On Thu, Apr 7, 2016 at 6:48 PM, Andres Freund <and...@anarazel.de> wrote:
> On 2016-04-07 18:40:14 +0530, Amit Kapila wrote: > > > This is the data with -b tpcb-like@1 with 20-min run for each version > and I > > could see almost similar results as the data posted in previous e-mail. > > > > Client Count/Patch_ver (tps) 256 > > clog_buf_128 40617 > > clog_buf_128 +group_clog_v8 51137 > > clog_buf_128 +content_lock 54188 > > > > For -b select-only@3, I have done quicktest for each version and > number is > > same 62K~63K for all version, why do you think this will improve > > select-only workload? > > What I was looking for was pgbench with both -btpcb-like@1 > -bselect-only@3 specified; i.e. a mixed read/write test. I have taken the data in the suggested way and the performance seems to be neutral for both the patches. Detailed data for all the runs for three versions is attached. Median of 3 20-minutes run. Client Count/Patch_ver (tps) 256 clog_buf_128 110630 clog_buf_128 +group_clog_v8 111575 clog_buf_128 +content_lock 96581 Now, from above data, it appears that content lock patch has some regression, but if you see in detailed data attached with this mail, the highest TPS is close to other patches, but still on the lesser side. > In my > measurement that's where Simon's approach shines (not surprising if you > look at the way it works), and it's of immense practical importance - > most workloads are mixed. > > I have tried above tests two times, but didn't notice any gain with content lock approach. I think by now, we have done many tests with both approaches and we find that in some cases, it is slightly better and in most cases it is neutral and in some cases it is worse than group clog approach. I feel we should go with group clog approach now as that has been tested and reviewed multiple times and in future if we find that other approach is giving substantial gain, then we can anyway change it. With Regards, Amit Kapila. EnterpriseDB: http://www.enterprisedb.com
Start of script for 300 8GB ============== group_clog_v8 ============= ================================================ 300, 8GB, 256, 256, 1200 Reading - 1 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 135434825 latency average: 2.268 ms tps = 112860.743738 (including connections establishing) tps = 112861.829048 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 33376619 transactions (24.6% of total, tps = 27813.452292) - latency average = 6.866 ms - latency stddev = 4.977 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 98885323 transactions (73.0% of total, tps = 82403.260007) - latency average = 0.739 ms - latency stddev = 1.681 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 2 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 133946949 latency average: 2.293 ms tps = 111573.984461 (including connections establishing) tps = 111575.116415 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 32930394 transactions (24.6% of total, tps = 27430.078071) - latency average = 6.903 ms - latency stddev = 5.022 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 98074548 transactions (73.2% of total, tps = 81693.298551) - latency average = 0.759 ms - latency stddev = 1.711 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 3 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 123963990 latency average: 2.478 ms tps = 103301.868530 (including connections establishing) tps = 103302.907780 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 30533935 transactions (24.6% of total, tps = 25444.587086) - latency average = 7.218 ms - latency stddev = 5.409 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 90442530 transactions (73.0% of total, tps = 75367.712378) - latency average = 0.901 ms - latency stddev = 2.139 ms CHECKPOINT
Start of script for 300 8GB ============== clog_bufs_128 ============= ================================================ 300, 8GB, 256, 256, 1200 Reading - 1 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 133674543 latency average: 2.298 ms tps = 111374.553715 (including connections establishing) tps = 111376.035237 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 32932644 transactions (24.6% of total, tps = 27438.721284) - latency average = 7.006 ms - latency stddev = 4.972 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 97971096 transactions (73.3% of total, tps = 81627.263120) - latency average = 0.727 ms - latency stddev = 1.572 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 2 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 132757684 latency average: 2.314 ms tps = 110629.233802 (including connections establishing) tps = 110630.216250 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 32723094 transactions (24.6% of total, tps = 27268.710238) - latency average = 7.007 ms - latency stddev = 5.006 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 97352576 transactions (73.3% of total, tps = 81125.555727) - latency average = 0.754 ms - latency stddev = 1.684 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 3 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 128866667 latency average: 2.384 ms tps = 107372.372880 (including connections establishing) tps = 107373.747375 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 31780534 transactions (24.7% of total, tps = 26479.705159) - latency average = 7.110 ms - latency stddev = 5.054 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 94503526 transactions (73.3% of total, tps = 78740.826223) - latency average = 0.809 ms - latency stddev = 1.742 ms CHECKPOINT
Start of script for 300 8GB ============== content_lock_v1 ============= ================================================ 300, 8GB, 256, 256, 1200 Reading - 1 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 115898576 latency average: 2.651 ms tps = 96580.732081 (including connections establishing) tps = 96581.841866 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 28580217 transactions (24.7% of total, tps = 23816.498668) - latency average = 7.425 ms - latency stddev = 5.537 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 85000250 transactions (73.3% of total, tps = 70832.504207) - latency average = 1.062 ms - latency stddev = 2.361 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 2 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 127856329 latency average: 2.403 ms tps = 106543.872725 (including connections establishing) tps = 106545.272892 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 31526708 transactions (24.7% of total, tps = 26271.500135) - latency average = 6.975 ms - latency stddev = 5.150 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 93449930 transactions (73.1% of total, tps = 77872.699193) - latency average = 0.883 ms - latency stddev = 1.961 ms CHECKPOINT ================================================ 300, 8GB, 256, 256, 1200 Reading - 3 transaction type: multiple scripts scaling factor: 300 query mode: prepared number of clients: 256 number of threads: 256 duration: 1200 s number of transactions actually processed: 115599417 latency average: 2.657 ms tps = 96322.521966 (including connections establishing) tps = 96323.716924 (excluding connections establishing) SQL script 1: <builtin: TPC-B (sort of)> - weight = 1 (targets 25.0% of total) - 28480781 transactions (24.6% of total, tps = 23731.440216) - latency average = 7.392 ms - latency stddev = 5.801 ms SQL script 2: <builtin: select only> - weight = 3 (targets 75.0% of total) - 84703628 transactions (73.3% of total, tps = 70578.790797) - latency average = 1.084 ms - latency stddev = 2.561 ms CHECKPOINT
-- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers