600 0.10 83.82
pkcfftmlt_14_ (./fft.cmp.f) 2505 8 2457600 0.10 83.92
pkcfftmlt_14_ (./fft.cmp.f) 2464 8 2457600 0.10 84.02
pkcfftmlt_14_ (./fft.cmp.f) 2483 8 2457600 0.10 84.11
pkcfftmlt_14_ (./fft.cmp.f) 2493 8 2457600 0.10 84.21
pkcfftmlt_14_ (./fft.cmp.f) 2479 8 2457600 0.10 84.31
pkcfftmlt_14_ (./fft.cmp.f) 2471 8 2457600 0.10 84.41
pkcfftmlt_14_ (./fft.cmp.f) 2459 8 2457600 0.10 84.51
pklescyl_14_ (./pipe.cmp.f) 6881 68 2359304 0.09 84.60
pklescyl_26_ (./pipe.cmp.f) 3379 56 2359301 0.09 84.69
pklescyl_26_ (./pipe.cmp.f) 3385 36 2359296 0.09 84.79
pklescyl_26_ (./pipe.cmp.f) 3391 36 2359296 0.09 84.88
pklescyl_14_ (./pipe.cmp.f) 6887 36 2359296 0.09 84.98
pklescyl_26_ (./pipe.cmp.f) 3366 36 2359296 0.09 85.07
pklescyl_14_ (./pipe.cmp.f) 6893 36 2359296 0.09 85.17
pklescyl_14_ (./pipe.cmp.f) 6868 36 2359296 0.09 85.26
pkcfftmlt_5_ (./fft.cmp.f) 3874 64 2294480 0.09 85.35
pkcfftmlt_5_ (./fft.cmp.f) 3860 48 2294160 0.09 85.44
pklescyl_15_ (./pipe.cmp.f) 6466 28 2293760 0.09 85.54
pklescyl_11_ (./pipe.cmp.f) 7396 172 2240581 0.09 85.62
pklescyl_22_ (./pipe.cmp.f) 4232 52 2232320 0.09 85.71
pklescyl_26_ (./pipe.cmp.f) 3367 40 2097154 0.08 85.80
pklescyl_14_ (./pipe.cmp.f) 6869 40 2097154 0.08 85.88
pklescyl_26_ (./pipe.cmp.f) 3371 32 2097152 0.08 85.97
pklescyl_14_ (./pipe.cmp.f) 6873 32 2097152 0.08 86.05
pkcfftmlt_5_ (./fft.cmp.f) 3855 104 1967680 0.08 86.13
pkcfftmlt_5_ (./fft.cmp.f) 3851 96 1967520 0.08 86.21
pkcfftmlt_6_ (./fft.cmp.f) 3702 32 1967520 0.08 86.29
pkcfftmlt_6_ (./fft.cmp.f) 3703 24 1967040 0.08 86.36
pkcfftmlt_5_ (./fft.cmp.f) 3864 48 1966560 0.08 86.44
pkcfftmlt_5_ (./fft.cmp.f) 3869 44 1966480 0.08 86.52
pkcfftmlt_5_ (./fft.cmp.f) 3879 36 1966320 0.08 86.60
pkcfftmlt_6_ (./fft.cmp.f) 3752 8 1966080 0.08 86.68
pkcfftmlt_6_ (./fft.cmp.f) 3717 8 1966080 0.08 86.76
pkcfftmlt_6_ (./fft.cmp.f) 3735 8 1966080 0.08 86.84
pkcfftmlt_6_ (./fft.cmp.f) 3749 8 1966080 0.08 86.91
pkcfftmlt_6_ (./fft.cmp.f) 3754 8 1966080 0.08 86.99
pkcfftmlt_6_ (./fft.cmp.f) 3732 8 1966080 0.08 87.07
pkcfftmlt_6_ (./fft.cmp.f) 3727 8 1966080 0.08 87.15
pkcfftmlt_6_ (./fft.cmp.f) 3741 8 1966080 0.08 87.23
pklescyl_24_ (./pipe.cmp.f) 3746 24 1966080 0.08 87.31
pkcfftmlt_6_ (./fft.cmp.f) 3731 8 1966080 0.08 87.39
pkcfftmlt_6_ (./fft.cmp.f) 3759 8 1966080 0.08 87.46
pkcfftmlt_6_ (./fft.cmp.f) 3726 8 1966080 0.08 87.54
pkcfftmlt_6_ (./fft.cmp.f) 3740 8 1966080 0.08 87.62
pkcfftmlt_6_ (./fft.cmp.f) 3744 8 1966080 0.08 87.70
pkcfftmlt_6_ (./fft.cmp.f) 3736 8 1966080 0.08 87.78
pkcfftmlt_6_ (./fft.cmp.f) 3745 8 1966080 0.08 87.86
pkcfftmlt_6_ (./fft.cmp.f) 3750 8 1966080 0.08 87.93
