Superpage Benchmarking Ideas

This is a junk page for keeping ideas and notes about superpage benchmarking.

double size pte

fragmentation

run trans or similar many times, clear page counts each time and see if over time we end up with less allocations/more failures.

analyse similarly with the mem_map dumps

induce fragmentation: PeterChubb mentioned network stack has a habit of grabbing single pages. otherwise make a memory fragmenter

profile

see where we are spending time

lmbench

superpages has some issues; i can't yet get multiple lmbench runs out without crashing.

                 L M B E N C H  2 . 0   S U M M A R Y
                 ------------------------------------

Processor, Processes - times in microseconds - smaller is better
----------------------------------------------------------------
                                 null     null                       open    signal   signal    fork    execve  /bin/sh
kernel                           call      I/O     stat    fstat    close   install   handle  process  process  process
-----------------------------  -------  -------  -------  -------  -------  -------  -------  -------  -------  -------
2.6.15.1                         0.035    0.239    1.296    0.315    2.706    0.308    1.671     71.5    362.4   1723.0
  s.d. (5 runs)                  0.000  5.00000    0.005    0.000    0.032    0.000    0.019      0.0     17.7     36.2
2.6.15.1-lvhpt                   0.035  0.24091    1.303    0.316    2.694    0.300    1.663     75.3    389.2   1792.1
  s.d. (5 runs)                  0.000  0.00216    0.008    0.000    0.022    0.000    0.008      0.0     32.9     57.1
2.6.15.1-pte                     0.035  0.23792    1.312    0.315    2.783    0.295    1.699     76.1    373.1   1765.4
  s.d. (5 runs)                  0.000  9.08295    0.009    0.000    0.158    0.001    0.017      0.0      7.1     19.9
2.6.15.1-sp                      0.035  0.24045    1.304    0.315    2.709    0.297    1.699     76.2    446.8   1838.3
  s.d. (2 runs)                  0.000        0    0.000    0.000    0.000    0.000    0.000      0.0      0.0      0.0

File select - times in microseconds - smaller is better
-------------------------------------------------------
                                select   select   select   select   select   select   select   select
kernel                           10 fd   100 fd   250 fd   500 fd   10 tcp  100 tcp  250 tcp  500 tcp
-----------------------------  -------  -------  -------  -------  -------  -------  -------  -------
2.6.15.1                         1.075    6.257   14.842   29.249    1.570  10.9343  26.4194  52.1962
  s.d.                           0.001    0.054    0.131    0.315    0.003  0.00732  0.00681  0.01550
2.6.15.1-lvhpt                   1.048    6.192   14.720   28.904    1.535  10.8704  26.3714  52.1452
  s.d.                           0.001    0.010    0.047    0.135    0.002  0.00443  0.00586  0.00845
2.6.15.1-pte                     1.081    6.340   15.152   29.412    1.564  10.8974  26.4105  52.1716
  s.d.                           0.027    0.197    0.448    1.065    0.021  0.00450  0.00401  0.00786
2.6.15.1-sp                      1.100    6.354   15.046   29.476    1.579  10.9364  26.4279  52.1981
  s.d.                           0.000    0.000    0.000    0.000    0.000        0        0        0

Context switching with 0K - times in microseconds - smaller is better
---------------------------------------------------------------------
                                2proc/0k   4proc/0k   8proc/0k  16proc/0k  32proc/0k  64proc/0k  96proc/0k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           1.394      3.172      3.244      2.866      2.462      3.186      3.424
  s.d.                             0.043      2.376      1.003      0.898      0.348      0.574      0.316
2.6.15.1-lvhpt                     1.490      3.298      2.940      2.306      2.254      2.264      2.260
  s.d.                             0.039      2.417      1.211      0.591      0.408      0.143      0.203
2.6.15.1-pte                       3.272      4.152      3.834      2.746      2.238      2.132      2.318
  s.d.                             3.918      2.380      0.019      0.015      0.029      0.104      0.267
2.6.15.1-sp                        1.410      1.410      1.510      1.570      2.200      2.100      2.390
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

