Re: Benchmark()

[Gleb Paharenko]

Hello.



In your query BENCHMARK() doesn't execute the query, it is

passed just like a string to the second argument of the BENCHMARK().





Raimundo Sierra wrote:

> It is not clear to me what benchmark really does. Example:

> 

> SELECT SQL_NO_CACHE * FROM mytable where myVarcharRow like 'sometext%'

> or otherVarcharRow like 'someothertext';

> 

> takes approx. 0.3 seconds. If you benchmark this query, which to my

> understanding should run and not just parse the query, it is faster??

> 

> SELECT BENCHMARK(1000, "SELECT SQL_NO_CACHE * FROM mytable where

> myVarcharRow like 'sometext %' or otherVarcharRow like 'othertext %'");

> 

> returns 1 row in set (0.12 sec)

> 

> Another similar example is:

> 

> SELECT * FROM Table;

> returns: 59770 rows in set (0.23 sec)

> SELECT BENCHMARK(10, "SELECT * FROM Table");

> returns: 1 row in set (0.01 sec)

> 



-- 
For technical support contracts, goto https://order.mysql.com/?ref=ensita
This email is sponsored by Ensita.NET http://www.ensita.net/
   __  ___ ___   __
  /  |/  /_ __/ __/ __ \/ /Gleb Paharenko
 / /|_/ / // /\ \/ /_/ / /__   [EMAIL PROTECTED]
/_/  /_/\_, /___/\___\_\___/   MySQL AB / Ensita.NET
   <___/   www.mysql.com




-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

RE: Benchmark()

[emierzwa]

The document says it evaluates the 'expr', so I'm guessing it is only
function calls and arithmetic evaluation. Although I have tried selects
without any luck. It could probably be documented a little more clearly,
or event better yet...a supported feature.

Ed

-Original Message-
From: Raimundo Sierra [mailto:[EMAIL PROTECTED] 
Sent: Wednesday, December 14, 2005 9:37 PM
To: mysql@lists.mysql.com
Subject: Benchmark()

It is not clear to me what benchmark really does. Example:

SELECT SQL_NO_CACHE * FROM mytable where myVarcharRow like 'sometext%' 
or otherVarcharRow like 'someothertext';

takes approx. 0.3 seconds. If you benchmark this query, which to my 
understanding should run and not just parse the query, it is faster??

SELECT BENCHMARK(1000, "SELECT SQL_NO_CACHE * FROM mytable where 
myVarcharRow like 'sometext %' or otherVarcharRow like 'othertext %'");

returns 1 row in set (0.12 sec)

Another similar example is:

SELECT * FROM Table;
returns: 59770 rows in set (0.23 sec)
SELECT BENCHMARK(10, "SELECT * FROM Table");
returns: 1 row in set (0.01 sec)

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark of MyISAM vs Innodb vs Innod without FKs?!

[Gleb Paharenko]

Hello.



Usually only benchmarks will show a real picture

for you. Create foreign keys, perform some tests. Then

temporary disable FKs using SET FOREIGN_KEY_CHECKS=0 and 

repeat the performance measurement. Super Smack is a good

tool for such kind of analysis. See:

  http://dev.mysql.com/doc/mysql/en/custom-benchmarks.html





Kevin Burton <[EMAIL PROTECTED]> wrote:

>ere's some thing I've been thinking about.

>

>I want to use INNODB without FKs. I don't need or want referential integrity 

>in my app (due to a schema and performance issue).

>

>Basically I just create FKs in my OR layer and my app enforces the rules. 

>The column is still an _ID column so I visually know a FK when I see one but 

>INNODB doesn't have to do any runtime checks on insert.

>

>My question is whether INNODB will be faster without them. If so by how 

>much. If it's close to the speed of MyISAM then I'll be a happy camper.

>

>Thoughts?

>

>Kevin



-- 
For technical support contracts, goto https://order.mysql.com/?ref=ensita
This email is sponsored by Ensita.NET http://www.ensita.net/
   __  ___ ___   __
  /  |/  /_ __/ __/ __ \/ /Gleb Paharenko
 / /|_/ / // /\ \/ /_/ / /__   [EMAIL PROTECTED]
/_/  /_/\_, /___/\___\_\___/   MySQL AB / Ensita.NET
   <___/   www.mysql.com




-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Gabriel Ricard]

On Friday, November 7, 2003, at 09:05  AM, David Steinbrunner wrote:


The Mac was HFS+ journaled. Disk: the stock Apple-supplied one. The
Linux machine was a default SuSE 8.0 installation. ext2 as the
filesystem? No idea about journaling. No SCSI or RAID, just an
internal IDE disk. Both machines are really consumer-level machines,
no heavy-duty server hardware. That's about all I can tell you, I
really don't know that much about the details of the hardware. Sorry,
not much help here, I know.
Hmm.  The Mac had journaling but the Linux box did not (ext2 is not
journaled).  That could have influenced things too.  I know virtually
nothing about HFS+ journaling.
You can turn off journaling in 10.3 in Disk Utility by going to File 
and selecting  Disable Journaling.  This page has some visuals:
http://diveintoosx.org/panther/disk_management.html

There is a Terminal command that does the same thing that I don't know 
off the top of my head.
diskutil disableJournal disk0s1

- Gabriel

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[David Steinbrunner]

The Mac was HFS+ journaled. Disk: the stock Apple-supplied one. The
Linux machine was a default SuSE 8.0 installation. ext2 as the
filesystem? No idea about journaling. No SCSI or RAID, just an
internal IDE disk. Both machines are really consumer-level machines,
no heavy-duty server hardware. That's about all I can tell you, I
really don't know that much about the details of the hardware. Sorry,
not much help here, I know.
Hmm.  The Mac had journaling but the Linux box did not (ext2 is not
journaled).  That could have influenced things too.  I know virtually
nothing about HFS+ journaling.
You can turn off journaling in 10.3 in Disk Utility by going to File 
and selecting  Disable Journaling.  This page has some visuals:
http://diveintoosx.org/panther/disk_management.html

There is a Terminal command that does the same thing that I don't know 
off the top of my head.

--
David Steinbrunner
MFM Communication Software, Inc.
http://www.mfm.com
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jeremy Zawodny]

On Thu, Nov 06, 2003 at 07:36:55PM +0100, Jan Pieter Kunst wrote:
> Jeremy D. Zawodny:
> 
> >  > I was wondering if there is something I can do, configuration-wise, to
> >>  do something about those very slow 'inserts' (and 'updates') on the Mac?
> >>
> >>  Thanks in advance for any insight,
> >
> >Did it appear to be disk or CPU bound?
> 
> Sorry, newbie here. I don't know how I can tell. The RUN file says 
> this about the 'insert':
> 
> insert: Total time: 6811 wallclock secs (881.86 usr  0.00 sys +  0.00 
> cusr  0.00 csys = 881.86 CPU)
> 
> Does that mean anything in this regard?

Run the benchmark again.  Notice if the CPU is consistently around
100% or not.  If not, it's disk bound.  Otherwise, it's CPU bound.

> >What are the filesystems like on each?  Journaling on either?  Hard
> >disks and cache?  RAID/SCSI controllers?
> 
> The Mac was HFS+ journaled. Disk: the stock Apple-supplied one. The 
> Linux machine was a default SuSE 8.0 installation. ext2 as the 
> filesystem? No idea about journaling. No SCSI or RAID, just an 
> internal IDE disk. Both machines are really consumer-level machines, 
> no heavy-duty server hardware. That's about all I can tell you, I 
> really don't know that much about the details of the hardware. Sorry, 
> not much help here, I know.

Hmm.  The Mac had journaling but the Linux box did not (ext2 is not
journaled).  That could have influenced things too.  I know virtually
nothing about HFS+ journaling.

> Anyway, I now have Panther (Mac OS X 10.3) and the Perl DBI/DBD stuff 
> installed on my G5 and I'll run the benchmarks tonight. (I earlier 
> ran the benchmarks under Jaguar (Mac OS X 10.2) on the G5, and I also 
> got a relative 'spike' in the insert-test compared to the other 
> tests, just as on the G4.) I'll report the Panther-results later.

Cool.

Jeremy
-- 
Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
<[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/

MySQL 4.0.15-Yahoo-SMP: up 53 days, processed 2,006,750,860 queries (431/sec. avg)

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jan Pieter Kunst]

I just installed Panther on my G5 at home. Unfortunately, for some 
reason I can't get the Perl module DBD::mysql to install (using 
CPAN, had no problems doing this in Jaguar) so I can't run the 
benchmark suite for now.