pkcfftmlt_6_ (./fft.cmp.f) 3743 8 1966080 0.08 88.01
pklescyl_23_ (./pipe.cmp.f) 3838 24 1966080 0.08 88.09
pkcfftmlt_6_ (./fft.cmp.f) 3734 8 1966080 0.08 88.17
pklescyl_18_ (./pipe.cmp.f) 5122 36 1658880 0.07 88.24
pkcfftmlt_5_ (./fft.cmp.f) 3884 40 1638800 0.07 88.30
pklescyl_16_ (./pipe.cmp.f) 5418 52 1638440 0.07 88.37
pklescyl_15_ (./pipe.cmp.f) 6486 20 1638400 0.07 88.43
pklescyl_18_ (./pipe.cmp.f) 5119 28 1617920 0.06 88.50
pklescyl_22_ (./pipe.cmp.f) 4241 20 1612800 0.06 88.56
pklescyl_15_ (./pipe.cmp.f) 5937 36 1589940 0.06 88.63
pklescyl_11_ (./pipe.cmp.f) 7399 128 1585220 0.06 88.69
pklescyl_12_ (./pipe.cmp.f) 7250 116 1581185 0.06 88.75
pklescyl_14_ (./pipe.cmp.f) 6877 24 1572864 0.06 88.82
pklescyl_14_ (./pipe.cmp.f) 6897 24 1572864 0.06 88.88
fkdat_ (./pipe.cmp.f) 14122 64 1572864 0.06 88.94
pklescyl_26_ (./pipe.cmp.f) 3375 24 1572864 0.06 89.00
pklescyl_26_ (./pipe.cmp.f) 3395 24 1572864 0.06 89.07
pklescyl_15_ (./pipe.cmp.f) 5938 60 1525280 0.06 89.13
fkdat_ (./pipe.cmp.f) 14144 36 1474560 0.06 89.19
fkdat_ (./pipe.cmp.f) 14150 36 1474560 0.06 89.25
pklescyl_15_ (./pipe.cmp.f) 6119 48 1455360 0.06 89.30
pklescyl_22_ (./pipe.cmp.f) 4315 48 1433925 0.06 89.36
pklescyl_22_ (./pipe.cmp.f) 4236 52 1413760 0.06 89.42
pklescyl_22_ (./pipe.cmp.f) 4244 60 1411855 0.06 89.47
pklescyl_15_ (./pipe.cmp.f) 6178 44 1408640 0.06 89.53
pkcfftmlt_13_ (./fft.cmp.f) 2596 36 1395360 0.06 89.59
pkcfftmlt_13_ (./fft.cmp.f) 2594 32 1395020 0.06 89.64
pkcfftmlt_13_ (./fft.cmp.f) 2595 28 1394680 0.06 89.70
pkcfftmlt_13_ (./fft.cmp.f) 2604 20 1394000 0.06 89.75
pkcfftmlt_13_ (./fft.cmp.f) 2605 20 1394000 0.06 89.81
pkcfftmlt_13_ (./fft.cmp.f) 2603 20 1394000 0.06 89.86
pkcfftmlt_13_ (./fft.cmp.f) 2607 4 1392640 0.06 89.92
pkcfftmlt_13_ (./fft.cmp.f) 2617 4 1392640 0.06 89.98
pkcfftmlt_13_ (./fft.cmp.f) 2627 4 1392640 0.06 90.03
pkcfftmlt_13_ (./fft.cmp.f) 2599 4 1392640 0.06 90.09
pkcfftmlt_13_ (./fft.cmp.f) 2598 4 1392640 0.06 90.14
pkcfftmlt_13_ (./fft.cmp.f) 2608 4 1392640 0.06 90.20
pkcfftmlt_13_ (./fft.cmp.f) 2609 4 1392640 0.06 90.25
pkcfftmlt_13_ (./fft.cmp.f) 2629 4 1392640 0.06 90.31
pkcfftmlt_13_ (./fft.cmp.f) 2619 4 1392640 0.06 90.37
pkcfftmlt_13_ (./fft.cmp.f) 2618 4 1392640 0.06 90.42
pkcfftmlt_13_ (./fft.cmp.f) 2628 4 1392640 0.06 90.48
pklescyl_22_ (./pipe.cmp.f) 4269 56 1391685 0.06 90.53
fkdat_ (./pipe.cmp.f) 14162 24 1376256 0.05 90.59
pklescyl_24_ (./pipe.cmp.f) 3747 36 1372485 0.05 90.64
pklescyl_22_ (./pipe.cmp.f) 4268 52 1371525 0.05 90.70
pklescyl_15_ (./pipe.cmp.f) 6177 36 1368960 0.05 90.75
pklescyl_15_ (./pipe.cmp.f) 6398 36 1368960 0.05 90.81
pklescyl_15_ (./pipe.cmp.f) 6301 36 1368960 0.05 90.86
