Superpage Benchmarking Ideas
Contents
This is a junk page for keeping ideas and notes about superpage benchmarking.
double size pte
- lmbench; does it add overheads?
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
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_ALLianw@baci:~$ cat counters-vanilla.txt
15520549328 BE_L1D_FPU_BUBBLE_L1D_HPW
1737 BE_L1D_FPU_BUBBLE_L1D_TLB
41027595689 BACK_END_BUBBLE_ALLcaliper 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 *
