Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-22 Thread Sokolov Yura
Simplified a bit and more commented patch version is in attach. Algorithm were switched to linear probing, it makes code simpler and clearer. Flag usages were toggled: now it indicates that hash table were built, it also makes code clearer. XidInXip and some other functions got comments. --

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-14 Thread Robert Haas
On Sun, Aug 13, 2017 at 11:05 AM, Sokolov Yura wrote: > BTW, ad-hoc hash tables already exist: at least recourse owner uses one. Yeah. :-( -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company -- Sent via pgsql-hackers

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-13 Thread Sokolov Yura
В Fri, 11 Aug 2017 14:05:08 -0400 Robert Haas пишет: > On Thu, Aug 10, 2017 at 11:12 AM, Alexander Korotkov > wrote: > > These results look very cool! > > I think CSN is eventually inevitable, but it's a long distance > > feature. Thus, this

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-11 Thread Robert Haas
On Thu, Aug 10, 2017 at 11:12 AM, Alexander Korotkov wrote: > These results look very cool! > I think CSN is eventually inevitable, but it's a long distance feature. > Thus, this optimization could make us a good serve before we would have CSN. > Do you expect there are

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-11 Thread Sokolov Yura
В Thu, 10 Aug 2017 18:12:34 +0300 Alexander Korotkov пишет: > On Thu, Aug 10, 2017 at 3:30 PM, Sokolov Yura > wrote: > > > On 2017-07-31 18:56, Sokolov Yura wrote: > > > >> Good day, every one. > >> > >> In attempt to improve

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-10 Thread Alexander Korotkov
On Thu, Aug 10, 2017 at 3:30 PM, Sokolov Yura wrote: > On 2017-07-31 18:56, Sokolov Yura wrote: > >> Good day, every one. >> >> In attempt to improve performance of YCSB on zipfian distribution, >> it were found that significant time is spent in XidInMVCCSnapshot in

Re: [HACKERS] Lazy hash table for XidInMVCCSnapshot (helps Zipfian a bit)

2017-08-10 Thread Sokolov Yura
On 2017-07-31 18:56, Sokolov Yura wrote: Good day, every one. In attempt to improve performance of YCSB on zipfian distribution, it were found that significant time is spent in XidInMVCCSnapshot in scanning snapshot->xip array. While overall CPU time is not too noticable, it has measurable