pklescyl_15_ (./pipe.cmp.f) 6339 28 1352960 0.05 90.91
pklescyl_22_ (./pipe.cmp.f) 4310 40 1352960 0.05 90.97
pklescyl_22_ (./pipe.cmp.f) 4312 36 1352640 0.05 91.02
pklescyl_15_ (./pipe.cmp.f) 6118 24 1352640 0.05 91.08
pklescyl_22_ (./pipe.cmp.f) 4313 40 1352330 0.05 91.13
pklescyl_22_ (./pipe.cmp.f) 4311 36 1352010 0.05 91.19
pklescyl_22_ (./pipe.cmp.f) 4314 28 1352000 0.05 91.24
pklescyl_15_ (./pipe.cmp.f) 6489 40 1351700 0.05 91.29
pklescyl_15_ (./pipe.cmp.f) 6300 28 1329280 0.05 91.35
pklescyl_15_ (./pipe.cmp.f) 6397 28 1329280 0.05 91.40
pklescyl_15_ (./pipe.cmp.f) 6361 20 1327360 0.05 91.45
pklescyl_15_ (./pipe.cmp.f) 6264 20 1327360 0.05 91.51
pklescyl_11_ (./pipe.cmp.f) 7398 104 1319042 0.05 91.56
pklescyl_23_ (./pipe.cmp.f) 3843 36 1310745 0.05 91.61
pklescyl_16_ (./pipe.cmp.f) 5421 36 1310745 0.05 91.66
pklescyl_16_ (./pipe.cmp.f) 5422 32 1310740 0.05 91.72
pklescyl_24_ (./pipe.cmp.f) 3752 28 1310735 0.05 91.77
pklescyl_15_ (./pipe.cmp.f) 6011 4 1310720 0.05 91.82
pklescyl_26_ (./pipe.cmp.f) 3369 20 1310720 0.05 91.87
pklescyl_15_ (./pipe.cmp.f) 6182 16 1310720 0.05 91.92
pklescyl_15_ (./pipe.cmp.f) 6081 4 1310720 0.05 91.98
pklescyl_15_ (./pipe.cmp.f) 6010 4 1310720 0.05 92.03
pklescyl_15_ (./pipe.cmp.f) 5985 4 1310720 0.05 92.08
pklescyl_22_ (./pipe.cmp.f) 4215 16 1310720 0.05 92.13
pklescyl_23_ (./pipe.cmp.f) 3844 16 1310720 0.05 92.19
pklescyl_15_ (./pipe.cmp.f) 6040 4 1310720 0.05 92.24
pklescyl_15_ (./pipe.cmp.f) 6052 4 1310720 0.05 92.29
pklescyl_22_ (./pipe.cmp.f) 4302 16 1310720 0.05 92.34
pklescyl_15_ (./pipe.cmp.f) 6061 4 1310720 0.05 92.40
pklescyl_15_ (./pipe.cmp.f) 6115 4 1310720 0.05 92.45
pklescyl_23_ (./pipe.cmp.f) 3845 16 1310720 0.05 92.50
pklescyl_22_ (./pipe.cmp.f) 4305 16 1310720 0.05 92.55
pklescyl_15_ (./pipe.cmp.f) 5974 4 1310720 0.05 92.61
pklescyl_14_ (./pipe.cmp.f) 6871 20 1310720 0.05 92.66
pklescyl_15_ (./pipe.cmp.f) 5983 4 1310720 0.05 92.71
pklescyl_15_ (./pipe.cmp.f) 6020 4 1310720 0.05 92.76
pklescyl_18_ (./pipe.cmp.f) 5095 16 1310720 0.05 92.82
pklescyl_15_ (./pipe.cmp.f) 6325 4 1310720 0.05 92.87
pklescyl_22_ (./pipe.cmp.f) 4308 16 1310720 0.05 92.92
pklescyl_15_ (./pipe.cmp.f) 6402 16 1310720 0.05 92.97
pklescyl_15_ (./pipe.cmp.f) 6053 4 1310720 0.05 93.02
pklescyl_15_ (./pipe.cmp.f) 6100 4 1310720 0.05 93.08
pklescyl_15_ (./pipe.cmp.f) 5966 4 1310720 0.05 93.13
pklescyl_15_ (./pipe.cmp.f) 6305 16 1310720 0.05 93.18
pklescyl_15_ (./pipe.cmp.f) 6025 4 1310720 0.05 93.23
pklescyl_22_ (./pipe.cmp.f) 4266 16 1290240 0.05 93.29
pklescyl_15_ (./pipe.cmp.f) 6142 20 1249615 0.05 93.34
pklescyl_15_ (./pipe.cmp.f) 6143 16 1249610 0.05 93.39
