ppc-gpu and ppc-cpu are ppc for GPU v47
ppc-asm and ppc-ori are ppc in Assembly v7

used 0.1651 x5 DOM radius with AHA
options ASENS, TILT, MKOW, ANGW, LONG were disabled in the ppc for GPU
run on GeForce GTX 295 1.296 GHz (device 5) and i7-920 2.67 GHz


f2k muon:

ppc-gpu: 54414
Device time: 16089.9 (in-kernel: 15932.7...16078.8) [ms]

real	0m16.731s
user	0m2.480s
sys	0m0.748s

ppc-cpu: 54679

real	43m44.305s
user	43m44.044s
sys	0m0.388s

ppc-asm: 54201

real	32m1.581s
user	32m1.652s
sys	0m0.056s

ppc-ori: 54572

real	118m14.072s
user	118m13.095s
sys	0m0.960s


flasher:

ppc-gpu: 1260983
Device time: 39761.7 (in-kernel: 39511.9...39728.3) [ms]

real	0m40.012s
user	0m2.872s
sys	0m0.216s

ppc-cpu: 1260828

real	98m16.291s
user	98m15.488s
sys	0m1.136s

ppc-asm: 1244803

real	78m45.683s
user	78m45.183s
sys	0m0.924s

ppc-ori: 1258907

real	204m45.798s
user	204m40.235s
sys	0m6.288s


on Tesla C1060 1.296 GHz:
Running on 30 MPs x 448 threads
Kernel uses: l=0 r=35 s=3992 c=62400

f2k muon: 54549
Device time: 21925.3 (in-kernel: 21275.7...21913.2) [ms]

real	0m23.414s
user	0m2.656s
sys	0m1.106s

flasher: 1259636
Device time: 55561.8 (in-kernel: 55178.2...55526.2) [ms]

real	0m56.615s
user	0m2.058s
sys	0m0.897s


on Tesla C2050 1.147 GHz:
Running on 14 MPs x 768 threads
Kernel uses: l=0 r=41 s=3960 c=62400

f2k muon: 54251
Device time: 14199.1 (in-kernel: 14116.4...14200.5) [ms]

real	0m17.514s
user	0m2.879s
sys	0m2.971s

flasher: 1260565
Device time: 34892.5 (in-kernel: 34768.1...34895.8) [ms]

real	0m37.505s
user	0m2.037s
sys	0m2.440s


on GeForce GTX 480 1.401 GHz:
Running on 15 MPs x 768 threads
Kernel uses: l=0 r=40 s=3960 c=62400

f2k muon: 54067
Device time: 10825.0 (in-kernel: 10767.6...10828.2) [ms]

real	0m12.884s
user	0m2.058s
sys	0m1.694s

flasher: 1260004
Device time: 26533.1 (in-kernel: 26438.7...26539.8) [ms]

real	0m28.030s
user	0m2.072s
sys	0m1.427s


script:

time cat mmc.1.f2k | ./gpu/ppc-gpu 5 | grep HIT | wc
time cat mmc.1.f2k | ./gpu/ppc-cpu 5 | grep HIT | wc
time cat mmc.1.f2k | ./appc 0 | grep HIT | wc
time cat mmc.1.f2k | ./ppc 0 | grep HIT | wc

time WFLA=405 ./gpu/ppc-gpu 63 20 1.e9 5 | wc
time WFLA=405 ./gpu/ppc-cpu 63 20 1.e9 5 | wc
time ./appc 63 20 1.e9 0 | grep -v 'HIT 63 20 ' | wc
time ./ppc 63 20 1.e9 0 | wc

awk 'BEGIN {
  mc=43*60+44.305; ma=32*60+1.581;  mo=118*60+14.072;
  fc=98*60+16.291; fa=78*60+45.683; fo=204*60+45.798;
  mt=23.414; mg=16.731; mf=17.514; mx=12.884;
  ft=56.615; fg=40.012; ff=37.505; fx=28.030;
  printf "Original:\t1/%3.2f\t1/%3.2f\n", fo/fc, mo/mc
  printf "CPU c++:\t%3.2f\t%3.2f\n", fc/fc, mc/mc
  printf "Assembly:\t%3.2f\t%3.2f\n", fc/fa, mc/ma
  printf "GTX 295:\t%3.0f\t%3.0f\n", fc/fg, mc/mg
  printf "GTX/Ori:\t%3.0f\t%3.0f\n", fo/fg, mo/mg
  printf "C1060:  \t%3.0f\t%3.0f\n", fc/ft, mc/mt
  printf "C2050:  \t%3.0f\t%3.0f\n", fc/ff, mc/mf
  printf "GTX 480:\t%3.0f\t%3.0f\n", fc/fx, mc/mx
}'

Original:	1/2.08	1/2.70
CPU c++:	1.00	1.00
Assembly:	1.25	1.37
GTX 295:	147	157
GTX/Ori:	307	424
C1060:  	104	112
C2050:  	157	150
GTX 480:	210	204