Hi,

> No, Roman - it is not related either to kernel nor to jvm at all. But I

think

> I know to what parameter it is really related. Why don't you think that

this is a

> page size of db. When the buffer size equals to db page size there is no

need

> to wast time or resources on synchronization of data transmition. When

they are

> not equal everything immedeately is slowing down. E.g my db page size is

8kb.

> And the best result - about 6sec - was obtained at value of 8*1024 of

buffers sizes.

> Other values - 16, 32, 64 *1024 - give the same and worse result about

18sec

> , i.e.3 times slower.

1. My assumption was wrong - I checked with 16Kb page size - the best
results
were with receive/send buffer size again at 8192. Does anybody know why?
2. Another my mistake - on other values of buffers size results are slower
not in 3 times but in 30 times approximately. This is big difference. I'm
sorry for
misprint.

Thanks,
Sergei Volin.