pklescyl_15_ (./pipe.cmp.f) 6362 16 1249610 0.05 93.44
pklescyl_15_ (./pipe.cmp.f) 6265 16 1249610 0.05 93.49
pklescyl_15_ (./pipe.cmp.f) 6363 16 1249610 0.05 93.54
pklescyl_15_ (./pipe.cmp.f) 6266 16 1249610 0.05 93.59
pklescyl_15_ (./pipe.cmp.f) 6141 4 1249280 0.05 93.64
pklescyl_15_ (./pipe.cmp.f) 6350 4 1249280 0.05 93.69
pklescyl_15_ (./pipe.cmp.f) 6130 4 1249280 0.05 93.74
pklescyl_15_ (./pipe.cmp.f) 6352 4 1249280 0.05 93.78
pklescyl_15_ (./pipe.cmp.f) 6144 4 1249280 0.05 93.83
pklescyl_15_ (./pipe.cmp.f) 6145 4 1249280 0.05 93.88
pklescyl_15_ (./pipe.cmp.f) 6268 4 1249280 0.05 93.93
pklescyl_15_ (./pipe.cmp.f) 6146 4 1249280 0.05 93.98
pklescyl_15_ (./pipe.cmp.f) 6132 4 1249280 0.05 94.03
pklescyl_15_ (./pipe.cmp.f) 6253 4 1249280 0.05 94.08
pklescyl_15_ (./pipe.cmp.f) 6255 4 1249280 0.05 94.13
pklescyl_15_ (./pipe.cmp.f) 6365 4 1249280 0.05 94.18
pklescyl_15_ (./pipe.cmp.f) 6364 4 1249280 0.05 94.23
pklescyl_15_ (./pipe.cmp.f) 6366 4 1249280 0.05 94.28
pklescyl_15_ (./pipe.cmp.f) 6269 4 1249280 0.05 94.33
pklescyl_15_ (./pipe.cmp.f) 6267 4 1249280 0.05 94.38
pkcfftmlt_14_ (./fft.cmp.f) 2438 44 1231800 0.05 94.43
pkcfftmlt_14_ (./fft.cmp.f) 2440 32 1230900 0.05 94.48
pkcfftmlt_14_ (./fft.cmp.f) 2439 32 1230900 0.05 94.53
pkcfftmlt_14_ (./fft.cmp.f) 2443 20 1230000 0.05 94.58
pkcfftmlt_14_ (./fft.cmp.f) 2445 20 1230000 0.05 94.63
pkcfftmlt_14_ (./fft.cmp.f) 2444 20 1230000 0.05 94.68
pkcfftmlt_14_ (./fft.cmp.f) 2455 4 1228800 0.05 94.73
pkcfftmlt_14_ (./fft.cmp.f) 2449 4 1228800 0.05 94.78
pkcfftmlt_14_ (./fft.cmp.f) 2453 4 1228800 0.05 94.83
pkcfftmlt_14_ (./fft.cmp.f) 2461 4 1228800 0.05 94.88
pkcfftmlt_14_ (./fft.cmp.f) 2499 4 1228800 0.05 94.92
pkcfftmlt_14_ (./fft.cmp.f) 2494 4 1228800 0.05 94.97
pkcfftmlt_14_ (./fft.cmp.f) 2503 4 1228800 0.05 95.02
pkcfftmlt_14_ (./fft.cmp.f) 2504 4 1228800 0.05 95.07
pkcfftmlt_14_ (./fft.cmp.f) 2501 4 1228800 0.05 95.12
fkdat_ (./pipe.cmp.f) 14159 28 1146880 0.05 95.17
fkdat_ (./pipe.cmp.f) 14156 28 1146880 0.05 95.21
fkdat_ (./pipe.cmp.f) 14153 28 1146880 0.05 95.26
pklescyl_25_ (./pipe.cmp.f) 3618 108 1089604 0.04 95.30
pklescyl_13_ (./pipe.cmp.f) 7120 108 1089604 0.04 95.35
pklescyl_25_ (./pipe.cmp.f) 3617 100 1085570 0.04 95.39
pklescyl_13_ (./pipe.cmp.f) 7119 100 1085570 0.04 95.43
pklescyl_25_ (./pipe.cmp.f) 3616 96 1085506 0.04 95.48
pklescyl_13_ (./pipe.cmp.f) 7118 96 1085506 0.04 95.52
pklescyl_25_ (./pipe.cmp.f) 3615 88 1085504 0.04 95.56
pklescyl_13_ (./pipe.cmp.f) 7117 88 1085504 0.04 95.61
pklescyl_28_ (./pipe.cmp.f) 2889 124 1077382 0.04 95.65
pklescyl_26_ (./pipe.cmp.f) 3407 24 1056768 0.04 95.69