Context switching with 4K - times in microseconds - smaller is better
---------------------------------------------------------------------
                                2proc/4k   4proc/4k   8proc/4k  16proc/4k  32proc/4k  64proc/4k  96proc/4k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           3.528      5.380      3.754      2.986      2.924      4.178      4.668
  s.d.                             3.948      1.991      0.975      0.474      0.229      0.295      0.529
2.6.15.1-lvhpt                     1.852      4.536      3.852      3.084      2.776      3.010      3.638
  s.d.                             0.026      2.411      1.833      0.916      0.447      0.186      0.788
2.6.15.1-pte                       3.618      2.828      2.558      2.642      2.566      2.854      3.338
  s.d.                             4.033      1.986      0.980      0.547      0.254      0.110      0.640
2.6.15.1-sp                        1.810      1.840      4.200      3.240      2.850      2.590      2.810
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

Context switching with 8K - times in microseconds - smaller is better
---------------------------------------------------------------------
                                2proc/8k   4proc/8k   8proc/8k  16proc/8k  32proc/8k  64proc/8k  96proc/8k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           2.072      2.148      2.360      2.798      3.252      4.228      4.792
  s.d.                             0.018      0.055      0.067      0.433      0.231      0.406      0.524
2.6.15.1-lvhpt                     2.126      2.240      2.876      2.852      3.054      3.290      3.810
  s.d.                             0.018      0.039      0.953      0.436      0.263      0.116      0.733
2.6.15.1-pte                       3.938      4.898      3.722      3.346      3.196      3.634      4.236
  s.d.                             3.937      2.399      1.189      0.613      0.302      0.232      0.594
2.6.15.1-sp                        2.110      2.170      2.400      2.710      2.880      3.470      3.290
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

Context switching with 16K - times in microseconds - smaller is better
----------------------------------------------------------------------
                               2proc/16k  4proc/16k  8proc/16k  16prc/16k  32prc/16k  64prc/16k  96prc/16k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           2.624      3.586      3.816      4.240      4.424      5.620      6.136
  s.d.                             0.015      2.015      1.182      0.512      0.175      0.552      0.438
2.6.15.1-lvhpt                     2.650      2.754      3.428      3.780      4.108      4.518      5.066
  s.d.                             0.026      0.025      0.930      0.329      0.114      0.469      0.610
2.6.15.1-pte                       2.672      2.824      3.420      3.700      4.166      4.954      5.726
  s.d.                             0.011      0.024      0.934      0.353      0.259      0.581      0.644
2.6.15.1-sp                       11.500      7.130      5.090      4.560      4.390      5.320      5.950
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

Context switching with 32K - times in microseconds - smaller is better
----------------------------------------------------------------------
                               2proc/32k  4proc/32k  8proc/32k  16prc/32k  32prc/32k  64prc/32k  96prc/32k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           3.654      3.918      4.998      5.802      6.202      7.050      8.574
  s.d.                             0.030      0.029      0.027      0.011      0.169      0.783      2.065
2.6.15.1-lvhpt                     3.736      4.038      6.288      6.122      6.056      6.892      7.544
  s.d.                             0.027      0.056      0.751      0.253      0.127      0.535      1.546
2.6.15.1-pte                       3.788      4.966      5.432      5.858      5.978      7.106      8.740
  s.d.                             0.071      1.926      0.838      0.426      0.286      0.225      2.674
2.6.15.1-sp                        3.760      3.990      5.100      6.490      6.190      6.140      6.810
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