JP


I had similar issues, but I was also using a custom build of MySQL, 
which was not in a standard location.
I found this solution:



JP

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jan Pieter Kunst]

Jeremy D. Zawodny:

 > I was wondering if there is something I can do, configuration-wise, to
 do something about those very slow 'inserts' (and 'updates') on the Mac?

 Thanks in advance for any insight,
Did it appear to be disk or CPU bound?
Sorry, newbie here. I don't know how I can tell. The RUN file says 
this about the 'insert':

insert: Total time: 6811 wallclock secs (881.86 usr  0.00 sys +  0.00 
cusr  0.00 csys = 881.86 CPU)

Does that mean anything in this regard?

What are the filesystems like on each?  Journaling on either?  Hard
disks and cache?  RAID/SCSI controllers?
The Mac was HFS+ journaled. Disk: the stock Apple-supplied one. The 
Linux machine was a default SuSE 8.0 installation. ext2 as the 
filesystem? No idea about journaling. No SCSI or RAID, just an 
internal IDE disk. Both machines are really consumer-level machines, 
no heavy-duty server hardware. That's about all I can tell you, I 
really don't know that much about the details of the hardware. Sorry, 
not much help here, I know.

Anyway, I now have Panther (Mac OS X 10.3) and the Perl DBI/DBD stuff 
installed on my G5 and I'll run the benchmarks tonight. (I earlier 
ran the benchmarks under Jaguar (Mac OS X 10.2) on the G5, and I also 
got a relative 'spike' in the insert-test compared to the other 
tests, just as on the G4.) I'll report the Panther-results later.

JP

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Gabriel Ricard]

On Thursday, November 6, 2003, at 11:38  AM, Jan Pieter Kunst wrote:

I'd be curious what kind of numbers Panther shows. Once I get my 
xServe setup, just arrived, I'll try running some tests myself.
I just installed Panther on my G5 at home. Unfortunately, for some 
reason I can't get the Perl module DBD::mysql to install (using CPAN, 
had no problems doing this in Jaguar) so I can't run the benchmark 
suite for now.

JP


I had similar issues, but I was also using a custom build of MySQL, 
which was not in a standard location.

Try this:

perl Makefile.PL --cflags=-I/path/to/include/mysql 
"--libs=-L/path/to/lib/mysql -lmysqlclient"
export MACOSX_DEPLOYMENT_TARGET=10.3
make

If you get errors during the make process, try editing the Makefile to 
make sure it has the appropriate paths in these lines (39 + 40):

LDDLFLAGS = -bundle -undefined dynamic_lookup -L/path/to/mysql/lib
LDFLAGS = -L/path/to/mysql/lib
The only other thing I could do to get it to compile was to copy the 
part of the failed command from the 'cc' to the end of the line, e.g.:

LD_RUN_PATH="/imax/lib/mysql" /usr/bin/perl myld 
MACOSX_DEPLOYMENT_TARGET=10.3 cc  -bundle -undefined dynamic_lookup 
dbdimp.o mysql.o  -o blib/arch/auto/DBD/mysql/mysql.bundle   
-L/imax/lib/mysql -lmysqlclient

becomes:

cc  -bundle -undefined dynamic_lookup dbdimp.o mysql.o  -o 
blib/arch/auto/DBD/mysql/mysql.bundle   -L/path/to/lib/mysql 
-lmysqlclient

then type make again and see if it finishes.

Klugey, but it worked for me.

- Gabriel

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jeremy Zawodny]

On Thu, Nov 06, 2003 at 10:19:59AM -0700, Van wrote:
> 
> Jeremy:
> 
> I believe that's what I was attempting to convey.  My suggestion was that the
> G4 should not be disk bound (limited by disk bandwidth).  The CPU on my DEC
> Alpha is a meagerly 300, but the disk can stream faster than my PIII true Intel
> 733MHz.  That's why I suggested he look at the disk / controller.  INSERT = the
> need for disk speed.  The CPU doesn't really have to process much unless it's
> IDE, and even then it would be minimal.
> 
> Somewhat correct?  In other words, I wouldn't take your bet.