fkdat_ (./pipe.cmp.f) 14109 20 1052672 0.04 95.73
pklescyl_14_ (./pipe.cmp.f) 6909 20 1052672 0.04 95.78
pklescyl_12_ (./pipe.cmp.f) 7252 84 1048581 0.04 95.82
pklescyl_12_ (./pipe.cmp.f) 7253 76 1048579 0.04 95.86
pklescyl_14_ (./pipe.cmp.f) 6870 20 1048577 0.04 95.90
pklescyl_26_ (./pipe.cmp.f) 3368 20 1048577 0.04 95.94
fkdat_ (./pipe.cmp.f) 14119 16 1048576 0.04 95.98
pklescyl_26_ (./pipe.cmp.f) 3403 16 1048576 0.04 96.03
pklescyl_26_ (./pipe.cmp.f) 3399 16 1048576 0.04 96.07
pklescyl_26_ (./pipe.cmp.f) 3372 16 1048576 0.04 96.11
pklescyl_14_ (./pipe.cmp.f) 6874 16 1048576 0.04 96.15
pklescyl_14_ (./pipe.cmp.f) 6901 16 1048576 0.04 96.19
fkdat_ (./pipe.cmp.f) 14163 16 1048576 0.04 96.24
pklescyl_14_ (./pipe.cmp.f) 6905 16 1048576 0.04 96.28
pkcfftmlt_6_ (./fft.cmp.f) 3704 20 984000 0.04 96.32
pkcfftmlt_6_ (./fft.cmp.f) 3708 20 984000 0.04 96.36
pkcfftmlt_6_ (./fft.cmp.f) 3707 16 983760 0.04 96.40
pkcfftmlt_6_ (./fft.cmp.f) 3709 16 983760 0.04 96.44
pkcfftmlt_6_ (./fft.cmp.f) 3761 4 983040 0.04 96.47
pkcfftmlt_6_ (./fft.cmp.f) 3762 4 983040 0.04 96.51
pkcfftmlt_6_ (./fft.cmp.f) 3723 4 983040 0.04 96.55
pkcfftmlt_6_ (./fft.cmp.f) 3718 4 983040 0.04 96.59
pkcfftmlt_6_ (./fft.cmp.f) 3714 4 983040 0.04 96.63
pkcfftmlt_6_ (./fft.cmp.f) 3722 4 983040 0.04 96.67
pkcfftmlt_6_ (./fft.cmp.f) 3725 4 983040 0.04 96.71
fkdat_ (./pipe.cmp.f) 14148 24 983040 0.04 96.75
pkcfftmlt_6_ (./fft.cmp.f) 3713 4 983040 0.04 96.79
pkcfftmlt_6_ (./fft.cmp.f) 3763 4 983040 0.04 96.83
pkcfftmlt_6_ (./fft.cmp.f) 3716 4 983040 0.04 96.87
fkdat_ (./pipe.cmp.f) 14147 24 983040 0.04 96.91
pkcfftmlt_6_ (./fft.cmp.f) 3753 4 983040 0.04 96.95
pkcfftmlt_6_ (./fft.cmp.f) 3758 4 983040 0.04 96.99
pklescyl_12_ (./pipe.cmp.f) 7251 84 860229 0.03 97.02
fkdat_ (./pipe.cmp.f) 14121 20 819200 0.03 97.05
fkdat_ (./pipe.cmp.f) 14111 60 790539 0.03 97.08
pklescyl_26_ (./pipe.cmp.f) 3382 32 786437 0.03 97.12
pklescyl_14_ (./pipe.cmp.f) 6879 20 786434 0.03 97.15
pklescyl_14_ (./pipe.cmp.f) 6885 20 786434 0.03 97.18
pklescyl_26_ (./pipe.cmp.f) 3377 20 786434 0.03 97.21
pklescyl_14_ (./pipe.cmp.f) 6884 20 786434 0.03 97.24
pklescyl_26_ (./pipe.cmp.f) 3383 16 786433 0.03 97.27
fkdat_ (./pipe.cmp.f) 14160 12 786432 0.03 97.30
pklescyl_14_ (./pipe.cmp.f) 6882 12 786432 0.03 97.34
pklescyl_14_ (./pipe.cmp.f) 6875 12 786432 0.03 97.37
pklescyl_14_ (./pipe.cmp.f) 6896 12 786432 0.03 97.40
pklescyl_14_ (./pipe.cmp.f) 6886 12 786432 0.03 97.43
pklescyl_14_ (./pipe.cmp.f) 6890 12 786432 0.03 97.46
pklescyl_11_ (./pipe.cmp.f) 7400 48 786432 0.03 97.49
pklescyl_26_ (./pipe.cmp.f) 3380 12 786432 0.03 97.52
pklescyl_26_ (./pipe.cmp.f) 3388 12 786432 0.03 97.56
