Re: [sqlite] Concurrent read performance

2013-01-14 Thread Howard Chu
Wayne Bradney wrote: Howard,Thanks for the links - mdb/sqlightning looks interesting.Before I dive in, however:1. As I understand it, sqlightning is a drop-in replacement for SQLite.Interop.dll, and I can still use System.Data.SQLite as my high-level interface? Right. 2. Is there a compiled

Re: [sqlite] Concurrent read performance

2013-01-14 Thread Wayne Bradney
qlite.org > Subject: Re: [sqlite] Concurrent read performance > > Wayne Bradney wrote: > >>> All access in SQLite is serialized. Apologies if I'm missing something > >>> fundamental here, but that's not what I'm seeing with a file-backed > >>> databa

Re: [sqlite] Concurrent read performance

2013-01-13 Thread Howard Chu
ite ported to use MDB as its backend is available here https://gitorious.org/mdb/sqlightning Date: Sat, 12 Jan 2013 17:48:56 + To: sqlite-users@sqlite.org Subject: Re: [sqlite] Concurrent read performance On 12 Jan 2013, at 5:38pm, Wayne Bradney <wayne_brad...@yahoo.com> wrote: &qu

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Wayne Bradney
still appear to be serialized. Would this approach require SQLITE_CONFIG_MULTITHREAD or SQLITE_OPEN_NOMUTEX, I wonder? > Date: Sat, 12 Jan 2013 13:39:35 -0700 > From: kmedc...@dessus.com > To: sqlite-users@sqlite.org > Subject: Re: [sqlite] Concurrent read performance > > >

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Keith Medcalf
e.org [mailto:sqlite-users- > boun...@sqlite.org] On Behalf Of Wayne Bradney > Sent: Saturday, 12 January, 2013 12:36 > To: General Discussion of SQLite Database > Subject: Re: [sqlite] Concurrent read performance > > Increasing cache_size didn't seem to have an effect. I think I'm goin

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Michael Black
ht be in terms of SQLite configuration of > memory-backed databases. > From: slav...@bigfraud.org >> Date: Sat, 12 Jan 2013 17:48:56 + >> To: sqlite-users@sqlite.org >> Subject: Re: [sqlite] Concurrent read performance >> >> >> On 12 Jan 20

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Wayne Bradney
Re: [sqlite] Concurrent read performance > > Also...does increasing cache_size help? > Are you able to use a RAM disk? > > > -Original Message- > From: sqlite-users-boun...@sqlite.org > [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Wayne Bradney > Se

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Wayne Bradney
but were surprised by the apparent lack of read-concurrency, and wanted to get some input on what our options might be in terms of SQLite configuration of memory-backed databases. > From: slav...@bigfraud.org > Date: Sat, 12 Jan 2013 17:48:56 + > To: sqlite-users@sqlite.org > Su

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Wayne Bradney
read_uncommitted didn't seem to have any effect > From: mdblac...@yahoo.com > To: sqlite-users@sqlite.org > Date: Sat, 12 Jan 2013 11:47:55 -0600 > Subject: Re: [sqlite] Concurrent read performance > > Did you try read-uncommitted? > Sounds promising... > > 2.

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Clemens Ladisch
Wayne Bradney wrote: > 1. when shared cache is enabled, all reads are serialized, Yes. > I guess I MUST use a file-backed database to get concurrent reads, > even though I don't need the persistence and don't want to take the > I/O hit. If the in-memory database works, you do have enough memory

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Michael Black
Also...does increasing cache_size help? Are you able to use a RAM disk? -Original Message- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Wayne Bradney Sent: Saturday, January 12, 2013 11:39 AM To: sqlite-users@sqlite.org Subject: [sqlite]

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Simon Slavin
On 12 Jan 2013, at 5:38pm, Wayne Bradney wrote: > "mode=memory=shared" > 1. when shared cache is enabled, all reads are serialized, and All access in SQLite is serialised. All transactions require locking the entire database. SQLite is very simple -- 'lite' -- so

Re: [sqlite] Concurrent read performance

2013-01-12 Thread Michael Black
Did you try read-uncommitted? Sounds promising... 2.2.1 http://www.sqlite.org/sharedcache.html PRAGMA read_uncommitted = ; -Original Message- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org] On Behalf Of Wayne Bradney Sent: Saturday, January 12, 2013 11:39