Okay, so we are on the same page. :-)

I was rather confused by what you were trying to say in the first
message.  When I read:

  I can't imagine a G4 would be disk bound relative to an Intel
  machine, unless there is something very wrong with the disk or
  controller.

I've seen cases when there was nothing "wrong" on either side, but one
was way faster than the other.

Jeremy
-- 
Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
<[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Van]

Jeremy:

I believe that's what I was attempting to convey.  My suggestion was that the
G4 should not be disk bound (limited by disk bandwidth).  The CPU on my DEC
Alpha is a meagerly 300, but the disk can stream faster than my PIII true Intel
733MHz.  That's why I suggested he look at the disk / controller.  INSERT = the
need for disk speed.  The CPU doesn't really have to process much unless it's
IDE, and even then it would be minimal.

Somewhat correct?  In other words, I wouldn't take your bet.

Van

=
http://artists.mp3s.com/artists/113/ded_serius.html
Linux rocks!!!   http://www.dedserius.com
=

On Thu, 6 Nov 2003, Jeremy Zawodny wrote:

> On Thu, Nov 06, 2003 at 09:56:22AM -0700, Van wrote:
> >
> > I can't imagine a G4 would be disk bound relative to an Intel machine, unless
> > there is something very wrong with the disk or controller.  Also, you might want
> > to defrag your disk on the Mac.
> >
> > G4s have much more disk bandwidth than any Intel I've ever seen.  Almost as much
> > as my DEC Alpha.  >:)
>
> Huh?
>
> CPUs have little if anything to do with disk I/O operations.  I'd bet
> you $1000 that a 1 CPU Pentium II with the right disk subsystem could
> easily outrun a much "faster" 4 CPU Pentium III system with a single
> IDE disk in it.
>
> Jeremy
> --
> Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
> <[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/
>
> MySQL 4.0.15-Yahoo-SMP: up 53 days, processed 2,004,157,804 queries (431/sec. avg)
>


-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jeremy Zawodny]

On Thu, Nov 06, 2003 at 09:56:22AM -0700, Van wrote:
> 
> I can't imagine a G4 would be disk bound relative to an Intel machine, unless
> there is something very wrong with the disk or controller.  Also, you might want
> to defrag your disk on the Mac.
> 
> G4s have much more disk bandwidth than any Intel I've ever seen.  Almost as much
> as my DEC Alpha.  >:)

Huh?

CPUs have little if anything to do with disk I/O operations.  I'd bet
you $1000 that a 1 CPU Pentium II with the right disk subsystem could
easily outrun a much "faster" 4 CPU Pentium III system with a single
IDE disk in it.

Jeremy
-- 
Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
<[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/

MySQL 4.0.15-Yahoo-SMP: up 53 days, processed 2,004,157,804 queries (431/sec. avg)

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Van]

I can't imagine a G4 would be disk bound relative to an Intel machine, unless
there is something very wrong with the disk or controller.  Also, you might want
to defrag your disk on the Mac.

G4s have much more disk bandwidth than any Intel I've ever seen.  Almost as much
as my DEC Alpha.  >:)

Van

=
http://artists.mp3s.com/artists/113/ded_serius.html
Linux rocks!!!   http://www.dedserius.com
=

On Thu, 6 Nov 2003, Jeremy Zawodny wrote:

> On Wed, Nov 05, 2003 at 07:57:00PM +0100, Jan Pieter Kunst wrote:
> > Hi everyone,
> >
> > I recently ran the MySQL benchmark suite on a Dual 1 GHz G4 running Mac
> > OS X Server 10.2.8, and an 800 MHz Intel machine running SuSE Linux 8.0.
> > Both installations used the same my.cnf file.
> >
> > The results are comparable in all benchmarks except one: the 'insert'.
> > In that one, the Mac is more than twice as slow. Below are the benchmark
> > results for both machines, and the my.cnf I used.
> >
> > I was wondering if there is something I can do, configuration-wise, to
> > do something about those very slow 'inserts' (and 'updates') on the Mac?
> >
> > Thanks in advance for any insight,
>
> Did it appear to be disk or CPU bound?
>
> What are the filesystems like on each?  Journaling on either?  Hard
> disks and cache?  RAID/SCSI controllers?
>
> Jeremy
> --
> Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
> <[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/
>
> MySQL 4.0.15-Yahoo-SMP: up 53 days, processed 2,000,832,379 queries (431/sec. avg)





-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jan Pieter Kunst]

I'd be curious what kind of numbers Panther shows. Once I get my 
xServe setup, just arrived, I'll try running some tests myself.
I just installed Panther on my G5 at home. Unfortunately, for some 
reason I can't get the Perl module DBD::mysql to install (using CPAN, 
had no problems doing this in Jaguar) so I can't run the benchmark 
suite for now.