pklescyl_26_ (./pipe.cmp.f) 3373 12 786432 0.03 97.59
pklescyl_26_ (./pipe.cmp.f) 3394 12 786432 0.03 97.62
pklescyl_26_ (./pipe.cmp.f) 3384 12 786432 0.03 97.65
fkdat_ (./pipe.cmp.f) 14113 20 741378 0.03 97.68
cfftmlt_ (./fft.cmp.f) 173 64 737420 0.03 97.71
pklescyl_14_ (./pipe.cmp.f) 6763 44 675904 0.03 97.74
pklescyl_26_ (./pipe.cmp.f) 3261 44 675904 0.03 97.76
pklescyl_26_ (./pipe.cmp.f) 3259 40 675840 0.03 97.79
pklescyl_26_ (./pipe.cmp.f) 3257 40 675840 0.03 97.82
pklescyl_14_ (./pipe.cmp.f) 6761 40 675840 0.03 97.84
pklescyl_14_ (./pipe.cmp.f) 6759 40 675840 0.03 97.87
pklescyl_26_ (./pipe.cmp.f) 3255 40 675840 0.03 97.90
pklescyl_14_ (./pipe.cmp.f) 6757 40 675840 0.03 97.92
pkcfftmlt_5_ (./fft.cmp.f) 3854 24 655680 0.03 97.95
pkcfftmlt_5_ (./fft.cmp.f) 3872 8 655360 0.03 97.98
pkcfftmlt_5_ (./fft.cmp.f) 3870 8 655360 0.03 98.00
pkcfftmlt_5_ (./fft.cmp.f) 3881 8 655360 0.03 98.03
fkdat_ (./pipe.cmp.f) 14157 16 655360 0.03 98.06
pkcfftmlt_5_ (./fft.cmp.f) 3882 8 655360 0.03 98.08
pkcfftmlt_5_ (./fft.cmp.f) 3880 8 655360 0.03 98.11
pkcfftmlt_5_ (./fft.cmp.f) 3858 8 655360 0.03 98.13
pklescyl_13_ (./pipe.cmp.f) 7116 68 593920 0.02 98.16
pklescyl_25_ (./pipe.cmp.f) 3614 68 593920 0.02 98.18
fkdat_ (./pipe.cmp.f) 14127 24 589824 0.02 98.21
fkdat_ (./pipe.cmp.f) 14133 24 589824 0.02 98.23
fkdat_ (./pipe.cmp.f) 14130 24 589824 0.02 98.25
fkdat_ (./pipe.cmp.f) 14136 24 589824 0.02 98.28
fkdat_ (./pipe.cmp.f) 14125 24 589824 0.02 98.30
pklescyl_26_ (./pipe.cmp.f) 3260 92 570374 0.02 98.32
pklescyl_14_ (./pipe.cmp.f) 6762 92 570374 0.02 98.34
pklescyl_27_ (./pipe.cmp.f) 2989 112 561415 0.02 98.37
pklescyl_26_ (./pipe.cmp.f) 3256 60 553474 0.02 98.39
pklescyl_14_ (./pipe.cmp.f) 6758 60 553474 0.02 98.41
pklescyl_14_ (./pipe.cmp.f) 6756 52 553472 0.02 98.43
pklescyl_26_ (./pipe.cmp.f) 3254 52 553472 0.02 98.46
pklescyl_26_ (./pipe.cmp.f) 3258 56 553410 0.02 98.48
pklescyl_14_ (./pipe.cmp.f) 6760 56 553410 0.02 98.50
pklescyl_14_ (./pipe.cmp.f) 6891 8 524288 0.02 98.52
pklescyl_14_ (./pipe.cmp.f) 6872 8 524288 0.02 98.54
pklescyl_26_ (./pipe.cmp.f) 3389 8 524288 0.02 98.56
pklescyl_14_ (./pipe.cmp.f) 6883 8 524288 0.02 98.58
pklescyl_26_ (./pipe.cmp.f) 3370 8 524288 0.02 98.60
pklescyl_26_ (./pipe.cmp.f) 3381 8 524288 0.02 98.63
pklescyl_14_ (./pipe.cmp.f) 6880 8 524288 0.02 98.65
pklescyl_26_ (./pipe.cmp.f) 3378 8 524288 0.02 98.67
pklescyl_11_ (./pipe.cmp.f) 7404 32 524288 0.02 98.69
pklescyl_12_ (./pipe.cmp.f) 7255 32 524288 0.02 98.71
fkdat_ (./pipe.cmp.f) 14124 20 491520 0.02 98.73
fkdat_ (./pipe.cmp.f) 14151 12 491520 0.02 98.75
fkdat_ (./pipe.cmp.f) 14154 12 491520 0.02 98.77
