[gridengine users] SQL spooling

2011-09-09 Thread Chi Chan
It is good that Ron Chen said that there is new features (IPV6, GPU integration) -- but any forks working on SQL spooling? We don't need GPU integration now (with AMD bulldozer shipping soon, the need for a seperate GPU for handling GPGPU type of workload may be even less), and our clusters wil

Re: [gridengine users] SQL spooling

2011-09-09 Thread Hung-Sheng Tsao (Lao Tsao 老曹) Ph.D.
recently oracle introduce a SQLlite frontend to berkleyDB not sure you can take advantage this new FE to query the berkleyDB in GE regards On 9/9/2011 2:51 PM, Chi Chan wrote: It is good that Ron Chen said that there is new features (IPV6, GPU integration) -- but any forks working on SQL spoo

Re: [gridengine users] SQL spooling

2011-09-11 Thread Reuti
Hi, Am 09.09.2011 um 20:51 schrieb Chi Chan: > It is good that Ron Chen said that there is new features (IPV6, GPU > integration) -- but any forks working on SQL spooling? > > We don't need GPU integration now (with AMD bulldozer shipping soon, the need > for a seperate GPU for handling GPGPU

Re: [gridengine users] SQL spooling

2011-09-11 Thread Rayson Ho
On Sun, Sep 11, 2011 at 7:04 PM, Reuti wrote: > I have a vague memory about this. In which version of SGE was it? Eventually > BDB was chosen to replace the flat file format. There are still a large number of sites using flatfile / classic spooling - for small to medium sites Berkeley DB is usua

Re: [gridengine users] SQL spooling

2011-09-26 Thread Rayson Ho
Andy used to tell us that the original SQL spooling was slower than Berkeley DB spooling and thus it was not used. However, the original SQL spooling was written for PostgreSQL, which is a client/server DB. An embedded SQLite DB should be much faster, as SQLite runs in the same process as the qmast

Re: [gridengine users] SQL spooling

2011-09-27 Thread Ron Chen
--- On Tue, 9/27/11, Rayson Ho wrote: > I just did a very quick test with SQLite 3.6.22, some of > the SQL > statements used by SGE were able to run against an SQLite > DB without > any changes - in the end, we are doing simple > INSERT/UPDATE/DELETE > SQL, plus some table maintenance code & table

Re: [gridengine users] SQL spooling

2011-10-22 Thread Chi Chan
this thread? IMHO, SQLite seems to be easier to maintain and far better as a spooling database. --Chi - 原始信件 寄件者: Ron Chen 收件者: users@gridengine.org; Rayson Ho 副本: 寄件日期: 2011/9/27 (二) 11:35 AM 主旨: Re: [gridengine users] SQL spooling --- On Tue, 9/27/11, Rayson Ho wrote: > I jus

Re: [gridengine users] SQL spooling

2011-10-24 Thread William Bryce
er as a spooling > database. > > > --Chi > > > > > - 原始信件 > 寄件者: Ron Chen > 收件者: users@gridengine.org; Rayson Ho > 副本: > 寄件日期: 2011/9/27 (二) 11:35 AM > 主旨: Re: [gridengine users] SQL spooling > > --- On Tue, 9/27/11, Rayson Ho w

Re: [gridengine users] SQL spooling

2011-10-24 Thread Rayson Ho
gt; >> >> >> ----- 原始信件 >> 寄件者: Ron Chen >> 收件者: users@gridengine.org; Rayson Ho >> 副本: >> 寄件日期: 2011/9/27 (二) 11:35 AM >> 主旨: Re: [gridengine users] SQL spooling >> >> --- On Tue, 9/27/11, Rayson Ho wrote: >>> I just did a v