JP

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Jeremy Zawodny]

On Wed, Nov 05, 2003 at 07:57:00PM +0100, Jan Pieter Kunst wrote:
> Hi everyone,
> 
> I recently ran the MySQL benchmark suite on a Dual 1 GHz G4 running Mac
> OS X Server 10.2.8, and an 800 MHz Intel machine running SuSE Linux 8.0.
> Both installations used the same my.cnf file.
> 
> The results are comparable in all benchmarks except one: the 'insert'.
> In that one, the Mac is more than twice as slow. Below are the benchmark
> results for both machines, and the my.cnf I used.
> 
> I was wondering if there is something I can do, configuration-wise, to
> do something about those very slow 'inserts' (and 'updates') on the Mac?
> 
> Thanks in advance for any insight,

Did it appear to be disk or CPU bound?

What are the filesystems like on each?  Journaling on either?  Hard
disks and cache?  RAID/SCSI controllers?

Jeremy
-- 
Jeremy D. Zawodny |  Perl, Web, MySQL, Linux Magazine, Yahoo!
<[EMAIL PROTECTED]>  |  http://jeremy.zawodny.com/

MySQL 4.0.15-Yahoo-SMP: up 53 days, processed 2,000,832,379 queries (431/sec. avg)

-- 
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Gabriel Ricard]

FYI, I've found 'iostat' to be quite useful in monitoring the drive 
transfer rates while benchmarking in OSX. Then again, in Panther all 
you really need to do is pop open Activity Monitor (formerly Process 
Viewer) which now has some nifty graphing for system status: cpu, disk, 
ram activity and usage, etc.

- Gabriel

On Wednesday, November 5, 2003, at 03:21  PM, Brent Baisley wrote:

I'd be curious what the specs of the hard drives are. Using the stock 
drives in the Mac means you are using a drive that's about average 
(2MB cache, 7200RPM). I would assume they are both ATA/IDE drives.  
But I would guess the bottleneck is the drive. Try running top when 
you are running your tests to see where the bottleneck is.

OSX is also a work in progress, I/O is a big area that Apple is 
improving on. I thought I remember reading that 10.2 was just reaching 
the throughput you would get in OS9.
I'd be curious what kind of numbers Panther shows. Once I get my 
xServe setup, just arrived, I'll try running some tests myself.

On Wednesday, November 5, 2003, at 01:57 PM, Jan Pieter Kunst wrote:

Hi everyone,

I recently ran the MySQL benchmark suite on a Dual 1 GHz G4 running 
Mac
OS X Server 10.2.8, and an 800 MHz Intel machine running SuSE Linux 
8.0.
Both installations used the same my.cnf file.

The results are comparable in all benchmarks except one: the 'insert'.
In that one, the Mac is more than twice as slow. Below are the 
benchmark
results for both machines, and the my.cnf I used.

I was wondering if there is something I can do, configuration-wise, to
do something about those very slow 'inserts' (and 'updates') on the 
Mac?

--
Brent Baisley
Systems Architect
Landover Associates, Inc.
Search & Advisory Services for Advanced Technology Environments
p: 212.759.6400/800.759.0577
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:
http://lists.mysql.com/[EMAIL PROTECTED]



--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Brent Baisley]

I'd be curious what the specs of the hard drives are. Using the stock 
drives in the Mac means you are using a drive that's about average (2MB 
cache, 7200RPM). I would assume they are both ATA/IDE drives.  But I 
would guess the bottleneck is the drive. Try running top when you are 
running your tests to see where the bottleneck is.

OSX is also a work in progress, I/O is a big area that Apple is 
improving on. I thought I remember reading that 10.2 was just reaching 
the throughput you would get in OS9.
I'd be curious what kind of numbers Panther shows. Once I get my xServe 
setup, just arrived, I'll try running some tests myself.

