On a quad xeon with 3.75 gb ram there was a big
regression in dbench and tiobench with many threads
between 2.4.20-pre1 and 2.4.20-pre2. Below is from
5 dbench runs.
There was a 23% regression at 192 processes
reiserfs
dbench 192 processes Average High Low
2.4.20-pre1 55.94 58.17 54.26
2.4.20-pre2 42.98 44.73 42.30
And 13% regression at 64 processes.
dbench 64 processes Average High Low
2.4.20-pre1 70.98 72.53 69.47
2.4.20-pre2 61.93 64.01 57.31
ext2 and ext3 did not change between 2.4.20-pre1 and pre2.
ext3
dbench 192 processes Average High Low
2.4.20-pre1 60.50 61.47 59.01
2.4.20-pre2 60.61 61.63 58.80
dbench 64 processes Average High Low
2.4.20-pre1 82.73 89.63 64.67
2.4.20-pre2 83.12 91.28 65.11
ext2
dbench 192 processes Average High Low
2.4.20-pre1 116.32 124.44 110.95
2.4.20-pre2 116.45 119.89 109.89
dbench 64 processes Average High Low
2.4.20-pre1 145.61 158.67 102.33
2.4.20-pre2 145.77 158.96 103.89
CPU utilization on reiserfs went down in the high thread count
tiobench tests, but it also had a throughput regression when there was
more than one thread most tests. ext2 and ext3 didn't have a regression.
tiobench-0.3.3
Unit information
================
File size = 12288 megabytes
Blk Size = 4096 bytes
Rate = megabytes per second
CPU% = percentage of CPU used during the test
Latency = milliseconds
Lat% = percent of requests that took longer than X seconds
CPU Eff = Rate divided by CPU% - throughput per cpu load
Sequential Reads reiserfs had the biggest drop in throughput.
Num Avg Maximum Lat% Lat% CPU
Kernel Thr Rate (CPU%) Latency Latency >2s >10s Eff
------------- --- ----------------------------------------------------------
2.4.20-pre1 1 35.11 23.02% 0.331 107.07 0.00000 0.00000 152
2.4.20-pre2 1 34.68 22.83% 0.335 119.57 0.00000 0.00000 152
2.4.20-pre1 32 27.27 20.12% 11.536 258436.68 0.01275 0.01135 136
2.4.20-pre2 32 8.00 6.05% 43.881 7568.44 0.00096 0.00000 132
2.4.20-pre1 64 26.50 19.63% 21.328 361756.08 0.02183 0.02053 135
2.4.20-pre2 64 8.42 6.46% 79.039 15721.84 0.18813 0.00000 130
2.4.20-pre1 128 27.72 21.17% 35.485 814716.92 0.02671 0.02549 131
2.4.20-pre2 128 9.30 7.10% 135.056 32520.09 3.92532 0.00035 131
2.4.20-pre1 256 26.55 19.95% 68.425 825224.03 0.04636 0.04515 133
2.4.20-pre2 256 8.47 6.49% 285.430 62013.50 4.68931 0.11676 131
Random Reads reiserfs went down when there is more than 1 thread.
Num Avg Maximum Lat% Lat% CPU
Kernel Thr Rate (CPU%) Latency Latency >2s >10s Eff
------------- --- ----------------------------------------------------------
2.4.20-pre1 1 0.68 0.84% 17.320 51.54 0.00000 0.00000 81
2.4.20-pre2 1 0.70 0.76% 16.787 53.99 0.00000 0.00000 92
2.4.20-pre1 32 3.09 6.80% 102.783 535.46 0.00000 0.00000 46
2.4.20-pre2 32 1.98 3.84% 147.780 16015.44 0.55000 0.00000 52
2.4.20-pre1 64 3.10 5.99% 178.641 10722.38 0.50403 0.00000 52
2.4.20-pre2 64 2.18 3.70% 229.511 16850.53 2.77218 0.00000 59
2.4.20-pre1 128 3.35 5.27% 226.876 9018.16 1.91532 0.00000 64
2.4.20-pre2 128 2.31 3.98% 378.511 17932.69 5.77117 0.00000 58
2.4.20-pre1 256 3.26 6.44% 249.264 5751.33 0.57292 0.00000 51
2.4.20-pre2 256 2.20 5.53% 724.340 19268.40 13.09896 0.00000 40
Sequential Writes reiserfs cpu utilization went down, but throughput dropped.
Num Avg Maximum Lat% Lat% CPU
Kernel Thr Rate (CPU%) Latency Latency >2s >10s Eff
------------- --- ----------------------------------------------------------
2.4.20-pre1 1 39.31 71.71% 0.282 6301.59 0.00012 0.00000 55
2.4.20-pre2 1 38.28 73.39% 0.290 8058.03 0.00019 0.00000 52
2.4.20-pre1 32 29.64 119.5% 11.171 16619.06 0.18094 0.00000 25
2.4.20-pre2 32 23.52 103.0% 10.242 17721.55 0.13629 0.00000 23
2.4.20-pre1 64 29.02 115.6% 22.065 38792.62 0.38675 0.00426 25
2.4.20-pre2 64 20.92 84.06% 19.449 41273.31 0.33824 0.00480 25
2.4.20-pre1 128 31.52 111.6% 36.264 104743.20 0.40728 0.11428 28
2.4.20-pre2 128 22.31 87.18% 36.251 78916.66 0.46791 0.10154 26
2.4.20-pre1 256 28.95 132.7% 75.959 163159.61 0.79301 0.24067 22
2.4.20-pre2 256 20.08 93.50% 74.113 114883.50 0.83399 0.23864 21
Random Writes reiserfs - slightly improved.
Num Avg Maximum Lat% Lat% CPU
Kernel Thr Rate (CPU%) Latency Latency >2s >10s Eff
------------- --- ----------------------------------------------------------
2.4.20-pre1 1 3.57 3.80% 0.091 0.57 0.00000 0.00000 94
2.4.20-pre2 1 3.63 3.79% 0.089 0.68 0.00000 0.00000 96
2.4.20-pre1 32 3.52 9.15% 0.654 268.43 0.00000 0.00000 38
2.4.20-pre2 32 3.57 8.15% 0.377 258.42 0.00000 0.00000 44
2.4.20-pre1 64 3.68 9.50% 0.567 196.77 0.00000 0.00000 39
2.4.20-pre2 64 3.78 9.76% 0.450 319.24 0.00000 0.00000 39
2.4.20-pre1 128 3.48 8.38% 2.923 626.60 0.00000 0.00000 42
2.4.20-pre2 128 3.54 8.98% 0.669 735.87 0.00000 0.00000 39
2.4.20-pre1 256 3.45 9.11% 3.416 572.87 0.00000 0.00000 38
2.4.20-pre2 256 3.59 10.12% 1.350 824.15 0.00000 0.00000 35
Bonnie++ did not have these regressions, but I run that as a single thread.
More on the tests at:
http://home.earthlink.net/~rwhron/kernel/bigbox.html
--
Randy Hron
