ASE 12.5 vs ASE 11.5 performance issues

Hi All,
Again my question is about performance of ASE 12.5 against 11.5.01
We have a huge production client-server application running on W2K server
with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
The database was placed on raw disk partitions (no RAID) and split between
different physical drives as follows: data, log, tempdb and master database
are located on different drives.
We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
separate tempdb cache and bind tempdb to it. Database was bind to default
data cache.

Then we want to switch to ASE 12.5
We take the same hardware configuration and decide to split database into
four physical devices. We create separate database devices for datasegment,
index segment for indexes, log segment and  tempdb. WE create database from
the scratch (using scripts, omitting index creation) and bcp all data into
database. Then we create indexes, triggers and stored procedures. We run
reorg rebuild for all tables in order to refresh statistics and tables. The
cache configuration is the same.

We expected that for some performance gain using parallel query processing,
worker processes etc. But we got TWICE SLOWDOWN in average performance and
in query execution!!!!

Can anybody give some advice what can we do else to rise performance?
Does creation of separate cache for syslogs can give some advantage?
Should we run update index statistics after executing  reorg rebuild?
How we can tune ASE 12.5 optimizer? May be he is based on different
algorithm?
How parallel processing can increase (decrease) performance?
Any other suggestions?

Regards

AlexM



-1
Alexander
4/10/2002 12:20:00 PM
sybase.ase.performance+tuning 2395 articles. 0 followers. Follow

28 Replies
1145 Views

Similar Articles

[PageSpeed] 55
Get it on Google Play
Get it on Apple App Store

Alex,
We would have to look at the configuration of ASE.  A show plan for one of
the queries thats running slower ( from both databases) and possiable
statisitcs IO and statistics time

Does you computers have more than one processer?  It is tough to do
parrallel processing without it.

By raw, are you going directly to the disks? or are you going to files on
the disk?


I would also be curious if you recreated the 11.5 database with bcps if it
would slow down..

David


"Alexander Mitchenko" <alex@alexm.carrier.kiev.ua> wrote in message
news:yBj6ZpI4BHA.206@forums.sybase.com...
> Hi All,
> Again my question is about performance of ASE 12.5 against 11.5.01
> We have a huge production client-server application running on W2K server
> with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
> The database was placed on raw disk partitions (no RAID) and split between
> different physical drives as follows: data, log, tempdb and master
database
> are located on different drives.
> We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> separate tempdb cache and bind tempdb to it. Database was bind to default
> data cache.
>
> Then we want to switch to ASE 12.5
> We take the same hardware configuration and decide to split database into
> four physical devices. We create separate database devices for
datasegment,
> index segment for indexes, log segment and  tempdb. WE create database
from
> the scratch (using scripts, omitting index creation) and bcp all data into
> database. Then we create indexes, triggers and stored procedures. We run
> reorg rebuild for all tables in order to refresh statistics and tables.
The
> cache configuration is the same.
>
> We expected that for some performance gain using parallel query
processing,
> worker processes etc. But we got TWICE SLOWDOWN in average performance and
> in query execution!!!!
>
> Can anybody give some advice what can we do else to rise performance?
> Does creation of separate cache for syslogs can give some advantage?
> Should we run update index statistics after executing  reorg rebuild?
> How we can tune ASE 12.5 optimizer? May be he is based on different
> algorithm?
> How parallel processing can increase (decrease) performance?
> Any other suggestions?
>
> Regards
>
> AlexM
>
>
>


1
David
4/10/2002 1:29:38 PM
"David Pascuzzi" <pascuzzi@sybase.com> wrote in message
news:TLD3EUJ4BHA.204@forums.sybase.com...
> Alex,
> We would have to look at the configuration of ASE.  A show plan for one of
> the queries thats running slower ( from both databases) and possiable
> statisitcs IO and statistics time

How do you want me to get them? Using monitor server?

>
> Does you computers have more than one processer?  It is tough to do
> parrallel processing without it.
>
I wrote about this - two PIII 1GHz

> By raw, are you going directly to the disks? or are you going to files on
> the disk?
>
I wrote about this - Raw disk partitions under W2K
>
> I would also be curious if you recreated the 11.5 database with bcps if it
> would slow down..
>

I wrote about this - We create database using scripts and bcp data into
database

> David
>
>
> "Alexander Mitchenko" <alex@alexm.carrier.kiev.ua> wrote in message
> news:yBj6ZpI4BHA.206@forums.sybase.com...
> > Hi All,
> > Again my question is about performance of ASE 12.5 against 11.5.01
> > We have a huge production client-server application running on W2K
server
> > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
procedures).
> > The database was placed on raw disk partitions (no RAID) and split
between
> > different physical drives as follows: data, log, tempdb and master
> database
> > are located on different drives.
> > We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> > separate tempdb cache and bind tempdb to it. Database was bind to
default
> > data cache.
> >
> > Then we want to switch to ASE 12.5
> > We take the same hardware configuration and decide to split database
into
> > four physical devices. We create separate database devices for
> datasegment,
> > index segment for indexes, log segment and  tempdb. WE create database
> from
> > the scratch (using scripts, omitting index creation) and bcp all data
into
> > database. Then we create indexes, triggers and stored procedures. We run
> > reorg rebuild for all tables in order to refresh statistics and tables.
> The
> > cache configuration is the same.
> >
> > We expected that for some performance gain using parallel query
> processing,
> > worker processes etc. But we got TWICE SLOWDOWN in average performance
and
> > in query execution!!!!
> >
> > Can anybody give some advice what can we do else to rise performance?
> > Does creation of separate cache for syslogs can give some advantage?
> > Should we run update index statistics after executing  reorg rebuild?
> > How we can tune ASE 12.5 optimizer? May be he is based on different
> > algorithm?
> > How parallel processing can increase (decrease) performance?
> > Any other suggestions?
> >
> > Regards
> >
> > AlexM
> >
> >
> >
>
>


1
Alexander
4/10/2002 2:49:09 PM
alex@alexm.carrier.kiev.ua wrote...
> Hi All,
> Again my question is about performance of ASE 12.5 against 11.5.01
> We have a huge production client-server application running on W2K server
> with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
> The database was placed on raw disk partitions (no RAID) and split between
> different physical drives as follows: data, log, tempdb and master database
> are located on different drives.

It is very unusual to run with raw partitions on Win2K.  From what I hear, the performance 
gain is minimal and not worth the hassle.

> We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> separate tempdb cache and bind tempdb to it. Database was bind to default
> data cache.

What is the default data cache set to?  The default is 8MB and you must explicitly increase 
it to utilize the max memory.
 
> Then we want to switch to ASE 12.5
> We take the same hardware configuration and decide to split database into
> four physical devices. We create separate database devices for datasegment,
> index segment for indexes, log segment and  tempdb. WE create database from
> the scratch (using scripts, omitting index creation) and bcp all data into
> database. Then we create indexes, triggers and stored procedures. We run
> reorg rebuild for all tables in order to refresh statistics and tables. 

Running REORG REBUILD was probably a complete waste of time.  Everything was in pristine 
allocations after the BCP and the creation of the indexes.

> The cache configuration is the same.
> 
> We expected that for some performance gain using parallel query processing,
> worker processes etc. But we got TWICE SLOWDOWN in average performance and
> in query execution!!!!
> 
> Can anybody give some advice what can we do else to rise performance?

Hard to say.  What is slow?  Inserts? Updates?  

> Does creation of separate cache for syslogs can give some advantage?

Yes, but not for queries.

> Should we run update index statistics after executing  reorg rebuild?

You can do this.  It will provide additional statistics information on columns that are not 
the leading column of each index.  If you don't have any composite indexes or primary keys 
then it would not help at all.

> How we can tune ASE 12.5 optimizer? May be he is based on different
> algorithm?
> How parallel processing can increase (decrease) performance?
> Any other suggestions?

It sounds like the server is not configured properly.  I migrated a client last month from 
11.0.3 to 12.5 and with a little bit of tuning they saw a fantastic performance 
improvement.

When you say you have a "huge" application, how "huge" is "huge"?  Maybe you need to add 
more memory to the server.  If the database is 1 GB then you probably have enough RAM.  If 
the database is 50 GB then more memory is in order.  Memory is cheap.  Geez, I run 512 MB 
on my laptop.  I would put at least 2GB in a server.  You could probably squeeze another 
100 MB of RAM out of the O/S for ASE.

Any reason why you didn't do a dump from the 11.5 server to the 12.5 server?  This method 
has some benefits.
-- 
Jim Egan [TeamSybase]
Senior Consultant
Sybase Professional Services
0
Jim
4/10/2002 3:21:32 PM
"Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
news:MPG.171e066a963386c098bcf6@forums.sybase.com...
> alex@alexm.carrier.kiev.ua wrote...
> It is very unusual to run with raw partitions on Win2K.  From what I hear,
the performance
> gain is minimal and not worth the hassle.
>

We did this for a long time and it is faster than on files for sure -
maintenance experienc

> > We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> > separate tempdb cache and bind tempdb to it. Database was bind to
default
> > data cache.
>
> What is the default data cache set to?  The default is 8MB and you must
explicitly increase
> it to utilize the max memory.

Default data cache size - 400 MB, tempdb cache - 100MB

> Running REORG REBUILD was probably a complete waste of time.  Everything
was in pristine
> allocations after the BCP and the creation of the indexes.

I mean not only during import, but in the future while maintaining database
weekly

> Hard to say.  What is slow?  Inserts? Updates?
>
For sure - queries

> > Does creation of separate cache for syslogs can give some advantage?
>
> Yes, but not for queries.
i understand

>
> > Should we run update index statistics after executing  reorg rebuild?
>
> You can do this.  It will provide additional statistics information on
columns that are not
> the leading column of each index.  If you don't have any composite indexes
or primary keys
> then it would not help at all.
We have a lot
>
> > How we can tune ASE 12.5 optimizer? May be he is based on different
> > algorithm?
> > How parallel processing can increase (decrease) performance?
> > Any other suggestions?
>
> It sounds like the server is not configured properly.  I migrated a client
last month from
> 11.0.3 to 12.5 and with a little bit of tuning they saw a fantastic
performance
> improvement.

What do we miss - i'll atach config file. Or may be the database contains
less tables?

> When you say you have a "huge" application, how "huge" is "huge"?  Maybe
you need to add
> more memory to the server.  If the database is 1 GB then you probably have
enough RAM.  If
> the database is 50 GB then more memory is in order.  Memory is cheap.
Geez, I run 512 MB
> on my laptop.  I would put at least 2GB in a server.  You could probably
squeeze another
> 100 MB of RAM out of the O/S for ASE.