pklescyl_26_ (./pipe.cmp.f) 3253 24 405504 0.02 98.78
pklescyl_14_ (./pipe.cmp.f) 6755 24 405504 0.02 98.80
pklescyl_12_ (./pipe.cmp.f) 7248 24 393216 0.02 98.82
pklescyl_11_ (./pipe.cmp.f) 7395 24 393216 0.02 98.83
fkdat_ (./pipe.cmp.f) 14134 16 393216 0.02 98.85
pklescyl_22_ (./pipe.cmp.f) 4235 24 368660 0.01 98.86
pkcfftmlt_5_ (./fft.cmp.f) 3852 28 328160 0.01 98.88
pkcfftmlt_5_ (./fft.cmp.f) 3861 20 328000 0.01 98.89
pkcfftmlt_5_ (./fft.cmp.f) 3862 20 328000 0.01 98.90
pkcfftmlt_5_ (./fft.cmp.f) 3853 20 328000 0.01 98.92
pkcfftmlt_5_ (./fft.cmp.f) 3863 20 328000 0.01 98.93
pkcfftmlt_5_ (./fft.cmp.f) 3865 4 327680 0.01 98.94
pkcfftmlt_5_ (./fft.cmp.f) 3856 4 327680 0.01 98.95
pkcfftmlt_5_ (./fft.cmp.f) 3877 4 327680 0.01 98.97
pkcfftmlt_5_ (./fft.cmp.f) 3871 4 327680 0.01 98.98
pkcfftmlt_5_ (./fft.cmp.f) 3866 4 327680 0.01 98.99
pkcfftmlt_5_ (./fft.cmp.f) 3885 4 327680 0.01 99.01
pkcfftmlt_5_ (./fft.cmp.f) 3886 4 327680 0.01 99.02
pkcfftmlt_5_ (./fft.cmp.f) 3857 4 327680 0.01 99.03
pkcfftmlt_5_ (./fft.cmp.f) 3887 4 327680 0.01 99.05
pkcfftmlt_5_ (./fft.cmp.f) 3875 4 327680 0.01 99.06
pkcfftmlt_5_ (./fft.cmp.f) 3867 4 327680 0.01 99.07
pkcfftmlt_5_ (./fft.cmp.f) 3876 4 327680 0.01 99.09
pklescyl_27_ (./pipe.cmp.f) 2973 36 317988 0.01 99.10
pklescyl_11_ (./pipe.cmp.f) 7370 36 317988 0.01 99.11
fkdat_ (./pipe.cmp.f) 14116 8 307200 0.01 99.12
pklescyl_27_ (./pipe.cmp.f) 2974 76 305060 0.01 99.14
pklescyl_11_ (./pipe.cmp.f) 7371 60 305056 0.01 99.15
fkdat_ (./pipe.cmp.f) 14128 12 294912 0.01 99.16
fkdat_ (./pipe.cmp.f) 14131 12 294912 0.01 99.17
fkdat_ (./pipe.cmp.f) 14139 12 294912 0.01 99.18
pklescyl_12_ (./pipe.cmp.f) 7249 36 274497 0.01 99.19
pklescyl_12_ (./pipe.cmp.f) 7254 28 262147 0.01 99.20
pklescyl_14_ (./pipe.cmp.f) 6892 8 262145 0.01 99.21
pklescyl_26_ (./pipe.cmp.f) 3390 8 262145 0.01 99.23
pklescyl_14_ (./pipe.cmp.f) 6894 4 262144 0.01 99.24
pklescyl_14_ (./pipe.cmp.f) 6889 4 262144 0.01 99.25
pklescyl_14_ (./pipe.cmp.f) 6888 4 262144 0.01 99.26
pklescyl_11_ (./pipe.cmp.f) 7402 16 262144 0.01 99.27
pklescyl_14_ (./pipe.cmp.f) 6876 4 262144 0.01 99.28
pklescyl_26_ (./pipe.cmp.f) 3392 4 262144 0.01 99.29
pklescyl_14_ (./pipe.cmp.f) 6878 4 262144 0.01 99.30
pklescyl_14_ (./pipe.cmp.f) 6898 4 262144 0.01 99.31
pklescyl_26_ (./pipe.cmp.f) 3387 4 262144 0.01 99.32
pklescyl_26_ (./pipe.cmp.f) 3386 4 262144 0.01 99.33
pklescyl_26_ (./pipe.cmp.f) 3374 4 262144 0.01 99.34
pklescyl_26_ (./pipe.cmp.f) 3376 4 262144 0.01 99.35
pklescyl_27_ (./pipe.cmp.f) 2988 16 262144 0.01 99.36
pklescyl_11_ (./pipe.cmp.f) 7403 16 262144 0.01 99.37
pklescyl_26_ (./pipe.cmp.f) 3396 4 262144 0.01 99.38