Context switching with 64K - times in microseconds - smaller is better
----------------------------------------------------------------------
                               2proc/64k  4proc/64k  8proc/64k  16prc/64k  32prc/64k  64prc/64k  96prc/64k
kernel                         ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch  ctx swtch
-----------------------------  ---------  ---------  ---------  ---------  ---------  ---------  ---------
2.6.15.1                           5.834      8.020      9.074      9.650     10.202     15.744     22.986
  s.d.                             0.036      0.057      0.015      0.321      0.149      3.656     19.157
2.6.15.1-lvhpt                     6.100      7.978      9.156      9.460      9.950     17.902     34.348
  s.d.                             0.146      0.090      0.036      0.332      0.120      5.026     20.734
2.6.15.1-pte                       6.182      8.042      9.210      9.380      9.796     14.582     22.698
  s.d.                             0.218      0.038      0.019      0.010      0.159      5.506     17.619
2.6.15.1-sp                        6.020      7.940      9.160      9.320      9.720     18.970     47.460
  s.d.                             0.000      0.000      0.000      0.000      0.000      0.000      0.000

File create/delete and VM system latencies in microseconds - smaller is better
----------------------------------------------------------------------------
                                 0K       0K       1K       1K       4K       4K      10K      10K     Mmap     Prot    Page
kernel                         Create   Delete   Create   Delete   Create   Delete   Create   Delete   Latency  Fault   Fault
------------------------------ -------  -------  -------  -------  -------  -------  -------  -------  -------  ------  ------
2.6.15.1                          0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00   1495.2   0.787    1.00
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00      4.9   0.018    0.00
2.6.15.1-lvhpt                    0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00   1844.4   0.832    1.00
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00      6.9   0.007    0.00
2.6.15.1-pte                      0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00   1863.6   0.796    1.00
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00     21.0   0.014    0.00
2.6.15.1-sp                       0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00   1844.0   0.802    1.00
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00      0.0   0.000    0.00

*Local* Communication latencies in microseconds - smaller is better
-------------------------------------------------------------------
kernel                           Pipe   AF/Unix     UDP   RPC/UDP     TCP   RPC/TCP  TCPconn
-----------------------------  -------  -------  -------  -------  -------  -------  -------
2.6.15.1                        90.727   49.233  66.2380  37.0575  60.6048  58.0874   37.781
  s.d.                           4.444    0.622  10.8649  3.72030  24.2038  14.9277    4.812
2.6.15.1-lvhpt                  93.986   50.529  66.7851  35.4007  65.4921  60.1085   39.369
  s.d.                           9.356    2.941  21.2655  0.25917  18.8449  16.8485    5.842
2.6.15.1-pte                    87.069   48.924  56.7465  35.6185  52.4286  59.6093   38.947
  s.d.                          14.144    1.833  17.2087  0.12100  25.1215  16.8455    5.862
2.6.15.1-sp                     99.612   52.960  29.5767  31.7871  79.6326  72.6858   35.869
  s.d.                           0.000    0.000        0        0        0        0    0.000

*Local* Communication bandwidths in MB/s - bigger is better
-----------------------------------------------------------
                                                             File     Mmap    Bcopy    Bcopy   Memory   Memory
kernel                           Pipe   AF/Unix    TCP     reread   reread   (libc)   (hand)     read    write
-----------------------------  -------  -------  -------  -------  -------  -------  -------  -------  -------
2.6.15.1                        599.92  1040.44  1563.56  1483.90   751.01   704.89   482.76   751.05   675.76
  s.d.                           41.18   167.53   285.83     2.89     0.01    16.33     2.70     0.09     0.18
2.6.15.1-lvhpt                  609.87   939.47  1559.23  1481.92   750.92   702.73   482.43   751.00   673.51
  s.d.                           69.60    85.51   290.42     1.93     0.08    15.22     3.90     0.08     0.10
2.6.15.1-pte                    603.70   952.69  1684.12  1480.82   750.86   709.33   482.20   750.89   672.95
  s.d.                           48.61    95.71   364.05     3.49     0.09    20.93     2.48     0.10     0.39
2.6.15.1-sp                     576.38  1052.59  1462.80  1487.41   750.91   710.82   482.95   750.91   672.86
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00

*Local* More Communication bandwidths in MB/s - bigger is better
----------------------------------------------------------------
                                  File     Mmap  Aligned  Partial  Partial  Partial  Partial
OS                                open     open    Bcopy    Bcopy     Mmap     Mmap     Mmap    Bzero
                                 close    close   (libc)   (hand)     read    write   rd/wrt     copy     HTTP
-----------------------------  -------  -------  -------  -------  -------  -------  -------  -------  -------
2.6.15.1                       1484.84   715.71   704.27   710.53   946.41  1563.26   739.75  2255.89   29.974
  s.d.                            1.72     0.45    11.70    23.02     0.07     2.46     0.17     8.51    0.304
2.6.15.1-lvhpt                 1484.12   710.04   716.83   713.90   946.27  1556.90   737.56  2231.20   29.412
  s.d.                            1.30     0.39    11.02    13.16     0.09     3.43     0.08     6.73    0.829
2.6.15.1-pte                   1481.12   708.77   709.41   707.99   946.28  1555.24   737.74  2235.57   29.908
  s.d.                            5.15     0.73    14.74    21.63     0.15     3.48     0.07     5.97    0.128
2.6.15.1-sp                    1486.90   708.53   710.67   706.19   946.48  1565.75   737.77  2233.65   28.810
  s.d.                            0.00     0.00     0.00     0.00     0.00     0.00     0.00     0.00    0.000

Memory latencies in nanoseconds - smaller is better
---------------------------------------------------
kernel                          Mhz     L1 $     L2 $    Main mem
-----------------------------  -----  -------  -------  ---------
2.6.15.1                        1501    1.333    4.002      99.87
  s.d.                             0    0.304    0.304       0.30
2.6.15.1-lvhpt                  1501    1.333    4.003     100.40
  s.d.                             0    0.829    0.829       0.83
2.6.15.1-pte                    1501    1.333    4.003     100.42
  s.d.                             0    0.128    0.128       0.13
2.6.15.1-sp                     1501    1.333    4.002      98.03
  s.d.                             0    0.000    0.000       0.00

kernel compile

with both double pte and superpage.

matrix transform

matrix.png

500 equates to 500 * 500 * 2 * 8 ~= 4MB, 3500 to about 186MB

Test divides this space into two and rotates the matrix from first half into second half. times by 8 because each element is a double

this run on a 1.5ghz itanium 2 with 6mb of cache

therefore (n^2 * 2 * 8) = (6 * 1024 * 1024) means we expect to run out of cache to do this in around 627 elements; which is where we see a large drop