On Wednesday, November 5, 2003, at 01:57 PM, Jan Pieter Kunst wrote:

Hi everyone,

I recently ran the MySQL benchmark suite on a Dual 1 GHz G4 running Mac
OS X Server 10.2.8, and an 800 MHz Intel machine running SuSE Linux 
8.0.
Both installations used the same my.cnf file.

The results are comparable in all benchmarks except one: the 'insert'.
In that one, the Mac is more than twice as slow. Below are the 
benchmark
results for both machines, and the my.cnf I used.

I was wondering if there is something I can do, configuration-wise, to
do something about those very slow 'inserts' (and 'updates') on the 
Mac?

--
Brent Baisley
Systems Architect
Landover Associates, Inc.
Search & Advisory Services for Advanced Technology Environments
p: 212.759.6400/800.759.0577
--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql
To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]

Re: Benchmark differences: Mac OS X - Linux

[Gabriel Ricard]

I'm not entirely sure what to do about the slow insert results, they 
are the slowest part no matter how you configure it, it seems. I've 
attached some benchmark results I ran on a dual 2GHz G5 for comparison. 
Both MyISAM and InnoDB.

Here are the insert results though:

MyISAM:
insert: Total time: 1437 wallclock secs (432.84 usr 358.12 sys +  0.00 
cusr  0.00 csys = 790.96 CPU)

InnoDB:
insert: Total time: 3270 wallclock secs (440.53 usr 361.82 sys +  0.00 
cusr  0.00 csys = 802.35 CPU)

- Gabriel

Benchmark DBD suite: 2.14
Date of test:2003-10-17 11:15:04
Running tests on:Darwin 7.0.0 Power Macintosh
Arguments:   
Comments:
Limits from: 
Server version:  MySQL 4.0.15aIMAX G5 debug log
Optimization:None
Hardware:PowerMac G5 2 x 2GHz , 4GB PC3200 RAM (8x256MB Kingston HyperX), 
160GB SATA HD

alter-table: Total time: 69 wallclock secs ( 0.01 usr  0.03 sys +  0.00 cusr  0.00 
csys =  0.04 CPU)
ATIS: Total time: 27 wallclock secs ( 9.31 usr  9.93 sys +  0.00 cusr  0.00 csys = 
19.24 CPU)
big-tables: Total time: 37 wallclock secs ( 8.10 usr 20.83 sys +  0.00 cusr  0.00 csys 
= 28.93 CPU)
connect: Total time: 182 wallclock secs (46.25 usr 54.08 sys +  0.00 cusr  0.00 csys = 
100.33 CPU)
create: Total time: 197 wallclock secs ( 5.71 usr  2.65 sys +  0.00 cusr  0.00 csys =  
8.36 CPU)
insert: Total time: 3270 wallclock secs (440.53 usr 361.82 sys +  0.00 cusr  0.00 csys 
= 802.35 CPU)
select: Total time: 222 wallclock secs (44.41 usr 32.55 sys +  0.00 cusr  0.00 csys = 
76.96 CPU)
transactions: Test skipped because the database doesn't support transactions
wisconsin: Total time: 28 wallclock secs ( 2.99 usr  2.73 sys +  0.00 cusr  0.00 csys 
=  5.72 CPU)

All 9 test executed successfully

Totals per operation:
Operation seconds usr sys cpu   tests
alter_table_add   19.000.000.000.00  20 
alter_table_drop  17.000.010.000.01  20 
connect   10.005.431.697.12   1 
connect+select_1_row  15.006.522.579.09   1 
connect+select_simple 12.006.181.878.05   1 
count 39.000.060.000.06 100 
count_distinct 1.000.190.030.221000 
count_distinct_big26.009.11   10.75   19.86 120 
count_distinct_group   2.001.060.601.661000 
count_distinct_group_on_key0.000.230.070.301000 
count_distinct_group_on_key_parts  2.001.070.621.691000 
count_distinct_key_prefix  0.000.140.100.241000 
count_group_on_key_parts   2.001.250.381.631000 
count_on_key  31.007.592.98   10.57   50100 
create+drop   57.001.570.712.28   1 
create_MANY_tables38.001.050.471.52   1 
create_index  16.000.000.000.00   8 
create_key+drop   61.001.920.562.48   1 
create_table   0.000.000.000.00  31 
delete_all_many_keys 171.000.000.010.01   1 
delete_big 0.000.000.000.00   1 
delete_big_many_keys 171.000.000.010.01 128 
delete_key 5.000.240.320.56   1 
delete_range  25.000.000.000.00  12 
drop_index16.000.000.000.00   8 
drop_table 1.000.000.000.00  28 
drop_table_when_MANY_tables   36.000.680.381.06   1 
insert   271.009.92   10.35   20.27  350768 
insert_duplicates 18.002.303.075.37  10 
insert_key   371.005.263.318.57  10 
insert_many_fields 7.000.290.040.332000 
insert_select_1_key7.000.000.000.00   1 
insert_select_2_keys  12.000.000.000.00   1 
min_max   12.000.000.010.01  60 
min_max_on_key22.00   12.484.77   17.25   85000 
multiple_value_insert 19.000.290.030.32  10 
order_by_big  67.00   26.36   33.24   59.60  10 
order_by_big_key  66.00   27.64   33.13   60.77  10 
order_by_big_key2 65.00   26.27   34.13   60.40  10 
order_by_big_key_desc 66.00   27.11   33.51   60.62  10 
order_by_big_key_diff 64.00   26.85   32.

Re: Benchmark Results or MySQL Articles Needed!

[Mark Wong]

On Wed, 2002-10-02 at 20:51, mos wrote:
> At 09:40 PM 10/2/2002, you wrote:
> >For science fair I'm doing a project on MySQL. For my introduction, I
> >need some articles, benchmarks, or anything else credible that I can put
> >in the bibliography. If anyone can help me out, I'd be extremely
> >appreciative! Thanks.
> 
> SS,
>Benchmarks are no longer believable because other database 
> manufacturers (not MySQL AB of course) can twist the tests so it appears 
> their database is faster than it is.  I did run into InnoDb benchmarks at 
> http://www.innodb.com/bench.html. Instead of benchmarks you might want to 
> report on who's using MySQL. See 
> http://www.mysql.com/press/user_stories/index.html and 
> http://www.innodb.com/userstories.html. Also the size of some of these 
> databases (1+Terabyte)  might impress people.
> 
> Mike

We're working on developing real-world database workloads at the Open
Source Development Lab (http://www.osdl.org/projects/performance/). 
While these are based on the TPC benchmarks, the intent is to evaluate
system performance as opposed to seeing which database is faster.

Our development here is currently focused on SAP DB, but this does not
mean we want it to exclusively work with that database.  We would love
(and appreciate) any help to adapt it to work with SAP DB.  Please
contact us if there is any interest.

Thanks,
Mark

-- 
Mark Wong - - [EMAIL PROTECTED]
Open Source Development Lab Inc - A non-profit corporation
15275 SW Koll Parkway - Suite H - Beaverton OR, 97006
(503)-626-2455 x 32 (office)
(503)-626-2436  (fax)
http://www.osdl.org/archive/markw/


-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Re: Benchmark Results or MySQL Articles Needed!

[mos]

At 09:40 PM 10/2/2002, you wrote:
>For science fair I'm doing a project on MySQL. For my introduction, I
>need some articles, benchmarks, or anything else credible that I can put
>in the bibliography. If anyone can help me out, I'd be extremely
>appreciative! Thanks.

SS,
   Benchmarks are no longer believable because other database 
manufacturers (not MySQL AB of course) can twist the tests so it appears 
their database is faster than it is.  I did run into InnoDb benchmarks at 
http://www.innodb.com/bench.html. Instead of benchmarks you might want to 
report on who's using MySQL. See 
http://www.mysql.com/press/user_stories/index.html and 
http://www.innodb.com/userstories.html. Also the size of some of these 
databases (1+Terabyte)  might impress people.

Mike


-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Re: benchmark to my MySQL

[Luuk de Boer]

if you have downloaded the src of mysql and installed it you can 
find the mysql benchmark in your /usr/local/sql-bench

Else you also have downloads seperately on the mysql site for only 
the benchmark (in rpm and others).

Grtz.