pklescyl_28_ (./pipe.cmp.f) 2888 16 212992 0.01 99.39
pklescyl_18_ (./pipe.cmp.f) 5108 44 184960 0.01 99.40
cfftmlt_ (./fft.cmp.f) 175 28 163940 0.01 99.40
cfftmlt_ (./fft.cmp.f) 174 12 163860 0.01 99.41
pklescyl_15_ (./pipe.cmp.f) 6150 28 143360 0.01 99.42
pklescyl_15_ (./pipe.cmp.f) 6149 28 143360 0.01 99.42
pklescyl_22_ (./pipe.cmp.f) 4285 112 143360 0.01 99.43
pklescyl_15_ (./pipe.cmp.f) 6369 28 143360 0.01 99.43
pklescyl_15_ (./pipe.cmp.f) 6273 28 143360 0.01 99.44
pklescyl_15_ (./pipe.cmp.f) 6370 28 143360 0.01 99.45
pklescyl_15_ (./pipe.cmp.f) 6272 28 143360 0.01 99.45
pklescyl_21_ (./pipe.cmp.f) 4585 28 141120 0.01 99.46
pklescyl_22_ (./pipe.cmp.f) 4335 148 137620 0.01 99.46
|
| > -< KAP is parallelizing the code on our systems >-
>
>There is a difference between not parallelizing and not speeding up. With
>the switches I used (given below) KAP was able to parallelize almost all of the
>loops. There were 2 loops which KAP couldn't parallelize. These loops
>had a CDIR IVDEP, which is a cray directive telling the compiler not to
>worry about data dependencies. I added a similar KAP directive
>(C*$* assert no recurrence(variable list)) to give our software the same chance
>as cray/sgi had.
>
Thanks, Karen, for your help. We are losing sales in the academy (mostly to
SGI) because they have people whose time is allocated for customer benchmarks,
and we (as far as I know) don't.
I have run the modified code on an AlphaServer 4100 5/400 with 2 CPUs. I got
a 20% speedup on one of the jobs (50 steps, NSIZE=64), which is good (SGI got
40% on 4 CPUs; We don't have a system with 4 CPUs available, but this
result is what the prospect expected, so it is enough). We are waiting for
the customer's decision (either AlphaServer 4100s or SGI Origin 2000s).
Thanks again,
Y.
BTW:
We will not work on this anymore, however, when running the 100x100x100 job
(20 steps), we got:
DECthreads bugcheck (version V3.13-435), terminating execution.
Failure initializing the manager thread tcb
(It didn't happen for the smaller datasizes, and it might be similar to
note 335, but we don't have time to test it right now).
|
| >
>We have received this benchmark from a customer who has an Alpha farm
>(of 12 DEC3000/400 connected by a Gigaswitch/FDDI). This customer will
>buy 3 AlphaServer 4100 (or Origin 2000s) based on this benchmark.
>
About two weeks ago we got a $125K order from this customer. Based on the
results, he decided that we have better perf the SGI, but also decided
to start with only one 4100, 4 466MHz CPUs, lots of memory and disks,
and connect it to his existing Gigaswitch.
Thanks again, Karen.
Regards,
Yossi.
|