matrix size

sp read

vanilla read

read diff

sp write

vanilla write

write diff

501

1435.21

1372.29

62.92

1374.66

1316.24

58.42

565

1144.32

848.51

295.81

602.56

544.19

58.37

629

942.39

698.75

243.64

768.92

584.79

184.13

693

470.17

404.62

65.55

462.59

350.56

112.03

757

475.03

386.20

88.83

457.78

345.01

112.77

821

465.54

375.22

90.32

458.13

338.19

119.94

885

464.79

370.58

94.21

455.70

329.03

126.67

949

462.82

365.46

97.36

455.66

303.76

151.9

1013

451.94

361.19

90.75

357.39

281.25

76.14

1077

463.06

364.13

98.93

454.04

294.81

159.23

1141

463.89

370.14

93.75

454.07

281.79

172.28

1205

462.72

366.84

95.88

449.36

262.58

186.78

1269

465.43

359.03

106.4

450.38

262.70

187.68

1333

464.61

357.30

107.31

452.36

247.36

205

1397

463.95

352.52

111.43

451.25

245.04

206.21

1461

460.56

345.36

115.2

370.97

240.99

129.98

1525

464.50

342.77

121.73

445.36

235.82

209.54

1589

460.41

340.58

119.83

434.92

236.37

198.55

1653

458.22

335.52

122.7

445.19

231.27

213.92

1717

462.12

327.25

134.87

442.95

230.52

212.43

1781

453.80

321.96

131.84

379.41

230.20

149.21

1845

458.43

320.97

137.46

421.74

229.62

192.12

1909

453.23

315.57

137.66

387.06

228.23

158.83

1973

450.93

309.81

141.12

384.15

223.49

160.66

2037

458.85

307.03

151.82

387.27

229.02

158.25

2101

450.54

300.65

149.89

372.74

218.54

154.2

2165

452.34

300.02

152.32

382.72

220.17

162.55

2229

448.87

300.56

148.31

354.71

220.79

133.92

2293

449.02

299.18

149.84

354.56

220.55

134.01

2357

451.86

299.81

152.05

338.34

220.42

117.92

2421

450.33

302.36

147.97

332.06

220.45

111.61

2485

446.83

300.22

146.61

327.87

220.10

107.77

2549

447.36

300.10

147.26

326.50

218.29

108.21

2613

446.39

299.87

146.52

336.32

220.06

116.26

2677

447.87

299.54

148.33

320.67

220.02

100.65

2741

445.63

299.09

146.54

321.23

219.29

101.94

2805

456.91

299.47

157.44

316.98

220.85

96.13

2869

446.53

300.60

145.93

318.40

218.93

99.47

2933

444.75

298.87

145.88

318.39

220.08

98.31

2997

447.51

297.49

150.02

318.19

218.79

99.4

3061

443.55

297.78

145.77

314.19

217.16

97.03

3125

445.09

298.49

146.6

314.93

218.13

96.8

3189

447.98

300.16

147.82

313.74

219.46

94.28

3253

451.90

297.11

154.79

310.59

219.58

91.01

3317

455.98

298.14

157.84

309.79

219.60

90.19

3381

444.89

295.70

149.19

313.41

219.73

93.68

3445

444.73

297.77

146.96

313.40

219.40

94

superpage stats from run

baci:~# cat /sys/superpage/superpage_stat_faulted
   16K | 58
   64K | 97
  256K | 35
 1024K | 106
 4096K | 187
16384K | 155

looking at pfmon for this case

defining virutal memory stalls as

ianw@baci:~$ cat counters-sp.txt
                    256025 BE_L1D_FPU_BUBBLE_L1D_HPW
                      1620 BE_L1D_FPU_BUBBLE_L1D_TLB
               32587754224 BACK_END_BUBBLE_ALL

ianw@baci:~$ cat counters-vanilla.txt
               15520549328 BE_L1D_FPU_BUBBLE_L1D_HPW
                      1737 BE_L1D_FPU_BUBBLE_L1D_TLB
               41027595689 BACK_END_BUBBLE_ALL

caliper also helps us out here

caliper runs

vanilla run

Data TLB Metrics Summed for Entire Run
-----------------------------------------------
                         PLM
Event Name              U..K  TH          Count
-----------------------------------------------
DATA_REFERENCES         x___   0     2814721300
DTLB_INSERTS_HPW        x___   0      744197664
L2DTLB_MISSES           x___   0      745547808
-----------------------------------------------
PLM: Privilege Level Mask
     U/K = user/kernel levels (U: level 3, K: level 0)
     The intermediate levels (1, 2) are unused on HP-UX or Linux
     x : the metric is measured at the given level (_ : not measured)
TH: event THreshold, determines the event counter behavior,
    TH == 0 : counter += event_count_in_cycle
    TH  > 0 : counter += (event_count_in_cycle >= threshold ? 1 : 0)
-----------------------------------------------
Total L1 data TLB references:
  2814721300 = DATA_REFERENCES

Percentage of data references covered by L1 and L2 DTLB:
  73.51 % = 100 * (1 - L2DTLB_MISSES / DATA_REFERENCES)

Percentage of data references covered by the HPW:
  26.44 % = 100 * (DTLB_INSERTS_HPW / DATA_REFERENCES)

Percentage of data references covered by software trap:
   0.05 % = 100 * ((L2DTLB_MISSES - DTLB_INSERTS_HPW) / DATA_REFERENCES)