I mean 400 tables, 2000 stored procedures and rather complicated analitical
queries

> Any reason why you didn't do a dump from the 11.5 server to the 12.5
server?  This method
> has some benefits.
I was not sure we can

AlexM


begin 666 FILESERVER.cfg
M(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C
M(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C#0HC#0HC"0E#;VYF
M:6=U<F%T:6]N($9I;&4@9F]R('1H92!3>6)A<V4@4U%,(%-E<G9E<@T*(PT*
M(PD)4&QE87-E(')E860@=&AE(%-Y<W1E;2!!9&UI;FES=')A=&EO;B!'=6ED
M92 H4T%'*0T*(PD)8F5F;W)E(&-H86YG:6YG(&%N>2!O9B!T:&4@=F%L=65S
M(&EN('1H:7,@9FEL92X-"B,-"B,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C
M(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C(R,C
M(R,C(R,C(PT*#0H-"@T*6T-O;F9I9W5R871I;VX@3W!T:6]N<UT-"@T*6T=E
M;F5R86P@26YF;W)M871I;VY=#0H-"EM"86-K=7 O4F5C;W9E<GE=#0H)<F5C
M;W9E<GD@:6YT97)V86P@:6X@;6EN=71E<R ]($1%1D%53%0-"@EP<FEN="!R
M96-O=F5R>2!I;F9O<FUA=&EO;B ]($1%1D%53%0-"@ET87!E(')E=&5N=&EO
M;B!I;B!D87ES(#T@1$5&055,5 T*#0I;0V%C:&4@36%N86=E<ET-"@EN=6UB
M97(@;V8@;V%M('1R:7!S(#T@1$5&055,5 T*"6YU;6)E<B!O9B!I;F1E>"!T
M<FEP<R ]($1%1D%53%0-"@EM96UO<GD@86QI9VYM96YT(&)O=6YD87)Y(#T@
M1$5&055,5 T*"6=L;V)A;"!A<WEN8R!P<F5F971C:"!L:6UI=" ]($1%1D%5
M3%0-"@EG;&]B86P@8V%C:&4@<&%R=&ET:6]N(&YU;6)E<B ](#(-"@T*6TYA
M;65D($-A8VAE.F1E9F%U;'0@9&%T82!C86-H95T-"@EC86-H92!S:7IE(#T@
M-# P,# P2PT*"6-A8VAE('-T871U<R ](&1E9F%U;'0@9&%T82!C86-H90T*
M"6-A8VAE(')E<&QA8V5M96YT('!O;&EC>2 ]($1%1D%53%0-"@EL;V-A;"!C
M86-H92!P87)T:71I;VX@;G5M8F5R(#T@1$5&055,5 T*#0I;3F%M960@0V%C
M:&4Z=&5M<&1B7V-A8VAE70T*"6-A8VAE('-I>F4@/2 S,# P,#!+#0H)8V%C
M:&4@<W1A='5S(#T@;6EX960@8V%C:&4-"@EC86-H92!R97!L86-E;65N="!P
M;VQI8WD@/2!$149!54Q4#0H);&]C86P@8V%C:&4@<&%R=&ET:6]N(&YU;6)E
M<B ]($1%1D%53%0-"@T*6TUE=&$M1&%T82!#86-H97-=#0H);G5M8F5R(&]F
M(&]P96X@9&%T86)A<V5S(#T@1$5&055,5 T*"6YU;6)E<B!O9B!O<&5N(&]B
M:F5C=',@/2 Q,# P#0H);W!E;B!O8FIE8W0@<W!I;FQO8VL@<F%T:6\@/2!$
M149!54Q4#0H);G5M8F5R(&]F(&]P96X@:6YD97AE<R ](#$P,# -"@EO<&5N
M(&EN9&5X(&AA<V@@<W!I;FQO8VL@<F%T:6\@/2!$149!54Q4#0H);W!E;B!I
M;F1E>"!S<&EN;&]C:R!R871I;R ]($1%1D%53%0-"@EP87)T:71I;VX@9W)O
M=7!S(#T@1$5&055,5 T*"7!A<G1I=&EO;B!S<&EN;&]C:R!R871I;R ]($1%
M1D%53%0-"@T*6T1I<VL@22]/70T*"61I<VL@:2]O('-T<G5C='5R97,@/2!$
M149!54Q4#0H);G5M8F5R(&]F(&QA<F=E(&DO;R!B=69F97)S(#T@1$5&055,
M5 T*"7!A9V4@=71I;&EZ871I;VX@<&5R8V5N=" ]($1%1D%53%0-"@EN=6UB
M97(@;V8@9&5V:6-E<R ]($1%1D%53%0-"@ED:7-A8FQE(&1I<VL@;6ER<F]R
M:6YG(#T@1$5&055,5 T*"6%L;&]W('-Q;"!S97)V97(@87-Y;F,@:2]O(#T@
M1$5&055,5 T*#0I;3&%N9W5A9V5S70T*"61I<V%B;&4@8VAA<F%C=&5R('-E
M="!C;VYV97)S:6]N<R ]($1%1D%53%0-"@T*6U5N:6-O9&5=#0H)96YA8FQE
M('5N:6-O9&4@;F]R;6%L:7IA=&EO;B ]($1%1D%53%0-"@EE;F%B;&4@<W5R
M<F]G871E('!R;V-E<W-I;F<@/2!$149!54Q4#0H)96YA8FQE('5N:6-O9&4@
M8V]N=F5R<VEO;G,@/2!$149!54Q4#0H)<VEZ92!O9B!U;FEL:6(@8V%C:&4@
M/2!$149!54Q4#0H-"EM.971W;W)K($-O;6UU;FEC871I;VY=#0H)9&5F875L
M="!N971W;W)K('!A8VME="!S:7IE(#T@1$5&055,5 T*"6UA>"!N971W;W)K
M('!A8VME="!S:7IE(#T@1$5&055,5 T*"7)E;6]T92!S97)V97(@<')E+7)E
M860@<&%C:V5T<R ]($1%1D%53%0-"@EN=6UB97(@;V8@<F5M;W1E(&-O;FYE
M8W1I;VYS(#T@1$5&055,5 T*"6YU;6)E<B!O9B!R96UO=&4@;&]G:6YS(#T@
M1$5&055,5 T*"6YU;6)E<B!O9B!R96UO=&4@<VET97,@/2!$149!54Q4#0H)
M;6%X(&YU;6)E<B!N971W;W)K(&QI<W1E;F5R<R ]($1%1D%53%0-"@ET8W @
M;F\@9&5L87D@/2!$149!54Q4#0H)86QL;W<@<V5N9&US9R ]($1%1D%53%0-
M"@ES>6)?<V5N9&US9R!P;W)T(&YU;6)E<B ]($1%1D%53%0-"@EA;&QO=R!R
M96UO=&4@86-C97-S(#T@1$5&055,5 T*#0I;3R]3(%)E<V]U<F-E<UT-"@EM
M87@@87-Y;F,@:2]O<R!P97(@96YG:6YE(#T@1$5&055,5 T*"6UA>"!A<WEN
M8R!I+V]S('!E<B!S97)V97(@/2!$149!54Q4#0H-"EM087)A;&QE;"!1=65R
M>5T-"@EN=6UB97(@;V8@=V]R:V5R('!R;V-E<W-E<R ](#$V#0H);65M;W)Y
M('!E<B!W;W)K97(@<')O8V5S<R ](#,R,# P#0H);6%X('!A<F%L;&5L(&1E
M9W)E92 ](#@-"@EM87@@<V-A;B!P87)A;&QE;"!D96=R964@/2 T#0H-"EM0
M:'ES:6-A;"!297-O=7)C97-=#0H-"EM0:'ES:6-A;"!-96UO<GE=#0H);6%X
M(&UE;6]R>2 ](#0X,# P, T*"6%D9&ET:6]N86P@;F5T=V]R:R!M96UO<GD@
M/2!$149!54Q4#0H)<VAA<F5D(&UE;6]R>2!S=&%R=&EN9R!A9&1R97-S(#T@
M1$5&055,5 T*"6%L;&]C871E(&UA>"!S:&%R960@;65M;W)Y(#T@1$5&055,
M5 T*"61Y;F%M:6,@86QL;V-A=&EO;B!O;B!D96UA;F0@/2!$149!54Q4#0H)
M;&]C:R!S:&%R960@;65M;W)Y(#T@1$5&055,5 T*"6AE87 @;65M;W)Y('!E
M<B!U<V5R(#T@1$5&055,5 T*#0I;4')O8V5S<V]R<UT-"@EM87@@;VYL:6YE
M(&5N9VEN97,@/2 R#0H);G5M8F5R(&]F(&5N9VEN97,@870@<W1A<G1U<" ]
M(#(-"@T*6U-13"!397)V97(@061M:6YI<W1R871I;VY=#0H)<')O8V5D=7)E
M(&-A8VAE('-I>F4@/2 V-# P, T*"61E9F%U;'0@9&%T86)A<V4@<VEZ92 ]
M($1%1D%53%0-"@EI9&5N=&ET>2!B=7)N:6YG('-E="!F86-T;W(@/2 Q#0H)
M86QL;W<@;F5S=&5D('1R:6=G97)S(#T@1$5&055,5 T*"6%L;&]W('5P9&%T
M97,@=&\@<WES=&5M('1A8FQE<R ]($1%1D%53%0-"@ED969A=6QT(&9I;&P@
M9F%C=&]R('!E<F-E;G0@/2!$149!54Q4#0H)9&5F875L="!E>'!?<F]W7W-I
M>F4@<&5R8V5N=" ]($1%1D%53%0-"@EN=6UB97(@;V8@;6%I;&)O>&5S(#T@
M1$5&055,5 T*"6YU;6)E<B!O9B!M97-S86=E<R ]($1%1D%53%0-"@EN=6UB
M97(@;V8@86QA<FUS(#T@1$5&055,5 T*"6YU;6)E<B!O9B!P<F4M86QL;V-A
M=&5D(&5X=&5N=',@/2!$149!54Q4#0H)979E;G0@8G5F9F5R<R!P97(@96YG
M:6YE(#T@1$5&055,5 T*"6-P=2!A8V-O=6YT:6YG(&9L=7-H(&EN=&5R=F%L
M(#T@1$5&055,5 T*"6DO;R!A8V-O=6YT:6YG(&9L=7-H(&EN=&5R=F%L(#T@
M1$5&055,5 T*"7-Q;"!S97)V97(@8VQO8VL@=&EC:R!L96YG=&@@/2!$149!
M54Q4#0H)<G5N;F%B;&4@<')O8V5S<R!S96%R8V@@8V]U;G0@/2!$149!54Q4
M#0H):2]O('!O;&QI;F<@<')O8V5S<R!C;W5N=" ]($1%1D%53%0-"@ET:6UE
M('-L:6-E(#T@1$5&055,5 T*"6-P=2!G<F%C92!T:6UE(#T@1$5&055,5 T*
M"6YU;6)E<B!O9B!S;W)T(&)U9F9E<G,@/2 Q,# P#0H)<VEZ92!O9B!A=71O
M(&ED96YT:71Y(&-O;'5M;B ]($1%1D%53%0-"@EI9&5N=&ET>2!G<F%B('-I
M>F4@/2!$149!54Q4#0H):&]U<V5K965P97(@9G)E92!W<FET92!P97)C96YT
M(#T@1$5&055,5 T*"65N86)L92!H;W5S96ME97!E<B!'0R ]($1%1D%53%0-
M"@EA;&QO=R!R97-O=7)C92!L:6UI=',@/2!$149!54Q4#0H);G5M8F5R(&]F