Luuk

On 4 Aug 2002, at 1:58, savaidis wrote:

> 
> Is there any standard method or benchmark to run localy or remotly to
> my host to help me to setup MySQL better or to check various
> computer/configurations?
> 
> 
> I could find anything on
> http://www.mysql.com/information/benchmarks.html links. Only at third
> part software but it is in C++, not .sql. What that means?
> 
> 
> Makis
> 
> 
> -
> Before posting, please check:
>http://www.mysql.com/manual.php   (the manual)
>http://lists.mysql.com/   (the list archive)
> 
> To request this thread, e-mail <[EMAIL PROTECTED]> To
> unsubscribe, e-mail <[EMAIL PROTECTED]>
> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
> 



-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Re: Benchmark

[Brent Baisley]

EWeek did a comparison review a few months back comparing MySQL, DB2, 
Oracle, MS SQL, and SyBase. Although it compared using a Java front end, 
so a lot of performance relied on JDBC support in the databases.
Anyway, here is the URL:
http://www.eweek.com/article2/0,3959,293,00.asp


On Tuesday, July 30, 2002, at 09:09 AM, Daniel Kiss wrote:

> Hi,
>
> Where can I find some benchmark test result documentation.
> First of all, I'm interested in comparing MySQL, Oracle and MSSQL speed 
> in different situations.
>
> Thanks,
>   Daniel
>
> -
> Before posting, please check:
>   http://www.mysql.com/manual.php   (the manual)
>   http://lists.mysql.com/   (the list archive)
>
> To request this thread, e-mail <[EMAIL PROTECTED]>
> To unsubscribe, e-mail  [EMAIL PROTECTED]>
> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
>
>
--
Brent Baisley
Systems Architect
Landover Associates, Inc.
Search & Advisory Services for Advanced Technology Environments
p: 212.759.6400/800.759.0577


-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Re: Benchmark

[Egor Egorov]

Daniel,
Tuesday, July 30, 2002, 4:09:00 PM, you wrote:

DK> Where can I find some benchmark test result documentation.
DK> First of all, I'm interested in comparing MySQL, Oracle and MSSQL speed in 
DK> different situations.

Check some links in the manual:
  http://www.mysql.com/doc/M/y/MySQL_Benchmarks.html
  http://www.mysql.com/information/benchmarks.html





-- 
For technical support contracts, goto https://order.mysql.com/?ref=ensita
This email is sponsored by Ensita.net http://www.ensita.net/
   __  ___ ___   __
  /  |/  /_ __/ __/ __ \/ /Egor Egorov
 / /|_/ / // /\ \/ /_/ / /__   [EMAIL PROTECTED]
/_/  /_/\_, /___/\___\_\___/   MySQL AB / Ensita.net
   <___/   www.mysql.com



-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

RE: Benchmark

[Salada, Duncan]

http://www.mysql.com/doc/M/y/MySQL_Benchmarks.html

Duncan
--
Duncan Salada | Titan | www.titan.com/testeval
Email: [EMAIL PROTECTED] | Voice: 301-925-3222x375 | Fax: 301-925-3216

> -Original Message-
> From: Kathy Reyes [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, March 19, 2002 1:51 PM
> To: mysql
> Subject: Benchmark
> 
> 
> Hello all its me again
> Does any one now of any database benchmarking utility that I 
> can use on
> mysql database???
> 
> 
> Kathy Reyes
> Live Technology  International
> 2975 Westchester Avenue
> Suite 103
> Purchase, NY 10577
> Tel: (914) 490-7096
> [EMAIL PROTECTED]
> www.LiveTechnology.com
> 
> 
> 
> 
> -
> Before posting, please check:
>http://www.mysql.com/manual.php   (the manual)
>http://lists.mysql.com/   (the list archive)
> 
> To request this thread, e-mail <[EMAIL PROTECTED]>
> To unsubscribe, e-mail 
> <[EMAIL PROTECTED]>
> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
> 

-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php

Re: Benchmark

[Christopher Thompson]

Have you looked at:
http://www.mysql.com/information/benchmarks.html

On Tuesday 19 March 2002 11:50 am, Kathy Reyes wrote:
> Hello all its me again
> Does any one now of any database benchmarking utility that I can use on
> mysql database???

-
Before posting, please check:
   http://www.mysql.com/manual.php   (the manual)
   http://lists.mysql.com/   (the list archive)

To request this thread, e-mail <[EMAIL PROTECTED]>
To unsubscribe, e-mail <[EMAIL PROTECTED]>
Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php