Percentage of L2 DTLB misses covered by the HPW:
  99.82 % = 100 * (DTLB_INSERTS_HPW / L2DTLB_MISSES)
-----------------------------------------------

superpage run

Data TLB Metrics Summed for Entire Run
-----------------------------------------------
                         PLM
Event Name              U..K  TH          Count
-----------------------------------------------
DATA_REFERENCES         x___   0     2813021018
DTLB_INSERTS_HPW        x___   0            397
L2DTLB_MISSES           x___   0           1729
-----------------------------------------------
PLM: Privilege Level Mask
     U/K = user/kernel levels (U: level 3, K: level 0)
     The intermediate levels (1, 2) are unused on HP-UX or Linux
     x : the metric is measured at the given level (_ : not measured)
TH: event THreshold, determines the event counter behavior,
    TH == 0 : counter += event_count_in_cycle
    TH  > 0 : counter += (event_count_in_cycle >= threshold ? 1 : 0)
-----------------------------------------------
Total L1 data TLB references:
  2813021018 = DATA_REFERENCES

Percentage of data references covered by L1 and L2 DTLB:
  100.00 % = 100 * (1 - L2DTLB_MISSES / DATA_REFERENCES)

Percentage of data references covered by the HPW:
   0.00 % = 100 * (DTLB_INSERTS_HPW / DATA_REFERENCES)

Percentage of data references covered by software trap:
   0.00 % = 100 * ((L2DTLB_MISSES - DTLB_INSERTS_HPW) / DATA_REFERENCES)

Percentage of L2 DTLB misses covered by the HPW:
  22.96 % = 100 * (DTLB_INSERTS_HPW / L2DTLB_MISSES)
-----------------------------------------------

SPECCPU 2000

this run isn't even close to be valid

          ========= VANILLA KERNEL ============ ========= SUPERPAGE KERNEL ==========
                       SPEC CINT2000 Summary   |        SPEC CINT2000 Summary
                    Fri Feb 24 13:16:01 2006   |         Fri Feb 24 14:18:16 2006
                                               |

                   Base      Base      Base    |                     Base      Base      Base    
   Benchmarks    Ref Time  Run Time   Ratio    |     Benchmarks    Ref Time  Run Time   Ratio    
   ------------  --------  --------  --------  |     ------------  --------  --------  -------- 
   164.gzip          1400   206       679    * |     164.gzip          1400   206       679    *
   175.vpr           1400   167       837    * |     175.vpr           1400   164       855    *
   176.gcc           1100        --          X |     176.gcc           1100        --          X
   181.mcf           1800   262       688    * |     181.mcf           1800   232       776    *
   186.crafty        1000   115       871    * |     186.crafty        1000   115       870    *
   197.parser        1800   249       724    * |     197.parser        1800   245       734    *
   252.eon           1300        --          X |     252.eon           1300        --          X
   253.perlbmk       1800   208       865    * |     253.perlbmk       1800   208              X
   254.gap           1100   199       553    * |     254.gap           1100   198       556    *
   255.vortex        1900   193              X |     255.vortex        1900   194              X
   256.bzip2         1500   211       712    * |     256.bzip2         1500   209       719    *
   300.twolf         3000   311       965    * |     300.twolf         3000   311       966    *
   =========================================== |     ===========================================
   164.gzip          1400   206       679    * |     164.gzip          1400   206       679    *
   175.vpr           1400   167       837    * |     175.vpr           1400   164       855    *
   176.gcc                                   X |     176.gcc                                   X
   181.mcf           1800   262       688    * |     181.mcf           1800   232       776    *
   186.crafty        1000   115       871    * |     186.crafty        1000   115       870    *
   197.parser        1800   249       724    * |     197.parser        1800   245       734    *
   252.eon                                   X |     252.eon                                   X
   253.perlbmk       1800   208       865    * |     253.perlbmk                               X
   254.gap           1100   199       553    * |     254.gap           1100   198       556    *
   255.vortex                                X |     255.vortex                                X
   256.bzip2         1500   211       712    * |     256.bzip2         1500   209       719    *
   300.twolf         3000   311       965    * |     300.twolf         3000   311       966    *

IA64wiki: Ia64SuperPages/BenchMarkingIdeas (last edited 2009-12-10 03:13:45 by localhost)

Gelato@UNSW is sponsored by
the University of New South Wales National ICT Australia The Gelato Federation Hewlett-Packard Company Australian Research Council
Please contact us with any questions or comments.