M(&%U>"!S8V%N(&1E<V-R:7!T;W)S(#T@1$5&055,5 T*"5-13"!097)F;6]N
M($EN=&5G<F%T:6]N(#T@1$5&055,5 T*"6%L;&]W(&)A8VMW87)D('-C86YS
M(#T@1$5&055,5 T*"6QI8V5N<V4@:6YF;W)M871I;VX@/2!$149!54Q4#0H)
M96YA8FQE('-O<G0M;65R9V4@:F]I;B!A;F0@2E1#(#T@1$5&055,5 T*"6%B
M<W1R86-T('!L86X@;&]A9" ]($1%1D%53%0-"@EA8G-T<F%C="!P;&%N(&1U
M;7 @/2!$149!54Q4#0H)86)S=')A8W0@<&QA;B!R97!L86-E(#T@1$5&055,
M5 T*"6%B<W1R86-T('!L86X@8V%C:&4@/2!$149!54Q4#0H)=&5X="!P<F5F
M971C:"!S:7IE(#T@1$5&055,5 T*"65N86)L92!(02 ]($1%1D%53%0-"@T*
M6U5S97(@16YV:7)O;FUE;G1=#0H);G5M8F5R(&]F('5S97(@8V]N;F5C=&EO
M;G,@/2!$149!54Q4#0H)<W1A8VL@<VEZ92 ]($1%1D%53%0-"@ES=&%C:R!G
M=6%R9"!S:7IE(#T@1$5&055,5 T*"7!E<FUI<W-I;VX@8V%C:&4@96YT<FEE
M<R ]($1%1D%53%0-"@EU<V5R(&QO9R!C86-H92!S:7IE(#T@1$5&055,5 T*
M"75S97(@;&]G(&-A8VAE('-P:6YL;V-K(')A=&EO(#T@1$5&055,5 T*#0I;
M3&]C:R!-86YA9V5R70T*"6YU;6)E<B!O9B!L;V-K<R ](#8P,# -"@ED96%D
M;&]C:R!C:&5C:VEN9R!P97)I;V0@/2!$149!54Q4#0H);&]C:R!S<&EN;&]C
M:R!R871I;R ]($1%1D%53%0-"@EL;V-K(&%D9')E<W,@<W!I;FQO8VL@<F%T
M:6\@/2!$149!54Q4#0H);&]C:R!T86)L92!S<&EN;&]C:R!R871I;R ]($1%
M1D%53%0-"@EL;V-K(&AA<VAT86)L92!S:7IE(#T@1$5&055,5 T*"6QO8VL@
M<V-H96UE(#T@1$5&055,5 T*"6QO8VL@=V%I="!P97)I;V0@/2!$149!54Q4
M#0H)<F5A9"!C;VUM:71T960@=VET:"!L;V-K(#T@1$5&055,5 T*"7!R:6YT
M(&1E861L;V-K(&EN9F]R;6%T:6]N(#T@1$5&055,5 T*"61E861L;V-K(')E
M=')I97,@/2!$149!54Q4#0H)<&%G92!L;V-K('!R;VUO=&EO;B!(5TT@/2!$
M149!54Q4#0H)<&%G92!L;V-K('!R;VUO=&EO;B!,5TT@/2!$149!54Q4#0H)
M<&%G92!L;V-K('!R;VUO=&EO;B!00U0@/2!$149!54Q4#0H)<F]W(&QO8VL@
M<')O;6]T:6]N($A732 ]($1%1D%53%0-"@ER;W<@;&]C:R!P<F]M;W1I;VX@
M3%=-(#T@1$5&055,5 T*"7)O=R!L;V-K('!R;VUO=&EO;B!00U0@/2!$149!
M54Q4#0H-"EM396-U<FET>2!296QA=&5D70T*"7-Y<W1E;7=I9&4@<&%S<W=O
M<F0@97AP:7)A=&EO;B ]($1%1D%53%0-"@EA=61I="!Q=65U92!S:7IE(#T@
M1$5&055,5 T*"6-U<G)E860@8VAA;F=E('<O(&]P96X@8W5R<V]R<R ]($1%
M1D%53%0-"@EA;&QO=R!P<F]C961U<F4@9W)O=7!I;F<@/2!$149!54Q4#0H)
M<V5L96-T(&]N('-Y<V-O;6UE;G1S+G1E>'0@/2!$149!54Q4#0H)875D:71I
M;F<@/2!$149!54Q4#0H)8W5R<F5N="!A=61I="!T86)L92 ]($1%1D%53%0-
M"@ES=7-P96YD(&%U9&ET('=H96X@9&5V:6-E(&9U;&P@/2!$149!54Q4#0H)
M96YA8FQE(')O=R!L979E;"!A8V-E<W,@/2!$149!54Q4#0H)8VAE8VL@<&%S
M<W=O<F0@9F]R(&1I9VET(#T@1$5&055,5 T*"6UI;FEM=6T@<&%S<W=O<F0@
M;&5N9W1H(#T@1$5&055,5 T*"6UA>&EM=6T@9F%I;&5D(&QO9VEN<R ]($1%
M1D%53%0-"@EE;F%B;&4@<W-L(#T@1$5&055,5 T*"75N:69I960@;&]G:6X@
M<F5Q=6ER960@/2!$149!54Q4#0H)=7-E('-E8W5R:71Y('-E<G9I8V5S(#T@
M1$5&055,5 T*"6US9R!C;VYF:61E;G1I86QI='D@<F5Q9" ]($1%1D%53%0-
M"@EM<V<@:6YT96=R:71Y(')E<60@/2!$149!54Q4#0H)<V5C=7)E(&1E9F%U
M;'0@;&]G:6X@/2!$149!54Q4#0H-"EM%>'1E;F1E9"!3=&]R960@4')O8V5D
M=7)E70T*"65S<"!U;FQO860@9&QL(#T@1$5&055,5 T*"65S<"!E>&5C=71I
M;VX@<')I;W)I='D@/2!$149!54Q4#0H)97-P(&5X96-U=&EO;B!S=&%C:W-I
M>F4@/2!$149!54Q4#0H)>'!?8VUD<VAE;&P@8V]N=&5X=" ]($1%1D%53%0-
M"@ES=&%R="!M86EL('-E<W-I;VX@/2!$149!54Q4#0H-"EM%<G)O<B!,;V==
M#0H)979E;G0@;&]G9VEN9R ]($1%1D%53%0-"@EL;V<@875D:70@;&]G;VX@
M<W5C8V5S<R ]($1%1D%53%0-"@EL;V<@875D:70@;&]G;VX@9F%I;'5R92 ]
M($1%1D%53%0-"@EE=F5N="!L;V<@8V]M<'5T97(@;F%M92 ]($1%1D%53%0-
M"@T*6U)E<"!!9V5N="!4:')E860@061M:6YI<W1R871I;VY=#0H)96YA8FQE
M(')E<"!A9V5N="!T:')E861S(#T@1$5&055,5 T*#0I;0V]M<&]N96YT($EN
M=&5G<F%T:6]N(%-E<G9I8V5S70T*"65N86)L92!C:7,@/2!$149!54Q4#0H)
M8VES(&-O;FYE8W0@=&EM96]U=" ]($1%1D%53%0-"@EC:7,@8G5L:R!I;G-E
M<G0@8F%T8V@@<VEZ92 ]($1%1D%53%0-"@EM87@@8VES(')E;6]T92!C;VYN
M96-T:6]N<R ]($1%1D%53%0-"@EC:7,@<&%C:V5T('-I>F4@/2!$149!54Q4
M#0H)8VES(&-U<G-O<B!R;W=S(#T@1$5&055,5 T*"65N86)L92!F:6QE(&%C
M8V5S<R ]($1%1D%53%0-"@EC:7,@8G5L:R!I;G-E<G0@87)R87D@<VEZ92 ]
M($1%1D%53%0-"@EE;F%B;&4@9G5L;"UT97AT('-E87)C:" ]($1%1D%53%0-
M"@EC:7,@<G!C(&AA;F1L:6YG(#T@1$5&055,5 T*#0I;2F%V82!397)V:6-E
M<UT-"@EE;F%B;&4@:F%V82 ]($1%1D%53%0-"@ES:7IE(&]F('!R;V-E<W,@
M;V)J96-T(&AE87 @/2!$149!54Q4#0H)<VEZ92!O9B!S:&%R960@8VQA<W,@
M:&5A<" ]($1%1D%53%0-"@ES:7IE(&]F(&=L;V)A;"!F:7AE9"!H96%P(#T@
M1$5&055,5 T*"6YU;6)E<B!O9B!J879A('-O8VME=',@/2!$149!54Q4#0H)
M96YA8FQE(&5N=&5R<')I<V4@:F%V82!B96%N<R ]($1%1D%53%0-"@T*6T14
M32!!9&UI;FES=')A=&EO;ET-"@EE;F%B;&4@1%1-(#T@1$5&055,5 T*"65N
M86)L92!X86-T(&-O;W)D:6YA=&EO;B ]($1%1D%53%0-"@EX86-T(&-O;W)D
M:6YA=&EO;B!I;G1E<G9A;" ]($1%1D%53%0-"@EN=6UB97(@;V8@9'1X('!A
M<G1I8VEP86YT<R ]($1%1D%53%0-"@ES=')I8W0@9'1M(&5N9F]R8V5M96YT
M(#T@1$5&055,5 T*"71X;B!T;R!P<W,@<F%T:6\@/2!$149!54Q4#0H)9'1M
M(&QO8VL@=&EM96]U="!P97)I;V0@/2!$149!54Q4#0H)9'1M(&1E=&%C:"!T
M:6UE;W5T('!E<FEO9" ]($1%1D%53%0-"@T*6T1I86=N;W-T:6-S70T*"61U
M;7 @;VX@8V]N9&ET:6]N<R ]($1%1D%53%0-"@EM87AI;75M(&1U;7 @8V]N
M9&ET:6]N<R ]($1%1D%53%0-"@EN=6UB97(@;V8@8V-B<R ]($1%1D%53%0-
M"@EC87!S('!E<B!C8V(@/2!$149!54Q4#0H)879E<F%G92!C87 @<VEZ92 ]
M($1%1D%53%0-"@T*6TUO;FET;W)I;F==#0H)96YA8FQE(&UO;FET;W)I;F<@
M/2!$149!54Q4#0H)<W%L('1E>'0@<&EP92!A8W1I=F4@/2!$149!54Q4#0H)
M<W%L('1E>'0@<&EP92!M87@@;65S<V%G97,@/2!$149!54Q4#0H)<&QA;B!T
M97AT('!I<&4@86-T:79E(#T@1$5&055,5 T*"7!L86X@=&5X="!P:7!E(&UA
M>"!M97-S86=E<R ]($1%1D%53%0-"@ES=&%T96UE;G0@<&EP92!A8W1I=F4@
M/2!$149!54Q4#0H)<W1A=&5M96YT('!I<&4@;6%X(&UE<W-A9V5S(#T@1$5&
M055,5 T*"65R<F]R;&]G('!I<&4@86-T:79E(#T@1$5&055,5 T*"65R<F]R
M;&]G('!I<&4@;6%X(&UE<W-A9V5S(#T@1$5&055,5 T*"61E861L;V-K('!I
M<&4@86-T:79E(#T@1$5&055,5 T*"61E861L;V-K('!I<&4@;6%X(&UE<W-A
M9V5S(#T@1$5&055,5 T*"7=A:70@979E;G0@=&EM:6YG(#T@1$5&055,5 T*
M"7!R;V-E<W,@=V%I="!E=F5N=',@/2!$149!54Q4#0H);V)J96-T(&QO8VMW
M86ET('1I;6EN9R ]($1%1D%53%0-"@E344P@8F%T8V@@8V%P='5R92 ]($1%
M1D%53%0-"@ES=&%T96UE;G0@<W1A=&ES=&EC<R!A8W1I=F4@/2!$149!54Q4
M#0H)<&5R(&]B:F5C="!S=&%T:7-T:6-S(&%C=&EV92 ]($1%1D%53%0-"@EM
A87@@4U%,('1E>'0@;6]N:71O<F5D(#T@1$5&055,5 T*
`
end

