Hi Thomas
I did some light testing merging 2.6.33.13 into real-time 2.6.33.9-rt31.
In addition I cherry-picked 3c955b407a084810f57260d61548cc92c14bc627
in order to compile on newer distros.
Please consider picking this up. You can get it from
git://git.kernel.org/pub/scm/linux/kernel/git/jkacur/jk-2.6.git rt/2.6.33
Thanks
John Kacur
Here is the result of cyclic test on one machine
sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
policy: fifo: loadavg: 0.00 0.00 0.00 1/541 3759
T: 0 ( 3728) P:80 I:10000 C: 10000 Min: 7 Act: 104 Avg: 114 Max: 470
T: 1 ( 3729) P:79 I:10500 C: 9524 Min: 4 Act: 93 Avg: 105 Max: 417
T: 2 ( 3730) P:78 I:11000 C: 9091 Min: 4 Act: 114 Avg: 125 Max: 420
T: 3 ( 3731) P:77 I:11500 C: 8696 Min: 4 Act: 58 Avg: 91 Max: 387
T: 4 ( 3732) P:76 I:12000 C: 8333 Min: 4 Act: 84 Avg: 114 Max: 425
T: 5 ( 3733) P:75 I:12500 C: 8000 Min: 4 Act: 133 Avg: 112 Max: 385
T: 6 ( 3734) P:74 I:13000 C: 7692 Min: 3 Act: 8 Avg: 65 Max: 377
T: 7 ( 3735) P:73 I:13500 C: 7407 Min: 3 Act: 91 Avg: 71 Max: 358
T: 8 ( 3736) P:72 I:14000 C: 7143 Min: 4 Act: 78 Avg: 98 Max: 323
T: 9 ( 3737) P:71 I:14500 C: 6897 Min: 7 Act: 133 Avg: 109 Max: 300
T:10 ( 3738) P:70 I:15000 C: 6667 Min: 5 Act: 115 Avg: 98 Max: 314
T:11 ( 3739) P:69 I:15500 C: 6452 Min: 3 Act: 6 Avg: 79 Max: 280
T:12 ( 3740) P:68 I:16000 C: 6250 Min: 3 Act: 68 Avg: 80 Max: 317
T:13 ( 3741) P:67 I:16500 C: 6061 Min: 4 Act: 142 Avg: 82 Max: 255
T:14 ( 3742) P:66 I:17000 C: 5882 Min: 4 Act: 132 Avg: 102 Max: 253
T:15 ( 3743) P:65 I:17500 C: 5714 Min: 4 Act: 135 Avg: 111 Max: 342
T:16 ( 3744) P:64 I:18000 C: 5556 Min: 4 Act: 16 Avg: 37 Max: 315
T:17 ( 3745) P:63 I:18500 C: 5405 Min: 3 Act: 12 Avg: 53 Max: 322
T:18 ( 3746) P:62 I:19000 C: 5263 Min: 3 Act: 62 Avg: 37 Max: 250
T:19 ( 3747) P:61 I:19500 C: 5128 Min: 3 Act: 65 Avg: 35 Max: 276
T:20 ( 3748) P:60 I:20000 C: 5000 Min: 3 Act: 180 Avg: 81 Max: 295
T:21 ( 3749) P:59 I:20500 C: 4878 Min: 3 Act: 76 Avg: 45 Max: 314
T:22 ( 3750) P:58 I:21000 C: 4762 Min: 3 Act: 133 Avg: 38 Max: 265
T:23 ( 3751) P:57 I:21500 C: 4651 Min: 3 Act: 132 Avg: 71 Max: 300
T:24 ( 3752) P:56 I:22000 C: 4546 Min: 3 Act: 6 Avg: 57 Max: 227
T:25 ( 3753) P:55 I:22500 C: 4445 Min: 3 Act: 72 Avg: 50 Max: 258
T:26 ( 3754) P:54 I:23000 C: 4348 Min: 4 Act: 97 Avg: 106 Max: 335
T:27 ( 3755) P:53 I:23500 C: 4255 Min: 3 Act: 67 Avg: 80 Max: 244
T:28 ( 3756) P:52 I:24000 C: 4167 Min: 3 Act: 7 Avg: 26 Max: 232
T:29 ( 3757) P:51 I:24500 C: 4082 Min: 3 Act: 6 Avg: 81 Max: 353
T:30 ( 3758) P:50 I:25000 C: 4000 Min: 3 Act: 11 Avg: 31 Max: 418
T:31 ( 3759) P:49 I:25500 C: 3922 Min: 3 Act: 6 Avg: 21 Max: 433
On Sat, 14 May 2011, John Kacur wrote:
> I did some light testing merging 2.6.33.13 into real-time 2.6.33.9-rt31.
> In addition I cherry-picked 3c955b407a084810f57260d61548cc92c14bc627
> in order to compile on newer distros.
Thanks.
> Here is the result of cyclic test on one machine
> sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
> policy: fifo: loadavg: 0.00 0.00 0.00 1/541 3759
>
> T: 0 ( 3728) P:80 I:10000 C: 10000 Min: 7 Act: 104 Avg: 114 Max: 470
The numbers are weird. How does that compare to older kernels on that
machine with the same test?
Thanks,
tglx
On Sat, May 14, 2011 at 6:28 PM, Thomas Gleixner <[email protected]> wrote:
> On Sat, 14 May 2011, John Kacur wrote:
>> I did some light testing merging 2.6.33.13 into real-time 2.6.33.9-rt31.
>> In addition I cherry-picked 3c955b407a084810f57260d61548cc92c14bc627
>> in order to compile on newer distros.
>
> Thanks.
>
>> Here is the result of cyclic test on one machine
>> sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
>> policy: fifo: loadavg: 0.00 0.00 0.00 1/541 3759
>>
>> T: 0 ( 3728) P:80 I:10000 C: ?10000 Min: ? ? ?7 Act: ?104 Avg: ?114 Max: ? ? 470
>
> The numbers are weird. How does that compare to older kernels on that
> machine with the same test?
Note this was on a machine with a straight Fedora install, and no -rt
packages or tuning. That being the case, the numbers are no better or
worse than recent -rt kernels. I can get you numbers with the last
kernel if you wish, on Monday. Perhaps Carsten would be interested in
running his tests?
John
John,
> On Sat, May 14, 2011 at 6:28 PM, Thomas Gleixner<[email protected]> wrote:
>> On Sat, 14 May 2011, John Kacur wrote:
>>> I did some light testing merging 2.6.33.13 into real-time 2.6.33.9-rt31.
>>> In addition I cherry-picked 3c955b407a084810f57260d61548cc92c14bc627
>>> in order to compile on newer distros.
>>
>> Thanks.
>>
>>> Here is the result of cyclic test on one machine
>>> sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
>>> policy: fifo: loadavg: 0.00 0.00 0.00 1/541 3759
>>>
>>> T: 0 ( 3728) P:80 I:10000 C: 10000 Min: 7 Act: 104 Avg: 114 Max: 470
>>
>> The numbers are weird. How does that compare to older kernels on that
>> machine with the same test?
>
> Note this was on a machine with a straight Fedora install, and no -rt
> packages or tuning. That being the case, the numbers are no better or
> worse than recent -rt kernels. I can get you numbers with the last
> kernel if you wish, on Monday. Perhaps Carsten would be interested in
> running his tests?
I created my own 2.6.33.13-rt31 kernel some days ago. It is running on a
pretest machine in our farm (rack #1, slot #4) since then ->
https://www.osadl.org/?id=994. For an overview about the kernels in the
farm (ordered by kernel version and release) see ->
https://www.osadl.org/?id=1001. Real-time and all other data before and
after the kernel upgrade to 2.6.33.13-rt31 are indistinguishable between
each other.
Nevertheless, it would be interesting to investigate the origin of the
numbers Thomas (correctly IMHO) classified as weird. The average and the
maximum latency could be from a Pentium 90, the minimum latency looks ok
for a state-of-the-art processor. What machine is this and how did you
manage to create such numbers? It doesn't look like SMIs were
responsible for that since they normally increase the maximum, not the
average. Was another task with prio 80 or higher running?
Thanks,
Carsten.
On Sat, May 14, 2011 at 5:59 PM, John Kacur <[email protected]> wrote:
>
> Hi Thomas
>
> I did some light testing merging 2.6.33.13 into real-time 2.6.33.9-rt31.
> In addition I cherry-picked 3c955b407a084810f57260d61548cc92c14bc627
> in order to compile on newer distros.
