Re: How fast is your server?
This WebDNA talk-list message is from 2002
It keeps the original formatting.
numero = 41834
interpreted = N
texte = A few points about the test.It seems to me like it'll be very dependant on disk I/O since you are writing to a db. I think a replace or math would be a better test of proc speed. We could probably define a broader test to check overall performance. And you'd prolly need to hit it a few times externally with ab or some other app to see any benefit from the dual proc machines. The dual proc mac doesn't really do that bad, considering: Webcat's running the command consecutively, on one proc and The PC's 1.3Ghz faster. Our 733 is obviously low on Ram. I think it's v. interesting the diff between our identical DP 533's, we just have more ram.I'm very eager to see what a dual proc 1ghz xserve with 2gb of ram will do (the config we are getting)I would expect a comparable dual xeon running linux would be a tad faster, but for us the xserve seems a logical choice because it'll integrate with our netinfo domain(s) , and from previous experience, I expect apple's code to get significantly faster. NT is not even a consideration b/c I personally had a machine defaced, and they just seem like bull's eye targets to me...Linux would certainly be an option for us, BUT the friendly support groups for OSX make us feel far more confident we can resolve a problem.And from a sales POV we are marketing ourselves as a mac-friendly place, so it'd be hypocritical not to run OSX.All that being said, I am DYING to see what an xserve will do w/this.On 7/23/02 3:34 AM, Andrew Simpson mashed the following keys :> > It is interesting that you what you say about the Xserve. > > When you can build a Fast as Buggery PC for a $1000.00 NZD is there any > point buying an xserve that might not be able to compete??? > > This is one of the hard choices we are struggling with at the moment - hence > the interest in the speed test. > > For added interest: > > My TiBook does it in 935 > 667mhz G4 > 512Mb > OSX 10.1.5, Apache. > > Anybody have a high spec Wintel or Intel/unix machine out there? > > > > > On 23/7/02 10:26 PM, Aaron Lynch
wrote: > >> On 7/23/02 2:17 AM, Andrew Simpson mashed the following keys : >> >>> >>> We just got a new PC in the office and thought we would compare it to the >>> speed of our mac / osx based webserver as a speed comparison. >>> >>> Using the following code: >>> >>> >>> [loop start=1&end=20000&advance=1] >>> [append db=id.db]ID=[index][/append] >>> [/loop] >>> [elapsedtime] >>> >>> The database was just a single file - (no header) with one field called ID. >>> >>> We found some interesting speed differences between the mac and pc. >>> >>> The PC completed this task in 243 ticks while the mac took 637 ticks. >>> >>> Similar speed differences were noted on tests with the general rule of thumb >>> being that the PC was about 2 to 2.5 time faster at processing. The PC was >>> faster still doing smaller jobs - being up to four times quicker when the >>> task wasn't so large - like creating 200 text variables for example. >>> >>> The two machines are as follows: >>> >>> ThePC >>> 1800+ Palimino XP >>> 256mb Ram >>> MS XP Pro, IIS 5 >>> >>> TheMac >>> Dual 533mhz G4 >>> 764mb Ram >>> OSX 10.1.5, Apache >>> >>> Both machines are running the latest version of Webcat >>> >>> >>> I thought it would be interesting for people out there to run this simple >>> test and tell us how it goes and the platform you tested on. >>> >>> Might give you either something to brag about or something to look at in >>> terms of your configuration. >>> >>> Can anybody do this test in under 100 ticks? >> >> Our results: >>> Dual 533mhz G4 >> 1.5G Ram >>> OSX 10.1.5, Apache >> result: 580 >> >> >> g4 733 >> 256MB ram >> OSX 10.1.5, Apache >> Result 693 >> >> g4 500 >> 1GB ram >> OS 9.2.2, Webstar >> Result 6302 <- BUT it's probably being backed up or something (I'll >> investigate in the morning) >> >> Just wait til we get our xserve.... :) >> >> Also, I added a delete command to your code after elapsed time so that >> subsequent hits have the roughly the same size DB to write to >> >> >> ------------------------------------------------------------- >> This message is sent to you because you are subscribed to >> the mailing list . >> To unsubscribe, E-mail to: >> To switch to the DIGEST mode, E-mail to >> >> Web Archive of this list is at: http://search.smithmicro.com/ ------------------------------------------------------------- This message is sent to you because you are subscribed to the mailing list . To unsubscribe, E-mail to: To switch to the DIGEST mode, E-mail to Web Archive of this list is at: http://search.smithmicro.com/
Associated Messages, from the most recent to the oldest:
A few points about the test.It seems to me like it'll be very dependant on disk I/O since you are writing to a db. I think a replace or math would be a better test of proc speed. We could probably define a broader test to check overall performance. And you'd prolly need to hit it a few times externally with ab or some other app to see any benefit from the dual proc machines. The dual proc mac doesn't really do that bad, considering: Webcat's running the command consecutively, on one proc and The PC's 1.3Ghz faster. Our 733 is obviously low on Ram. I think it's v. interesting the diff between our identical DP 533's, we just have more ram.I'm very eager to see what a dual proc 1ghz xserve with 2gb of ram will do (the config we are getting)I would expect a comparable dual xeon running linux would be a tad faster, but for us the xserve seems a logical choice because it'll integrate with our netinfo domain(s) , and from previous experience, I expect apple's code to get significantly faster. NT is not even a consideration b/c I personally had a machine defaced, and they just seem like bull's eye targets to me...Linux would certainly be an option for us, BUT the friendly support groups for OSX make us feel far more confident we can resolve a problem.And from a sales POV we are marketing ourselves as a mac-friendly place, so it'd be hypocritical not to run OSX.All that being said, I am DYING to see what an xserve will do w/this.On 7/23/02 3:34 AM, Andrew Simpson mashed the following keys :> > It is interesting that you what you say about the Xserve. > > When you can build a Fast as Buggery PC for a $1000.00 NZD is there any > point buying an xserve that might not be able to compete??? > > This is one of the hard choices we are struggling with at the moment - hence > the interest in the speed test. > > For added interest: > > My TiBook does it in 935 > 667mhz G4 > 512Mb > OSX 10.1.5, Apache. > > Anybody have a high spec Wintel or Intel/unix machine out there? > > > > > On 23/7/02 10:26 PM, Aaron Lynch wrote: > >> On 7/23/02 2:17 AM, Andrew Simpson mashed the following keys : >> >>> >>> We just got a new PC in the office and thought we would compare it to the >>> speed of our mac / osx based webserver as a speed comparison. >>> >>> Using the following code: >>> >>> >>> [loop start=1&end=20000&advance=1] >>> [append db=id.db]ID=[index][/append] >>> [/loop] >>> [elapsedtime] >>> >>> The database was just a single file - (no header) with one field called ID. >>> >>> We found some interesting speed differences between the mac and pc. >>> >>> The PC completed this task in 243 ticks while the mac took 637 ticks. >>> >>> Similar speed differences were noted on tests with the general rule of thumb >>> being that the PC was about 2 to 2.5 time faster at processing. The PC was >>> faster still doing smaller jobs - being up to four times quicker when the >>> task wasn't so large - like creating 200 text variables for example. >>> >>> The two machines are as follows: >>> >>> ThePC >>> 1800+ Palimino XP >>> 256mb Ram >>> MS XP Pro, IIS 5 >>> >>> TheMac >>> Dual 533mhz G4 >>> 764mb Ram >>> OSX 10.1.5, Apache >>> >>> Both machines are running the latest version of Webcat >>> >>> >>> I thought it would be interesting for people out there to run this simple >>> test and tell us how it goes and the platform you tested on. >>> >>> Might give you either something to brag about or something to look at in >>> terms of your configuration. >>> >>> Can anybody do this test in under 100 ticks? >> >> Our results: >>> Dual 533mhz G4 >> 1.5G Ram >>> OSX 10.1.5, Apache >> result: 580 >> >> >> g4 733 >> 256MB ram >> OSX 10.1.5, Apache >> Result 693 >> >> g4 500 >> 1GB ram >> OS 9.2.2, Webstar >> Result 6302 <- BUT it's probably being backed up or something (I'll >> investigate in the morning) >> >> Just wait til we get our xserve.... :) >> >> Also, I added a delete command to your code after elapsed time so that >> subsequent hits have the roughly the same size DB to write to >> >> >> ------------------------------------------------------------- >> This message is sent to you because you are subscribed to >> the mailing list . >> To unsubscribe, E-mail to: >> To switch to the DIGEST mode, E-mail to >> >> Web Archive of this list is at: http://search.smithmicro.com/ ------------------------------------------------------------- This message is sent to you because you are subscribed to the mailing list . To unsubscribe, E-mail to: To switch to the DIGEST mode, E-mail to Web Archive of this list is at: http://search.smithmicro.com/
Aaron Lynch
DOWNLOAD WEBDNA NOW!
Top Articles:
Talk List
The WebDNA community talk-list is the best place to get some help: several hundred extremely proficient programmers with an excellent knowledge of WebDNA and an excellent spirit will deliver all the tips and tricks you can imagine...
Related Readings:
all records returned. (1997)
Cookie set browser session. (1998)
WebCat2 - [format thousands] (1997)
Emailer file formats (1998)
[WebDNA] Quick Grep question (2009)
using showpage and showcart commands (1996)
Where's Cart Created ? (1997)
Locking up with WebCatalog... (1997)
WebMerchant 1.6 and https (1997)
timing out? (1997)
[isfile] ? (1997)
Secure Server (1997)
Tax on Shipping question (1998)
Big Databases (1997)
help needed: Non-english characters in WebCatalog (1997)
AAgghh!! Help, please. SSL strikes again. (1997)
General WebCatalog Questions (1997)
Requiring that certain fields be completed (1997)
PIXO Support (1997)
Shipping Help! (1998)