0
Alexander
4/10/2002 3:49:30 PM
"Alexander Mitchenko" <alex@alexm.carrier.kiev.ua> wrote in message
news:lVY9w8J4BHA.133@forums.sybase.com...
>
> "David Pascuzzi" <pascuzzi@sybase.com> wrote in message
> news:TLD3EUJ4BHA.204@forums.sybase.com...
> > Alex,
> > We would have to look at the configuration of ASE.  A show plan for one
of
> > the queries thats running slower ( from both databases) and possiable
> > statisitcs IO and statistics time
>
> How do you want me to get them? Using monitor server?
grab one of the queries that is running slow and then run it with isql to
see the behavior

Run one or more of the stored procedures and measure the behavior of it( I
am assuming that  with 2000 stored procs).. Do this twice for the statistics
IO and time so we know the state of the cache.

as far as the configureation get the .cfg file for both servers
>
> >
> > Does you computers have more than one processer?  It is tough to do
> > parrallel processing without it.
> >
> I wrote about this - two PIII 1GHz
sorry, I had took that to mean you have two different machines that are
configured the same.
>
> > By raw, are you going directly to the disks? or are you going to files
on
> > the disk?
> >
> I wrote about this - Raw disk partitions under W2K
But your disk init is for a  file... You probably are not going directly to
the devices..

Check the disks with a OS disk defragement utiltiy,(A defragmented disk may
or may not run better than a really fragemented one)

Are you .dat files on their own disks? or are they sharing disks?  if the
stored proc is doing a lot of writing is one server configured to let the OS
buffer the write and the other server is not?

> >
> > I would also be curious if you recreated the 11.5 database with bcps if
it
> > would slow down..
> >
>
> I wrote about this - We create database using scripts and bcp data into
> database
So, for this test you dropped and  recreated the 11.5 database  and
recreated its devices before running the test to compare performance.  That
makes actuall data placement less likely part of the problem.

A few other things.
   what network protcols are you using?
   are you client and server running on the same machine? or do you have a
different machine for the client?
   What are you using ct-lib/db-lib/odbc/jdbc/ to connect?
   How are you currently measuring that things are running slowwer?
  with the client application connected do a select * from
master..sysprocesses before and after the test for both machines ( I only
need this for the spid(s) that are used in the test)?
  Are you using rowlevel locking or page level locking?


David
>
> > David
> >
> >
> > "Alexander Mitchenko" <alex@alexm.carrier.kiev.ua> wrote in message
> > news:yBj6ZpI4BHA.206@forums.sybase.com...
> > > Hi All,
> > > Again my question is about performance of ASE 12.5 against 11.5.01
> > > We have a huge production client-server application running on W2K
> server
> > > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
> procedures).
> > > The database was placed on raw disk partitions (no RAID) and split
> between
> > > different physical drives as follows: data, log, tempdb and master
> > database
> > > are located on different drives.
> > > We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We
create
> > > separate tempdb cache and bind tempdb to it. Database was bind to
> default
> > > data cache.
> > >
> > > Then we want to switch to ASE 12.5
> > > We take the same hardware configuration and decide to split database
> into
> > > four physical devices. We create separate database devices for
> > datasegment,
> > > index segment for indexes, log segment and  tempdb. WE create database
> > from
> > > the scratch (using scripts, omitting index creation) and bcp all data
> into
> > > database. Then we create indexes, triggers and stored procedures. We
run
> > > reorg rebuild for all tables in order to refresh statistics and
tables.
> > The
> > > cache configuration is the same.
> > >
> > > We expected that for some performance gain using parallel query
> > processing,
> > > worker processes etc. But we got TWICE SLOWDOWN in average performance
> and
> > > in query execution!!!!
> > >
> > > Can anybody give some advice what can we do else to rise performance?
> > > Does creation of separate cache for syslogs can give some advantage?
> > > Should we run update index statistics after executing  reorg rebuild?
> > > How we can tune ASE 12.5 optimizer? May be he is based on different
> > > algorithm?
> > > How parallel processing can increase (decrease) performance?
> > > Any other suggestions?
> > >
> > > Regards
> > >
> > > AlexM
> > >
> > >
> > >
> >
> >
>
>


0
David
4/10/2002 3:54:39 PM
"David Pascuzzi" <pascuzzi@sybase.com> wrote in message
news:8dRAHlK4BHA.284@forums.sybase.com...
>
> grab one of the queries that is running slow and then run it with isql to
> see the behavior
>
> Run one or more of the stored procedures and measure the behavior of it( I
> am assuming that  with 2000 stored procs).. Do this twice for the
statistics
> IO and time so we know the state of the cache.
>
Do thi many times i think shopplan export is too lagre for this conference

> as far as the configureation get the .cfg file for both servers
already in my previous reply

> But your disk init is for a  file... You probably are not going directly
to
> the devices..

No we create database devices on raw disk partitions - this can be done
under Win NT
we work in such a way last four years
Just enter partition letter e.g. \\.\D: and enter partition size - you will
got it

>
> Check the disks with a OS disk defragement utiltiy,(A defragmented disk
may
> or may not run better than a really fragemented one)
>
> Are you .dat files on their own disks? or are they sharing disks?  if the
> stored proc is doing a lot of writing is one server configured to let the
OS
> buffer the write and the other server is not?
>
See before

> So, for this test you dropped and  recreated the 11.5 database  and
> recreated its devices before running the test to compare performance.
That
> makes actuall data placement less likely part of the problem.
>
> A few other things.
>    what network protcols are you using?
TCP/IP
>    are you client and server running on the same machine? or do you have a
> different machine for the client?
Does not make a difference

>    What are you using ct-lib/db-lib/odbc/jdbc/ to connect?
ct-lib

>    How are you currently measuring that things are running slowwer?
angry users running besides of me :)))

>   with the client application connected do a select * from
> master..sysprocesses before and after the test for both machines ( I only
> need this for the spid(s) that are used in the test)?
Why do you need them? e.g. spid = 1

>   Are you using rowlevel locking or page level locking?
We use datapages locking scheme

Regards

AlexM

>
>
> David
> >
> > > David
> > >
> > >
> > > "Alexander Mitchenko" <alex@alexm.carrier.kiev.ua> wrote in message
> > > news:yBj6ZpI4BHA.206@forums.sybase.com...
> > > > Hi All,
> > > > Again my question is about performance of ASE 12.5 against 11.5.01
> > > > We have a huge production client-server application running on W2K
> > server
> > > > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
> > procedures).
> > > > The database was placed on raw disk partitions (no RAID) and split
> > between
> > > > different physical drives as follows: data, log, tempdb and master
> > > database
> > > > are located on different drives.
> > > > We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We
> create
> > > > separate tempdb cache and bind tempdb to it. Database was bind to
> > default
> > > > data cache.
> > > >
> > > > Then we want to switch to ASE 12.5
> > > > We take the same hardware configuration and decide to split database
> > into
> > > > four physical devices. We create separate database devices for
> > > datasegment,
> > > > index segment for indexes, log segment and  tempdb. WE create
database
> > > from
> > > > the scratch (using scripts, omitting index creation) and bcp all
data
> > into
> > > > database. Then we create indexes, triggers and stored procedures. We
> run
> > > > reorg rebuild for all tables in order to refresh statistics and
> tables.
> > > The
> > > > cache configuration is the same.
> > > >
> > > > We expected that for some performance gain using parallel query
> > > processing,
> > > > worker processes etc. But we got TWICE SLOWDOWN in average
performance
> > and
> > > > in query execution!!!!
> > > >
> > > > Can anybody give some advice what can we do else to rise
performance?
> > > > Does creation of separate cache for syslogs can give some advantage?
> > > > Should we run update index statistics after executing  reorg
rebuild?
> > > > How we can tune ASE 12.5 optimizer? May be he is based on different
> > > > algorithm?
> > > > How parallel processing can increase (decrease) performance?
> > > > Any other suggestions?
> > > >
> > > > Regards
> > > >
> > > > AlexM
> > > >
> > > >
> > > >
> > >
> > >
> >
> >
>
>


0
Alexander
4/10/2002 5:24:33 PM
alex@alexm.carrier.kiev.ua wrote...
> I mean 400 tables, 2000 stored procedures and rather complicated analitical
> queries
> 

Well, this doesn't mean much.  I have walked into shops that had 1000 procedures but only 
300 to 400 were actually used.  All the other ones were procedures no longer used but the 
developers were afraid to drop them.

From your configuration I will guess the following:
- Open object setting is too low.  You have 2000 procedures but only allow 1000 of them to 
be open at any one time.  What about triggers?
- You have two processors but only one partition for each cache.  If you have spinlock 
contention then setting the global cache partition number to 2 would help reduce that.

In order to really see what's going on a sysmon report would be helpful.  Not that it will 
tell us enough to really solve your problem though we'll give it a try.
-- 
Jim Egan [TeamSybase]
Senior Consultant
Sybase Professional Services
0
Jim
4/11/2002 3:53:08 AM
"Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
news:MPG.171eb6926631f81a98bd02@forums.sybase.com...
> alex@alexm.carrier.kiev.ua wrote...
> > I mean 400 tables, 2000 stored procedures and rather complicated
analitical
> > queries
> >
>
> Well, this doesn't mean much.  I have walked into shops that had 1000
procedures but only
> 300 to 400 were actually used.  All the other ones were procedures no
longer used but the
> developers were afraid to drop them.
>
> From your configuration I will guess the following:
> - Open object setting is too low.  You have 2000 procedures but only allow
1000 of them to
> be open at any one time.  What about triggers?

All of these procedures are used, but not in the same moment. There is a
really "fat server with fat clien"
architecture. I do not point number of users - from 50 to 100.

> - You have two processors but only one partition for each cache.  If you
have spinlock
> contention then setting the global cache partition number to 2 would help
reduce that.
>
As i understand, this is already done (in the file)

Actually we notice performance descrease even in table and sp creation speed
and in the bcp in process -
i do not think that this is a point of parallel processing. I guess - may be
this is an optimiser ?

> In order to really see what's going on a sysmon report would be helpful.
Not that it will
> tell us enough to really solve your problem though we'll give it a try.

What should I run? I run "sp_configure memory" but it doesn't give me a lot
of information
Please help

> --
> Jim Egan [TeamSybase]
> Senior Consultant
> Sybase Professional Services


0
Alexander
4/11/2002 12:28:51 PM
> What should I run? I run "sp_configure memory" but it doesn't give me a lot
> of information

You can run sp_monitorconfig 'all' to check some basic resources.

What I usually do is using sp_lock to see which queries are running
for a long time, and then use the sybase central monitor to capture
the sql text of the query. With sp_showplan, you can look at the 
query plan to see if the right indexes are used.

Migrating from 11.9.2 to 12.0 caused also some performance problems.
Everyone complained, but it was only one query that wasn't using the
most optimal query plan.

Luc.

____________________________________________________________________
Luc Van der Veurst                                    ISUG Treasurer
Academic Hospital, VUB, Brussels, Belgium        http://www.isug.com
----                                                         ----
Sybase TechWave 2002: San Diego, August 5-9. THE PLACE TO BE !!!!!!!
Join ISUG first and get a $75 discount.    Visit http://www.isug.com

0
Luc
4/11/2002 1:19:14 PM
alex@alexm.carrier.kiev.ua wrote...
> > In order to really see what's going on a sysmon report would be helpful.
> Not that it will
> > tell us enough to really solve your problem though we'll give it a try.
> 
> What should I run? I run "sp_configure memory" but it doesn't give me a lot
> of information
> Please help

While your processes are running execute the following:
sp_sysmon "00:05:00"

This will run sysmon for 5 minutes and produce a lengthy set of output.  Post the output 
here and we'll take a look at it.
-- 
Jim Egan [TeamSybase]
Senior Consultant
Sybase Professional Services
0
Jim
4/11/2002 3:44:18 PM
Thank you very much, Jim
I find this dp_sysmon in documentation
the only question is the meaning of that four columns in the report - they
do not comment them:

per sec     per xact     count     % of total
--------    ---------    ------    ----------
can you please comment?

and I am still wondering about optimizer behavior. Somewhere I hear that
12.5 optimizer chooses tables order in the join in different way than 11.5

any comments?


"Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
news:MPG.171f5d40243c59a698bd08@forums.sybase.com...
> alex@alexm.carrier.kiev.ua wrote...
> > > In order to really see what's going on a sysmon report would be
helpful.
> > Not that it will
> > > tell us enough to really solve your problem though we'll give it a
try.
> >
> > What should I run? I run "sp_configure memory" but it doesn't give me a
lot
> > of information
> > Please help
>
> While your processes are running execute the following:
> sp_sysmon "00:05:00"
>
> This will run sysmon for 5 minutes and produce a lengthy set of output.
Post the output
> here and we'll take a look at it.
> --
> Jim Egan [TeamSybase]
> Senior Consultant
> Sybase Professional Services


0
Alexander
4/11/2002 3:59:56 PM
BTW

Have you ever use servers with pagesize different from default?
Does it really increase performance?

"Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
news:MPG.171f5d40243c59a698bd08@forums.sybase.com...
> alex@alexm.carrier.kiev.ua wrote...
> > > In order to really see what's going on a sysmon report would be
helpful.
> > Not that it will
> > > tell us enough to really solve your problem though we'll give it a
try.
> >
> > What should I run? I run "sp_configure memory" but it doesn't give me a
lot
> > of information
> > Please help
>
> While your processes are running execute the following:
> sp_sysmon "00:05:00"
>
> This will run sysmon for 5 minutes and produce a lengthy set of output.
Post the output
> here and we'll take a look at it.
> --
> Jim Egan [TeamSybase]
> Senior Consultant
> Sybase Professional Services


0
Alexander
4/11/2002 4:05:33 PM
This is a multi-part message in MIME format.
--------------D7CB83003F763CF12013B2A0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Alexander Mitchenko wrote:
> 
> Hi All,
> Again my question is about performance of ASE 12.5 against 11.5.01
> We have a huge production client-server application running on W2K server
> with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
> ...
> Then we want to switch to ASE 12.5
....
> We expected that for some performance gain using parallel query processing,
> worker processes etc. But we got TWICE SLOWDOWN in average performance and
> in query execution!!!!
> 
> Can anybody give some advice what can we do else to rise performance?
> Does creation of separate cache for syslogs can give some advantage?
> Should we run update index statistics after executing  reorg rebuild?
> How we can tune ASE 12.5 optimizer? May be he is based on different
> algorithm?
> How parallel processing can increase (decrease) performance?

Alexander,
First of all, you had/have your work cut out for you when deciding to
upgrade from ASE 11.5 directly to 12.5.  That encompasses 3 Major ASE
releases (11.9.x, 12.0, 12.5), each of which had some very major changes
in architecture and functionality.  Not the least of which were changes
that could affect performance of your system.  The laundry list of these
which can be derived from some reading that you will have to do.  Start
with the "What's New in ASE 12.5" guide at:

http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView

Pay particular attention to the changes to the way statistics are
collected and optimizer changes in 11.9.2, and 12.0.  The next bit of
reading that will help you understand your 12.5 ASE performance is the
"Performance and Tuning Guide".  Here is a link to that if you don't
already have one:

http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView

The upgrade process is probably something that should have been more
carefully planned for and tested in your case.  However, having said
that, if I were in your place, I'd immediately setup Monitor Server, and
Historical Server to find stored procedures which are not performing,
devices/tables/caches that are being hit hard, etc.

Here is all the info you need to setup, and use Monitor/Historical Server:

http://manuals.sybase.com/onlinebooks/group-mn/mng1150e

Your description of your problem is too generic to give specific advice
other than to use the tools that are available to you to troubleshoot
your problems.  If it turns out that you have issues with optimization,
post specific code, DDL, statistics and we'll do our best to help.
--------------D7CB83003F763CF12013B2A0
Content-Type: text/x-vcard; charset=us-ascii;
 name="ksherlo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Sherlock, Kevin
Content-Disposition: attachment;
 filename="ksherlo.vcf"

begin:vcard 
n:Sherlock;Kevin
x-mozilla-html:FALSE
url:http://qwestdex.com
org:Qwest DEX;Information Managment
adr:;;;Omaha;NE;68114;USA
version:2.1
email;internet:ksherlo_AT_qwest.com
title:Staff Information Systems Engineer
x-mozilla-cpt:;3
fn:Kevin Sherlock
end:vcard

--------------D7CB83003F763CF12013B2A0--

0
Sherlock
4/11/2002 4:46:48 PM
This is a multi-part message in MIME format.
--------------DAD8F4CC973E3EAA7BD10975
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Alexander Mitchenko wrote:
> 
> Thank you very much, Jim
> I find this dp_sysmon in documentation
> the only question is the meaning of that four columns in the report - they
> do not comment them:
> 
> per sec     per xact     count     % of total
> --------    ---------    ------    ----------
> can you please comment?

Well ok,
per sec - is the measurment of the identified metric in terms of an
average number of occurences of these items per clock second.
per xact - is the measurement of the identified metric in terms of an
avery number of occurences of these items per TSQL transaction
count - is the total number of the occurences of this metric during the
duration of the sp_sysmon report.
% of total - is the ratio of the count per total number of a superset of
this metric.

For example, if the line in the report is for Engine Yields:

CPU Yields by Engine  per sec    per xact   count   % of total
--------------------  ---------  ---------  ------  ---------- 
Engine 0              45.5       273.0      273     31.8 %
Engine 1              49.7       298.0      298     34.7 %
Engine 2              48.0       288.0      288     33.5 %
--------------------  ---------  ---------  ------
Total CPU Yields      143.2      859.0      859

Engine 0 yielded 45.5 times per second, 273 times per transaction in the
server (there was only one in this sample), and a total of 273 times
overall, which was 31.8% of the total engine yields over all of the engines.

> and I am still wondering about optimizer behavior. Somewhere I hear that
> 12.5 optimizer chooses tables order in the join in different way than 11.5
> any comments?

There were many, many, many changes to the optimizer/normalizer,
statistics architecture, cache configurations, etc between 11.5 and
12.5.  See my previous post in this thread about this subject.  Your
statement above is far too generalized.  Join order is a fuction of the
optimizer which underwent changes which implies that you MIGHT see
differences between 11.5 and 12.5.  In most cases for the better, but in
some cases not.  You need to pinpoint where (stored procedure, SQL) you
have performance issues and treat them on a case-by-case basis.
--------------DAD8F4CC973E3EAA7BD10975
Content-Type: text/x-vcard; charset=us-ascii;
 name="ksherlo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Sherlock, Kevin
Content-Disposition: attachment;
 filename="ksherlo.vcf"

begin:vcard 
n:Sherlock;Kevin
x-mozilla-html:FALSE
url:http://qwestdex.com
org:Qwest DEX;Information Managment
adr:;;;Omaha;NE;68114;USA
version:2.1
email;internet:ksherlo_AT_qwest.com
title:Staff Information Systems Engineer
x-mozilla-cpt:;3
fn:Kevin Sherlock
end:vcard

--------------DAD8F4CC973E3EAA7BD10975--

0
Sherlock
4/11/2002 5:04:49 PM
Thank you very much Kevin,

I guess I correctly use your first name :)))
"Sherlock, Kevin" <ksherlo@qwest.com.nospam> wrote in message
news:3CB5BDEF.5E563F5B@qwest.com.nospam...
> Alexander Mitchenko wrote:
> >
> > Hi All,
> > Again my question is about performance of ASE 12.5 against 11.5.01
> > We have a huge production client-server application running on W2K
server
> > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
procedures).
> > ...
> > Then we want to switch to ASE 12.5
> ...
> > We expected that for some performance gain using parallel query
processing,
> > worker processes etc. But we got TWICE SLOWDOWN in average performance
and
> > in query execution!!!!
> >
> > Can anybody give some advice what can we do else to rise performance?
> > Does creation of separate cache for syslogs can give some advantage?
> > Should we run update index statistics after executing  reorg rebuild?
> > How we can tune ASE 12.5 optimizer? May be he is based on different
> > algorithm?
> > How parallel processing can increase (decrease) performance?
>
> Alexander,
> First of all, you had/have your work cut out for you when deciding to
> upgrade from ASE 11.5 directly to 12.5.  That encompasses 3 Major ASE
> releases (11.9.x, 12.0, 12.5), each of which had some very major changes
> in architecture and functionality.  Not the least of which were changes
> that could affect performance of your system.  The laundry list of these
> which can be derived from some reading that you will have to do.  Start
> with the "What's New in ASE 12.5" guide at:
>
>
http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic
__BookView

This is a good advice - i think we can find a lot of userfull info in there

> Pay particular attention to the changes to the way statistics are
> collected and optimizer changes in 11.9.2, and 12.0.  The next bit of
> reading that will help you understand your 12.5 ASE performance is the
> "Performance and Tuning Guide".  Here is a link to that if you don't
> already have one:
>
>
http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__Bo
okView
>
> The upgrade process is probably something that should have been more
> carefully planned for and tested in your case.  However, having said
> that, if I were in your place, I'd immediately setup Monitor Server, and
> Historical Server to find stored procedures which are not performing,
> devices/tables/caches that are being hit hard, etc.
>
We do a lot of testing

> Here is all the info you need to setup, and use Monitor/Historical Server:
>
> http://manuals.sybase.com/onlinebooks/group-mn/mng1150e
>
> Your description of your problem is too generic to give specific advice
> other than to use the tools that are available to you to troubleshoot
> your problems.  If it turns out that you have issues with optimization,
> post specific code, DDL, statistics and we'll do our best to help.


0
Alexander
4/11/2002 5:14:12 PM
Thank you Kevin,

"Sherlock, Kevin" <ksherlo@qwest.com.nospam> wrote in message
news:3CB5C227.BB017845@qwest.com.nospam...
> Alexander Mitchenko wrote:
> >
> > Thank you very much, Jim
> > I find this dp_sysmon in documentation
> > the only question is the meaning of that four columns in the report -
they
> > do not comment them:
> >
> > per sec     per xact     count     % of total
> > --------    ---------    ------    ----------
> > can you please comment?
>
> Well ok,
> per sec - is the measurment of the identified metric in terms of an
> average number of occurences of these items per clock second.
> per xact - is the measurement of the identified metric in terms of an
> avery number of occurences of these items per TSQL transaction
> count - is the total number of the occurences of this metric during the
> duration of the sp_sysmon report.
> % of total - is the ratio of the count per total number of a superset of
> this metric.
>
> For example, if the line in the report is for Engine Yields:
>
> CPU Yields by Engine  per sec    per xact   count   % of total
> --------------------  ---------  ---------  ------  ----------
> Engine 0              45.5       273.0      273     31.8 %
> Engine 1              49.7       298.0      298     34.7 %
> Engine 2              48.0       288.0      288     33.5 %
> --------------------  ---------  ---------  ------
> Total CPU Yields      143.2      859.0      859
>
> Engine 0 yielded 45.5 times per second, 273 times per transaction in the
> server (there was only one in this sample), and a total of 273 times
> overall, which was 31.8% of the total engine yields over all of the
engines.
>
This explanation is perfect to understand - thank you again

> > and I am still wondering about optimizer behavior. Somewhere I hear that
> > 12.5 optimizer chooses tables order in the join in different way than
11.5
> > any comments?
>
> There were many, many, many changes to the optimizer/normalizer,
> statistics architecture, cache configurations, etc between 11.5 and
> 12.5.  See my previous post in this thread about this subject.  Your
> statement above is far too generalized.  Join order is a fuction of the
> optimizer which underwent changes which implies that you MIGHT see
> differences between 11.5 and 12.5.  In most cases for the better, but in
> some cases not.  You need to pinpoint where (stored procedure, SQL) you
> have performance issues and treat them on a case-by-case basis.
Unfortunately to many cases - a result of four years of work.
we are trying to find some tendencies

In any case, thanks a lot!




0
Alexander
4/11/2002 5:17:25 PM
--------------020106030600050108040805
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Heve you deleted all statistics and then recreate them after the 
upgrade, also reload all stored procedures and triggers.  

Sherlock, Kevin wrote:

>Alexander Mitchenko wrote:
>
>>Hi All,
>>Again my question is about performance of ASE 12.5 against 11.5.01
>>We have a huge production client-server application running on W2K server
>>with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
>>...
>>Then we want to switch to ASE 12.5
>>
>...
>
>>We expected that for some performance gain using parallel query processing,
>>worker processes etc. But we got TWICE SLOWDOWN in average performance and
>>in query execution!!!!
>>
>>Can anybody give some advice what can we do else to rise performance?
>>Does creation of separate cache for syslogs can give some advantage?
>>Should we run update index statistics after executing  reorg rebuild?
>>How we can tune ASE 12.5 optimizer? May be he is based on different
>>algorithm?
>>How parallel processing can increase (decrease) performance?
>>
>
>Alexander,
>First of all, you had/have your work cut out for you when deciding to
>upgrade from ASE 11.5 directly to 12.5.  That encompasses 3 Major ASE
>releases (11.9.x, 12.0, 12.5), each of which had some very major changes
>in architecture and functionality.  Not the least of which were changes
>that could affect performance of your system.  The laundry list of these
>which can be derived from some reading that you will have to do.  Start
>with the "What's New in ASE 12.5" guide at:
>
>http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView
>
>Pay particular attention to the changes to the way statistics are
>collected and optimizer changes in 11.9.2, and 12.0.  The next bit of
>reading that will help you understand your 12.5 ASE performance is the
>"Performance and Tuning Guide".  Here is a link to that if you don't
>already have one:
>
>http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView
>
>The upgrade process is probably something that should have been more
>carefully planned for and tested in your case.  However, having said
>that, if I were in your place, I'd immediately setup Monitor Server, and
>Historical Server to find stored procedures which are not performing,
>devices/tables/caches that are being hit hard, etc.
>
>Here is all the info you need to setup, and use Monitor/Historical Server:
>
>http://manuals.sybase.com/onlinebooks/group-mn/mng1150e
>
>Your description of your problem is too generic to give specific advice
>other than to use the tools that are available to you to troubleshoot
>your problems.  If it turns out that you have issues with optimization,
>post specific code, DDL, statistics and we'll do our best to help.
>


--------------020106030600050108040805
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<html>
<head>
</head>
<body>
Heve you deleted all statistics and then recreate them after the upgrade,
also reload all stored procedures and triggers. &nbsp;<br>
<br>
Sherlock, Kevin wrote:<br>
<blockquote type="cite" cite="mid:3CB5BDEF.5E563F5B@qwest.com.nospam">
  <pre wrap="">Alexander Mitchenko wrote:<br></pre>
  <blockquote type="cite">
    <pre wrap="">Hi All,<br>Again my question is about performance of ASE 12.5 against 11.5.01<br>We have a huge production client-server application running on W2K server<br>with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).<br>...<br>Then we want to switch to ASE 12.5<br></pre>
    </blockquote>
    <pre wrap=""><!---->...<br></pre>
    <blockquote type="cite">
      <pre wrap="">We expected that for some performance gain using parallel query processing,<br>worker processes etc. But we got TWICE SLOWDOWN in average performance and<br>in query execution!!!!<br><br>Can anybody give some advice what can we do else to rise performance?<br>Does creation of separate cache for syslogs can give some advantage?<br>Should we run update index statistics after executing  reorg rebuild?<br>How we can tune ASE 12.5 optimizer? May be he is based on different<br>algorithm?<br>How parallel processing can increase (decrease) performance?<br></pre>
      </blockquote>
      <pre wrap=""><!----><br>Alexander,<br>First of all, you had/have your work cut out for you when deciding to<br>upgrade from ASE 11.5 directly to 12.5.  That encompasses 3 Major ASE<br>releases (11.9.x, 12.0, 12.5), each of which had some very major changes<br>in architecture and functionality.  Not the least of which were changes<br>that could affect performance of your system.  The laundry list of these<br>which can be derived from some reading that you will have to do.  Start<br>with the "What's New in ASE 12.5" guide at:<br><br><a class="moz-txt-link-freetext" href="http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView">http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView</a><br><br>Pay particular attention to the changes to the way statistics are<br>collected and optimizer changes in 11.9.2, and 12.0.  The next bit of<br>reading that will help you understand your 12.5 ASE performance is the<br>"Perf
ormance and Tuning Guide".  Here is a link to that if you don't<br>already have one:<br><br><a class="moz-txt-link-freetext" href="http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView">http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView</a><br><br>The upgrade process is probably something that should have been more<br>carefully planned for and tested in your case.  However, having said<br>that, if I were in your place, I'd immediately setup Monitor Server, and<br>Historical Server to find stored procedures which are not performing,<br>devices/tables/caches that are being hit hard, etc.<br><br>Here is all the info you need to setup, and use Monitor/Historical Server:<br><br><a class="moz-txt-link-freetext" href="http://manuals.sybase.com/onlinebooks/group-mn/mng1150e">http://manuals.sybase.com/onlinebooks/group-mn/mng1150e</a><br><br>Your description of your problem is too generic to give specific advice<br>other tha
n to use the tools that are available to you to troubleshoot<br>your problems.  If it turns out that you have issues with optimization,<br>post specific code, DDL, statistics and we'll do our best to help.</pre>
      </blockquote>
      <br>
      </body>
      </html>

--------------020106030600050108040805--

0
Allen
4/12/2002 2:41:30 AM
alex@alexm.carrier.kiev.ua wrote...
> Have you ever use servers with pagesize different from default?
> Does it really increase performance?
> 

In certain cases it can indeed improve performance.  But there may be trade-offs and 
additional configuration that will be necessary to achieve top performance.
-- 
Jim Egan [TeamSybase]
Senior Consultant
Sybase Professional Services
0
Jim
4/12/2002 4:33:27 AM
agcornell@email.msn.com wrote...
> Heve you deleted all statistics and then recreate them after the 
> upgrade, also reload all stored procedures and triggers.
> 

NO, NO, NO!!!!  Deleting the statistics is not something you should do after an upgrade.  
After the upgrade there are valuable settings brought forward from the older version.  When 
you delete the statistics you'll wipe them out.
-- 
Jim Egan [TeamSybase]
Senior Consultant
Sybase Professional Services
0
Jim
4/12/2002 4:35:18 AM
Hi
How can we improve performance?

We're moving a database that was operating fairly well on a Sun 3000 with=20
Adaptive Server Enterprise v. 11.5 (32 bits) to a Sun 10000 with Adaptive=20
Server Enterprise v. 12.5 (64 bits) and Solaris v. 8
Our problem is the slower performance of the database on the new, more=20
powerful and faster machine with newer software (in both operative system=20
and sybase version)

A particular query with insert used to work for about 2 hours in a Sun 3000=
=20
with Adaptive Server Enterprise v. 11.5 (32 bits).=20
In a Sun 10000 with Adaptive Server Enterprise v. 12.5 (64 bits) and=20
Solaris v. 8 it takes near to 5 hours to complete.

We've noted that it starts really fast (as if it would take about 15=20
minutes) but slows down gradually. The priority changes from meduim to low.

We=B4ve tried many ways to improve the performance:
-Increase the max memory.
-Increase the number of engines at startup and online.
-Increase the procedure cache size.
-Increase the default data cache.
-Segment and create a named cache for tempdb.
-Add more network memory.
-Increase the size of the default network packet size.
-Create a segment for the biggest user table on the database, which we're=20
using for tests, and separate data and indexes on different devices.
-Create a set of 16K pools for various named caches.
-Create a new execution class for a special very time consuming query (an=20
47000 row insert)=20
-Create again index, stored procedure, triggers and tables.
-We install the patches that are available in Sybase.com for the version=20
12.5 to 64 bit

There is no contention because we're in tests and this query is the only=20
process running on the machine.

Now we have the same problem!.

->Is there any problems with this combination of versions (maybe we're=20
lacking additional software) of Adaptive Server and Solaris?

->Is there any combination of performance changes that works best?

Thanks
0
Richard_M
4/12/2002 2:53:56 PM
--------------CFCA308546045965EFF9E645
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

DO NOT DELETE STATISTICS AFTER UPGRADE!!!!
You will lose the inherited step counts from the old distribution page.
I'll post details later.

Is this myth still alive? I thought I killed it at Techwave and in
articles last year ;-)

Eric Miner
Sr. Engineer
Optimizer Group

Allen Cornell wrote:

> Heve you deleted all statistics and then recreate them after the
> upgrade, also reload all stored procedures and triggers.
>
> Sherlock, Kevin wrote:
>
>> Alexander Mitchenko wrote:
>>
>> > Hi All,
>> > Again my question is about performance of ASE 12.5 against 11.5.01
>> > We have a huge production client-server application running on W2K
>> > server
>> > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
>> > procedures).
>> > ...
>> > Then we want to switch to ASE 12.5
>> >
>> ...
>>
>> > We expected that for some performance gain using parallel query
>> > processing,
>> > worker processes etc. But we got TWICE SLOWDOWN in average
>> > performance and
>> > in query execution!!!!
>> >
>> > Can anybody give some advice what can we do else to rise
>> > performance?
>> > Does creation of separate cache for syslogs can give some
>> > advantage?
>> > Should we run update index statistics after executing  reorg
>> > rebuild?
>> > How we can tune ASE 12.5 optimizer? May be he is based on different
>> > algorithm?
>> > How parallel processing can increase (decrease) performance?
>> >
>> Alexander,
>> First of all, you had/have your work cut out for you when deciding
>> to
>> upgrade from ASE 11.5 directly to 12.5.  That encompasses 3 Major
>> ASE
>> releases (11.9.x, 12.0, 12.5), each of which had some very major
>> changes
>> in architecture and functionality.  Not the least of which were
>> changes
>> that could affect performance of your system.  The laundry list of
>> these
>> which can be derived from some reading that you will have to do.
>> Start
>> with the "What's New in ASE 12.5" guide at:
>>
>> http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView
>>
>> Pay particular attention to the changes to the way statistics are
>> collected and optimizer changes in 11.9.2, and 12.0.  The next bit
>> of
>> reading that will help you understand your 12.5 ASE performance is
>> the
>> "Perf
>> ormance and Tuning Guide".  Here is a link to that if you don't
>> already have one:
>>
>> http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView
>>
>> The upgrade process is probably something that should have been more
>> carefully planned for and tested in your case.  However, having said
>> that, if I were in your place, I'd immediately setup Monitor Server,
>> and
>> Historical Server to find stored procedures which are not
>> performing,
>> devices/tables/caches that are being hit hard, etc.
>>
>> Here is all the info you need to setup, and use Monitor/Historical
>> Server:
>>
>> http://manuals.sybase.com/onlinebooks/group-mn/mng1150e
>>
>> Your description of your problem is too generic to give specific
>> advice
>> other tha
>> n to use the tools that are available to you to troubleshoot
>> your problems.  If it turns out that you have issues with
>> optimization,
>> post specific code, DDL, statistics and we'll do our best to help.
>>

--------------CFCA308546045965EFF9E645
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
DO NOT DELETE STATISTICS AFTER UPGRADE!!!!
<br>You will lose the inherited step counts from the old distribution page.
I'll post details later.
<p>Is this myth still alive? I thought I killed it at Techwave and in articles
last year ;-)
<p>Eric Miner
<br>Sr. Engineer
<br>Optimizer Group
<p>Allen Cornell wrote:
<blockquote TYPE=CITE>Heve you deleted all statistics and then recreate
them after the upgrade, also reload all stored procedures and triggers.
<p>Sherlock, Kevin wrote:
<blockquote type="cite" cite="mid:3CB5BDEF.5E563F5B@qwest.com.nospam">
<pre wrap="">Alexander Mitchenko wrote:</pre>

<blockquote type="cite">
<pre wrap="">Hi All,
Again my question is about performance of ASE 12.5 against 11.5.01
We have a huge production client-server application running on W2K server
with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
....
Then we want to switch to ASE 12.5</pre>
</blockquote>

<pre wrap=""><!---->...</pre>

<blockquote type="cite">
<pre wrap="">We expected that for some performance gain using parallel query processing,
worker processes etc. But we got TWICE SLOWDOWN in average performance and
in query execution!!!!

Can anybody give some advice what can we do else to rise performance?
Does creation of separate cache for syslogs can give some advantage?
Should we run update index statistics after executing&nbsp; reorg rebuild?
How we can tune ASE 12.5 optimizer? May be he is based on different
algorithm?
How parallel processing can increase (decrease) performance?</pre>
</blockquote>

<pre wrap=""><!---->
Alexander,
First of all, you had/have your work cut out for you when deciding to
upgrade from ASE 11.5 directly to 12.5.&nbsp; That encompasses 3 Major ASE
releases (11.9.x, 12.0, 12.5), each of which had some very major changes
in architecture and functionality.&nbsp; Not the least of which were changes
that could affect performance of your system.&nbsp; The laundry list of these
which can be derived from some reading that you will have to do.&nbsp; Start
with the "What's New in ASE 12.5" guide at:

<a href="http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView" class="moz-txt-link-freetext">http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/whatsnew/@Generic__BookView

</a>Pay particular attention to the changes to the way statistics are
collected and optimizer changes in 11.9.2, and 12.0.&nbsp; The next bit of
reading that will help you understand your 12.5 ASE performance is the
"Perf
ormance and Tuning Guide".&nbsp; Here is a link to that if you don't
already have one:

<a href="http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView" class="moz-txt-link-freetext">http://manuals.sybase.com:80/onlinebooks/group-as/asg1250e/perf/@Generic__BookView

</a>The upgrade process is probably something that should have been more
carefully planned for and tested in your case.&nbsp; However, having said
that, if I were in your place, I'd immediately setup Monitor Server, and
Historical Server to find stored procedures which are not performing,
devices/tables/caches that are being hit hard, etc.

Here is all the info you need to setup, and use Monitor/Historical Server:

<a href="http://manuals.sybase.com/onlinebooks/group-mn/mng1150e" class="moz-txt-link-freetext">http://manuals.sybase.com/onlinebooks/group-mn/mng1150e

</a>Your description of your problem is too generic to give specific advice
other tha
n to use the tools that are available to you to troubleshoot
your problems.&nbsp; If it turns out that you have issues with optimization,
post specific code, DDL, statistics and we'll do our best to help.</pre>
</blockquote>
</blockquote>
</html>

--------------CFCA308546045965EFF9E645--

0
Eric
4/12/2002 3:53:52 PM
After upgrade run the update stats syntax you're always run - update
statistics table_name (optional name). This will gather and write the stats
in the new format. Once things are running OK at the query level then I
highly recommend you test the your queries with statistics on all columns of
your composite indexes or on non-indexed join columns - update index
statistics table_name [optional index name] and update statistics table_name
(col_name). Take a look at the P&T Guide for details.

I WILL NEVER RECOMMEND THE USE OF UPDATE ALL STATISTICS.
It will likely take a very long time to run and I have yet to see a table
that needed stats on all its columns.

BTW - if you're going to run update index stats or update stats on specific
columns make sure you have free space in tempdb.

Hope this helps,

Eric Miner
Sr. Engineer
Optimizer Group

Alexander Mitchenko wrote:

> Actually we create database from the scratch and bcp data in
>
> The real question is what update statistics we should use
> As follows from documentation update all statistics and update index
> statistics do different things
> so should we run both of them
>
> Thanks
>
> Alexander Mitchenko
>
> "Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
> news:MPG.172011f16c0d934498bd1a@forums.sybase.com...
> > agcornell@email.msn.com wrote...
> > > Heve you deleted all statistics and then recreate them after the
> > > upgrade, also reload all stored procedures and triggers.
> > >
> >
> > NO, NO, NO!!!!  Deleting the statistics is not something you should do
> after an upgrade.
> > After the upgrade there are valuable settings brought forward from the
> older version.  When
> > you delete the statistics you'll wipe them out.
> > --
> > Jim Egan [TeamSybase]
> > Senior Consultant
> > Sybase Professional Services

0
Eric
4/12/2002 4:01:02 PM
This is a multi-part message in MIME format.
--------------C1414E67D163A26614855148
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit

Richard_M. wrote:
> 
> Hi
> How can we improve performance?
> 
> We're moving a database that was operating fairly well on a Sun 3000 with
> Adaptive Server Enterprise v. 11.5 (32 bits) to a Sun 10000 with Adaptive
> Server Enterprise v. 12.5 (64 bits) and Solaris v. 8
> Our problem is the slower performance of the database on the new, more
> powerful and faster machine with newer software (in both operative system
> and sybase version)
> 
> A particular query with insert used to work for about 2 hours in a Sun 3000
> with Adaptive Server Enterprise v. 11.5 (32 bits).
> In a Sun 10000 with Adaptive Server Enterprise v. 12.5 (64 bits) and
> Solaris v. 8 it takes near to 5 hours to complete.

Please post the code and the table/index DDL, and statistics from
optdiag.  Let's start taking a look at details...

> We've noted that it starts really fast (as if it would take about 15
> minutes) but slows down gradually. The priority changes from meduim to low.
> 
> We�ve tried many ways to improve the performance:
> -Increase the max memory.

This does no good until that memory is used.

> -Increase the number of engines at startup and online.

Only affects throughput of multiple users, not just one connection at a time.

> -Increase the procedure cache size.

Ok, but only needed if you are observing a lot of procedure cache
misses.  What does your sp_sysmon output show there?

> -Increase the default data cache.

Ok, more memory is always good, but were you seeing a lot of cache
misses? 

> -Segment and create a named cache for tempdb.

> -Add more network memory.

Only needed for connections that use larger packet sizes than the
"default packet size"

> -Increase the size of the default network packet size.

Ok, but client still needs to ask for large packet sizes.  Did you do
this because of a network bottleneck?

> -Create a segment for the biggest user table on the database, which we're
> using for tests, and separate data and indexes on different devices.

Ok if that table is highly active.

> -Create a set of 16K pools for various named caches.

How many named caches do you have?  Post your configuration file.

> -Create a new execution class for a special very time consuming query (an
> 47000 row insert)

I've found that there are very few situations where this will help. 
Higher execution classes can backfire.  Post the code, DDL, etc. as
mentioned before.  47,000 rows is not that much.

> -Create again index, stored procedure, triggers and tables.
> -We install the patches that are available in Sybase.com for the version
> 12.5 to 64 bit
> 
> There is no contention because we're in tests and this query is the only
> process running on the machine.

You probably have a SQL, or table/index design issue here.  Post the
code, DDL, stats etc.

> 
> Now we have the same problem!.
> 
> ->Is there any problems with this combination of versions (maybe we're
> lacking additional software) of Adaptive Server and Solaris?
> 
> ->Is there any combination of performance changes that works best?

