Performance testing 1C

Want to measure the performance 1С82 in several configurations


+ all the same in virtualcad KVM, XEN, VirtualBox

There is either a test of Gileva
or take the working database(10Gb), to pureprofile all documents for the year 2012, Zemarite time.

Any other ideas how to test?
2 answers
October 3rd 19 at 02:44
Test Gileva pure synthetic is not always adequate and does not measure concurrency. Same with the reposting will be downloaded rphost 1 and respectively one core.

When we conducted tests, wrote a special treatment, which performs standard operations of our (creating documents, generating reports, changing properties in the references and so on) for a few thousand operations elements and measured the execution time. Plus to run several computers (the number of processes rphost on the server 1C or more).

Plus we must consider that different configurations require different — some need fast disk subsystem (which trade at a constant sampling of reference), the other main bottleneck processor (the accounting is the same).

Also recommend to test different configurations with the network and the disk subsystem, for example a) on the same DB server+server 1C+storage, b) all on different servers), the server 1S + DB server+storage, and so on.

PS: THE 1C is the fun, Hyper trading can be several times the results change when switching on and off, consider this point.
PS2: If the results are not under NDA, I would like to see them.
The results will definitely be published — I don't see private information.
Iron — I have no power testing is expected in i5(3.4 Ghz), 8Gb RAM. Small company — small server.
Post the DB and сервер1С also unlikely other server, even cut the V-world from anyone )))
To think anything is possible, but the results are not impressive. But I'll try anyway, thanks - Neil commented on October 3rd 19 at 02:47
Multi-threaded performance testing of 1C server — DBMS: description: mirror for downloading:

Additionally it is interesting the change in performance when switching to kernel 3.10 — there is promised growth in the test pgbench 2 times. - Nichole.Lueilwitz commented on October 3rd 19 at 02:50
del - Nichole.Lueilwitz commented on October 3rd 19 at 02:53
Thank you! Looks like what you need. I will try - Neil commented on October 3rd 19 at 02:56
Fragster, Thank you for the reference! :) Got stuck for some time in Gliste ))
- Juwan_Kub64 commented on October 3rd 19 at 02:59
October 3rd 19 at 02:46
Still have a question on testing, when the server 1C: enterprise and DBMS reside on different machines (on different virtualtech one host) and influence of various "physical" networks (including the presence of routers on the way) on this same performance.

Find more questions by tags 1CLinuxSQL ServerPostgreSQLTesting softwareVirtualization