>
> Please consider picking this up. You can get it from
> git://git.kernel.org/pub/scm/linux/kernel/git/jkacur/jk-2.6.git rt/2.6.33
>
> Thanks
>
> John Kacur
>
> Here is the result of cyclic test on one machine
> sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
> policy: fifo: loadavg: 0.00 0.00 0.00 1/541 3759
>
> T: 0 ( 3728) P:80 I:10000 C: ?10000 Min: ? ? ?7 Act: ?104 Avg: ?114 Max: ? ? 470
> T: 1 ( 3729) P:79 I:10500 C: ? 9524 Min: ? ? ?4 Act: ? 93 Avg: ?105 Max: ? ? 417
> T: 2 ( 3730) P:78 I:11000 C: ? 9091 Min: ? ? ?4 Act: ?114 Avg: ?125 Max: ? ? 420
> T: 3 ( 3731) P:77 I:11500 C: ? 8696 Min: ? ? ?4 Act: ? 58 Avg: ? 91 Max: ? ? 387
> T: 4 ( 3732) P:76 I:12000 C: ? 8333 Min: ? ? ?4 Act: ? 84 Avg: ?114 Max: ? ? 425
> T: 5 ( 3733) P:75 I:12500 C: ? 8000 Min: ? ? ?4 Act: ?133 Avg: ?112 Max: ? ? 385
> T: 6 ( 3734) P:74 I:13000 C: ? 7692 Min: ? ? ?3 Act: ? ?8 Avg: ? 65 Max: ? ? 377
> T: 7 ( 3735) P:73 I:13500 C: ? 7407 Min: ? ? ?3 Act: ? 91 Avg: ? 71 Max: ? ? 358
> T: 8 ( 3736) P:72 I:14000 C: ? 7143 Min: ? ? ?4 Act: ? 78 Avg: ? 98 Max: ? ? 323
> T: 9 ( 3737) P:71 I:14500 C: ? 6897 Min: ? ? ?7 Act: ?133 Avg: ?109 Max: ? ? 300
> T:10 ( 3738) P:70 I:15000 C: ? 6667 Min: ? ? ?5 Act: ?115 Avg: ? 98 Max: ? ? 314
> T:11 ( 3739) P:69 I:15500 C: ? 6452 Min: ? ? ?3 Act: ? ?6 Avg: ? 79 Max: ? ? 280
> T:12 ( 3740) P:68 I:16000 C: ? 6250 Min: ? ? ?3 Act: ? 68 Avg: ? 80 Max: ? ? 317
> T:13 ( 3741) P:67 I:16500 C: ? 6061 Min: ? ? ?4 Act: ?142 Avg: ? 82 Max: ? ? 255
> T:14 ( 3742) P:66 I:17000 C: ? 5882 Min: ? ? ?4 Act: ?132 Avg: ?102 Max: ? ? 253
> T:15 ( 3743) P:65 I:17500 C: ? 5714 Min: ? ? ?4 Act: ?135 Avg: ?111 Max: ? ? 342
> T:16 ( 3744) P:64 I:18000 C: ? 5556 Min: ? ? ?4 Act: ? 16 Avg: ? 37 Max: ? ? 315
> T:17 ( 3745) P:63 I:18500 C: ? 5405 Min: ? ? ?3 Act: ? 12 Avg: ? 53 Max: ? ? 322
> T:18 ( 3746) P:62 I:19000 C: ? 5263 Min: ? ? ?3 Act: ? 62 Avg: ? 37 Max: ? ? 250
> T:19 ( 3747) P:61 I:19500 C: ? 5128 Min: ? ? ?3 Act: ? 65 Avg: ? 35 Max: ? ? 276
> T:20 ( 3748) P:60 I:20000 C: ? 5000 Min: ? ? ?3 Act: ?180 Avg: ? 81 Max: ? ? 295
> T:21 ( 3749) P:59 I:20500 C: ? 4878 Min: ? ? ?3 Act: ? 76 Avg: ? 45 Max: ? ? 314
> T:22 ( 3750) P:58 I:21000 C: ? 4762 Min: ? ? ?3 Act: ?133 Avg: ? 38 Max: ? ? 265
> T:23 ( 3751) P:57 I:21500 C: ? 4651 Min: ? ? ?3 Act: ?132 Avg: ? 71 Max: ? ? 300
> T:24 ( 3752) P:56 I:22000 C: ? 4546 Min: ? ? ?3 Act: ? ?6 Avg: ? 57 Max: ? ? 227
> T:25 ( 3753) P:55 I:22500 C: ? 4445 Min: ? ? ?3 Act: ? 72 Avg: ? 50 Max: ? ? 258
> T:26 ( 3754) P:54 I:23000 C: ? 4348 Min: ? ? ?4 Act: ? 97 Avg: ?106 Max: ? ? 335
> T:27 ( 3755) P:53 I:23500 C: ? 4255 Min: ? ? ?3 Act: ? 67 Avg: ? 80 Max: ? ? 244
> T:28 ( 3756) P:52 I:24000 C: ? 4167 Min: ? ? ?3 Act: ? ?7 Avg: ? 26 Max: ? ? 232
> T:29 ( 3757) P:51 I:24500 C: ? 4082 Min: ? ? ?3 Act: ? ?6 Avg: ? 81 Max: ? ? 353
> T:30 ( 3758) P:50 I:25000 C: ? 4000 Min: ? ? ?3 Act: ? 11 Avg: ? 31 Max: ? ? 418
> T:31 ( 3759) P:49 I:25500 C: ? 3922 Min: ? ? ?3 Act: ? ?6 Avg: ? 21 Max: ? ? 433
Yeah, the numbers are a bit weird. I unfortunately choose odd hardware
to report numbers on, since it is known to have problems with -rt.
It was more a correctness test than an -rt performance test.
However, here are the numbers with 2.6.33.9-rt31 to compare and contrast.
uname -r
2.6.33.9-rt31-00001-g76582bb
sudo ./cyclictest -t32 -p 80 -n -i 10000 -l 10000
policy: fifo: loadavg: 0.10 0.31 0.15 1/543 3471
T: 0 ( 3440) P:80 I:10000 C: 5442 Min: 4 Act: 88 Avg: 138 Max: 486
T: 1 ( 3441) P:79 I:10500 C: 5183 Min: 6 Act: 70 Avg: 133 Max: 462
T: 2 ( 3442) P:78 I:11000 C: 4947 Min: 4 Act: 86 Avg: 127 Max: 399
T: 3 ( 3443) P:77 I:11500 C: 4732 Min: 5 Act: 120 Avg: 105 Max: 368
T: 4 ( 3444) P:76 I:12000 C: 4535 Min: 5 Act: 90 Avg: 126 Max: 424
T: 5 ( 3445) P:75 I:12500 C: 4354 Min: 4 Act: 89 Avg: 113 Max: 419
T: 6 ( 3446) P:74 I:13000 C: 4186 Min: 4 Act: 14 Avg: 89 Max: 428
T: 7 ( 3447) P:73 I:13500 C: 4031 Min: 3 Act: 9 Avg: 55 Max: 402
T: 8 ( 3448) P:72 I:14000 C: 3887 Min: 3 Act: 84 Avg: 65 Max: 375
T: 9 ( 3449) P:71 I:14500 C: 3753 Min: 4 Act: 68 Avg: 101 Max: 271
T:10 ( 3450) P:70 I:15000 C: 3628 Min: 3 Act: 130 Avg: 82 Max: 334
T:11 ( 3451) P:69 I:15500 C: 3511 Min: 3 Act: 110 Avg: 79 Max: 310
T:12 ( 3452) P:68 I:16000 C: 3401 Min: 3 Act: 9 Avg: 71 Max: 280
T:13 ( 3453) P:67 I:16500 C: 3298 Min: 3 Act: 7 Avg: 70 Max: 255
T:14 ( 3454) P:66 I:17000 C: 3201 Min: 5 Act: 112 Avg: 97 Max: 296
T:15 ( 3455) P:65 I:17500 C: 3110 Min: 4 Act: 69 Avg: 92 Max: 312
T:16 ( 3456) P:64 I:18000 C: 3023 Min: 4 Act: 138 Avg: 94 Max: 311
T:17 ( 3457) P:63 I:18500 C: 2942 Min: 4 Act: 86 Avg: 108 Max: 289
T:18 ( 3458) P:62 I:19000 C: 2864 Min: 3 Act: 15 Avg: 31 Max: 335
T:19 ( 3459) P:61 I:19500 C: 2791 Min: 4 Act: 71 Avg: 89 Max: 274
T:20 ( 3460) P:60 I:20000 C: 2721 Min: 4 Act: 8 Avg: 74 Max: 287
T:21 ( 3461) P:59 I:20500 C: 2655 Min: 3 Act: 68 Avg: 62 Max: 339
T:22 ( 3462) P:58 I:21000 C: 10000 Min: 3 Act: 186 Avg: 94 Max: 407
T:23 ( 3463) P:57 I:21500 C: 10000 Min: 3 Act: 93 Avg: 64 Max: 373
T:24 ( 3464) P:56 I:22000 C: 10000 Min: 4 Act: 132 Avg: 82 Max: 356
T:25 ( 3465) P:55 I:22500 C: 10000 Min: 3 Act: 133 Avg: 93 Max: 323
T:26 ( 3466) P:54 I:23000 C: 10000 Min: 3 Act: 97 Avg: 64 Max: 333
T:27 ( 3467) P:53 I:23500 C: 10000 Min: 3 Act: 134 Avg: 70 Max: 320
T:28 ( 3468) P:52 I:24000 C: 10000 Min: 3 Act: 123 Avg: 59 Max: 226
T:29 ( 3469) P:51 I:24500 C: 10000 Min: 3 Act: 143 Avg: 73 Max: 241
T:30 ( 3470) P:50 I:25000 C: 10000 Min: 3 Act: 15 Avg: 82 Max: 241
T:31 ( 3471) P:49 I:25500 C: 10000 Min: 3 Act: 10 Avg: 49 Max: 294