my test results of the .18 version
Hi
I thought everyone would be interested in my test results of the .18 version.
As per my post at http://crlab.com/forums/viewtopic.php?t=732 , the .17 version had slowed down a lot from a much older version that I had.
Here are the results including the latest testing - all numbers are in Ms, and the triplets are V3.10.0.5, V3.50.17, V3.50.18. They bounce around a bit each run, but are pretty consistent. I'm using a table of 100,000 records in a local database.
Read Speed (ie select * where..)
1353, 3446 (!), 1067 - The major select bug has been fixed in V3.50.18!
Write spead (ie insert ..)
40043, 39022, 43055 ie V3.50.18 has slowed down inserts by about 10%. It is still much faster than the micoOLAP product though. Any thoughts why .18 might have slowed down on inserts?
Aggregate (ie select sum(..) group by ..)
136, 157, 136 - problem fixed.
In summary, you HAVE to download this release and install it!
I thought everyone would be interested in my test results of the .18 version.
As per my post at http://crlab.com/forums/viewtopic.php?t=732 , the .17 version had slowed down a lot from a much older version that I had.
Here are the results including the latest testing - all numbers are in Ms, and the triplets are V3.10.0.5, V3.50.17, V3.50.18. They bounce around a bit each run, but are pretty consistent. I'm using a table of 100,000 records in a local database.
Read Speed (ie select * where..)
1353, 3446 (!), 1067 - The major select bug has been fixed in V3.50.18!
Write spead (ie insert ..)
40043, 39022, 43055 ie V3.50.18 has slowed down inserts by about 10%. It is still much faster than the micoOLAP product though. Any thoughts why .18 might have slowed down on inserts?
Aggregate (ie select sum(..) group by ..)
136, 157, 136 - problem fixed.
In summary, you HAVE to download this release and install it!
Nope, not him and not connected to devshed.ben wrote:r34, is your name "rudy". This r34 looks like a mysql expert guy in DevShed.
The r34 comes from the car I drive, a R34. If you are a car nut you'll know what one of them is, though if you are in the USA you would probably never have seen one... Think REAL fast...
And yes, to the implied question, I do know a thing or two about programming, and databases
> Write spead (ie insert ..)
> 40043, 39022, 43055 ie V3.50.18 has slowed down inserts by about 10%. It is
> still much faster than the micoOLAP
> product though. Any thoughts why .18 might have slowed down on inserts?
Most likely, the problem just is in improvement of Fetch performance. We have already corrected a small error that caused a delay in Insert execution.
If you send us complete small sample we could give you more certain answer.
> 40043, 39022, 43055 ie V3.50.18 has slowed down inserts by about 10%. It is
> still much faster than the micoOLAP
> product though. Any thoughts why .18 might have slowed down on inserts?
Most likely, the problem just is in improvement of Fetch performance. We have already corrected a small error that caused a delay in Insert execution.
If you send us complete small sample we could give you more certain answer.
Here are the results including the latest testing - all numbers are in Ms, and the triplets are V3.10.0.5, V3.50.17, V3.50.18, V3.50.19. They bounce around a bit each run, but are pretty consistent. I'm using a table of 100,000 records in a local database.
Read Speed (ie select * where..)
1353, 3446 (!), 1067, 1040 - There seems to be a slight, but consistent, improvement from V.18 to V.19
Write spead (ie insert ..)
40043, 39022, 43055, 38000 ie not only has V3.50.19 fixed the V.18 write slow down, it now appears to be the fastest release yet .
Aggregate (ie select sum(..) group by ..)
136, 157, 136, 138 No change from v.18
In summary, you should download the V.19 release and install it
On another note, the writes above do one record at a time. This is the slow way of course, and batching them up 1000 at a time speeds up the write by a factor of three on my local database, and many times that on my remote database...
Given releases are coming out so fast - not a complaint, as we all like fast fixes to problems! - is it possible for you to add a Version() function somewhere so that our application programs can tell what version they are using????
Read Speed (ie select * where..)
1353, 3446 (!), 1067, 1040 - There seems to be a slight, but consistent, improvement from V.18 to V.19
Write spead (ie insert ..)
40043, 39022, 43055, 38000 ie not only has V3.50.19 fixed the V.18 write slow down, it now appears to be the fastest release yet .
Aggregate (ie select sum(..) group by ..)
136, 157, 136, 138 No change from v.18
In summary, you should download the V.19 release and install it
On another note, the writes above do one record at a time. This is the slow way of course, and batching them up 1000 at a time speeds up the write by a factor of three on my local database, and many times that on my remote database...
Given releases are coming out so fast - not a complaint, as we all like fast fixes to problems! - is it possible for you to add a Version() function somewhere so that our application programs can tell what version they are using????