Depends on too many factors and how you use ASE.  There is no such
answer to that question without a full understanding of your
application, database design, etc.
--------------C1414E67D163A26614855148
Content-Type: text/x-vcard; charset=us-ascii;
 name="ksherlo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Sherlock, Kevin
Content-Disposition: attachment;
 filename="ksherlo.vcf"

begin:vcard 
n:Sherlock;Kevin
x-mozilla-html:FALSE
url:http://qwestdex.com
org:Qwest DEX;Information Managment
adr:;;;Omaha;NE;68114;USA
version:2.1
email;internet:ksherlo_AT_qwest.com
title:Staff Information Systems Engineer
x-mozilla-cpt:;3
fn:Kevin Sherlock
end:vcard

--------------C1414E67D163A26614855148--

0
Sherlock
4/12/2002 5:06:47 PM
Actually we create database from the scratch and bcp data in

The real question is what update statistics we should use
As follows from documentation update all statistics and update index
statistics do different things
so should we run both of them

Thanks

Alexander Mitchenko


"Jim Egan" <dontspam.dbaguru@eganomics.com> wrote in message
news:MPG.172011f16c0d934498bd1a@forums.sybase.com...
> agcornell@email.msn.com wrote...
> > Heve you deleted all statistics and then recreate them after the
> > upgrade, also reload all stored procedures and triggers.
> >
>
> NO, NO, NO!!!!  Deleting the statistics is not something you should do
after an upgrade.
> After the upgrade there are valuable settings brought forward from the
older version.  When
> you delete the statistics you'll wipe them out.
> --
> Jim Egan [TeamSybase]
> Senior Consultant
> Sybase Professional Services


0
Alexander
4/12/2002 5:09:14 PM
This is a multi-part message in MIME format.
--------------E8CA6289B0497F5BF45C82FA
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

One quick question to verify one common mistake made when upgrading from
pre-12.0 servers.  Do you have any filesystem devices in your server,
especially in tempdb?  Do these filesystem devices have DSYNC on?  To
find out, do a "sp_helpdevice" and look for the DSYNC setting for your
filesystem based devices (if you have them).

Alexander Mitchenko wrote:
> 
> Hi All,
> Again my question is about performance of ASE 12.5 against 11.5.01
> We have a huge production client-server application running on W2K server
> with ASE 11.5.0.1 database (more than 400 tables, 2000 stored procedures).
> The database was placed on raw disk partitions (no RAID) and split between
> different physical drives as follows: data, log, tempdb and master database
> are located on different drives.
> We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> separate tempdb cache and bind tempdb to it. Database was bind to default
> data cache.
> 
> Then we want to switch to ASE 12.5
> We take the same hardware configuration and decide to split database into
> four physical devices. We create separate database devices for datasegment,
> index segment for indexes, log segment and  tempdb. WE create database from
> the scratch (using scripts, omitting index creation) and bcp all data into
> database. Then we create indexes, triggers and stored procedures. We run
> reorg rebuild for all tables in order to refresh statistics and tables. The
> cache configuration is the same.
> 
> We expected that for some performance gain using parallel query processing,
> worker processes etc. But we got TWICE SLOWDOWN in average performance and
> in query execution!!!!
> 
> Can anybody give some advice what can we do else to rise performance?
> Does creation of separate cache for syslogs can give some advantage?
> Should we run update index statistics after executing  reorg rebuild?
> How we can tune ASE 12.5 optimizer? May be he is based on different
> algorithm?
> How parallel processing can increase (decrease) performance?
> Any other suggestions?
> 
> Regards
> 
> AlexM
--------------E8CA6289B0497F5BF45C82FA
Content-Type: text/x-vcard; charset=us-ascii;
 name="ksherlo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Sherlock, Kevin
Content-Disposition: attachment;
 filename="ksherlo.vcf"

begin:vcard 
n:Sherlock;Kevin
x-mozilla-html:FALSE
url:http://qwestdex.com
org:Qwest DEX;Information Managment
adr:;;;Omaha;NE;68114;USA
version:2.1
email;internet:ksherlo_AT_qwest.com
title:Staff Information Systems Engineer
x-mozilla-cpt:;3
fn:Kevin Sherlock
end:vcard

--------------E8CA6289B0497F5BF45C82FA--

0
Sherlock
4/16/2002 8:57:17 PM
Yes! This is a case. We do not switch DSYNC off and thus lose approximately
10% of performance
Thanks a lot !

"Sherlock, Kevin" <ksherlo@qwest.com.nospam> wrote in message
news:3CBC8F7F.BD7D7C79@qwest.com.nospam...
> One quick question to verify one common mistake made when upgrading from
> pre-12.0 servers.  Do you have any filesystem devices in your server,
> especially in tempdb?  Do these filesystem devices have DSYNC on?  To
> find out, do a "sp_helpdevice" and look for the DSYNC setting for your
> filesystem based devices (if you have them).
>
> Alexander Mitchenko wrote:
> >
> > Hi All,
> > Again my question is about performance of ASE 12.5 against 11.5.01
> > We have a huge production client-server application running on W2K
server
> > with ASE 11.5.0.1 database (more than 400 tables, 2000 stored
procedures).
> > The database was placed on raw disk partitions (no RAID) and split
between
> > different physical drives as follows: data, log, tempdb and master
database
> > are located on different drives.
> > We have two PIII, 1GHz, 1 GB of memory and give 800MB for ASE. We create
> > separate tempdb cache and bind tempdb to it. Database was bind to
default
> > data cache.
> >
> > Then we want to switch to ASE 12.5
> > We take the same hardware configuration and decide to split database
into
> > four physical devices. We create separate database devices for
datasegment,
> > index segment for indexes, log segment and  tempdb. WE create database
from
> > the scratch (using scripts, omitting index creation) and bcp all data
into
> > database. Then we create indexes, triggers and stored procedures. We run
> > reorg rebuild for all tables in order to refresh statistics and tables.
The
> > cache configuration is the same.
> >
> > We expected that for some performance gain using parallel query
processing,
> > worker processes etc. But we got TWICE SLOWDOWN in average performance
and
> > in query execution!!!!
> >
> > Can anybody give some advice what can we do else to rise performance?
> > Does creation of separate cache for syslogs can give some advantage?
> > Should we run update index statistics after executing  reorg rebuild?
> > How we can tune ASE 12.5 optimizer? May be he is based on different
> > algorithm?
> > How parallel processing can increase (decrease) performance?
> > Any other suggestions?
> >
> > Regards
> >
> > AlexM


0
Alexander
4/17/2002 12:07:41 PM
This is a multi-part message in MIME format.
--------------7D1FF26AFA37EDA8477CDE36
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Alexander Mitchenko wrote:
> 
> Yes! This is a case. We do not switch DSYNC off and thus lose approximately
> 10% of performance
> Thanks a lot !

So you are back to using filesystem devices which are buffered IO by the
OS now just as you were when using version 11.5.  I assume that you are
aware of the risk of using such devices in terms of recoverability.  Do
you have corporate data (user created databases) on these devices,
and/or is your master device a filesystem device also?  This is a risky
way to do business.  Or do you just have the tempdb device(s) on filesystems?
--------------7D1FF26AFA37EDA8477CDE36
Content-Type: text/x-vcard; charset=us-ascii;
 name="ksherlo.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Sherlock, Kevin
Content-Disposition: attachment;
 filename="ksherlo.vcf"

begin:vcard 
n:Sherlock;Kevin
x-mozilla-html:FALSE
url:http://qwestdex.com
org:Qwest DEX;Information Managment
adr:;;;Omaha;NE;68114;USA
version:2.1
email;internet:ksherlo_AT_qwest.com
title:Staff Information Systems Engineer
x-mozilla-cpt:;3
fn:Kevin Sherlock
end:vcard

--------------7D1FF26AFA37EDA8477CDE36--

0
Sherlock
4/17/2002 2:32:39 PM
We work on WinNT
WE place tempdb database on several filesystem files located on different
hardware devices
System databases are placed by default on files too. Only user databases we
place on raw partitions and our estimate is that
thus they are faster

Regards

AlexM

"Sherlock, Kevin" <ksherlo@qwest.com.nospam> wrote in message
news:3CBD8785.F92DA9E7@qwest.com.nospam...
> Alexander Mitchenko wrote:
> >
> > Yes! This is a case. We do not switch DSYNC off and thus lose
approximately
> > 10% of performance
> > Thanks a lot !
>
> So you are back to using filesystem devices which are buffered IO by the
> OS now just as you were when using version 11.5.  I assume that you are
> aware of the risk of using such devices in terms of recoverability.  Do
> you have corporate data (user created databases) on these devices,
> and/or is your master device a filesystem device also?  This is a risky
> way to do business.  Or do you just have the tempdb device(s) on
filesystems?


0
Alexander
4/23/2002 2:38:52 PM
Reply:

Similar Artilces:

Performance issues after upgrate ASE 11.5 to ASE 12.5
Hi How can we improve performance? We're moving a database that was operating fairly well on a HP-UX 10.20 with Adaptive Server Enterprise v. 11.5 (32 bits) to a HP-UX 11.0 with Adaptive Server Enterprise v. 12.5 (32 bits) Our problem is the slower performance of the database on the new, more powerful and faster machine with newer software (in both operative system and sybase version) We've noted that it starts really fast but slows down gradually. We�ve tried many ways to improve the performance: -Increase the max memory. -Increase the number of engines at startup...

Performance issues after upgrate ASE 11.5 to ASE 12.5 #2
Hi How can we improve performance? We're moving a database that was operating fairly well on a Sun 3000 with=20 Adaptive Server Enterprise v. 11.5 (32 bits) to a Sun 10000 with Adaptive=20 Server Enterprise v. 12.5 (64 bits) and Solaris v. 8 Our problem is the slower performance of the database on the new, more=20 powerful and faster machine with newer software (in both operative system=20 and sybase version) A particular query with insert used to work for about 2 hours in a Sun 3000= =20 with Adaptive Server Enterprise v. 11.5 (32 bits).=20 In a Sun 10000 with Adaptive ...

Dump/Load (from ASE 12.5.4 to ASE 15.x) - VS.
Hi All, We are in the process of planning the upgrade of our ASE 12.5.4 dataservers to ASE 15.x. What is the most recommended way to upgrade ASE dataservers from ASE 12.5.4 to ASE 15.x? Potential options: 1) Install new instances of ASE 15.x and load database dumps from our ASE 12.5.4 dataservers to the newly installed ASE 15.x dataservers. - OR - 2) Use the ASE upgrade utility (sqlupgrade) to upgrade our existing ASE 12.5.4 dataservers to ASE 15.x. What are the Pros vs. Cons for loading a 12.5.4 database into a 15.x ASE dataserver versus performing an ASE dataserv...

ASE Workplace 12.5 vs ASE Enterprise 12.5
Can anyone tell me what are the differences in the 2 packaging of ASE? TIA Himadri Hi Himadri, The main and I guess the only difference is price which is for WP ~= 4 times less. Category depends from machine level, for example for Sun the upper bound for WP is Sun E450 + 3 new XXXR. Go to http://eshop.sybase.com where you can see prices and in the help the up-to-date list of machines with category HTH, Eugene Himadri Laifangbam wrote: > Can anyone tell me what are the differences in the 2 packaging of ASE? > > TIA > Himadri For NT any machines up t...

Migration from ASE 11.5 to ASE 12.5
Hello I am planning the migration from Sybase 11.5 to 12.5. Can someone point me to some documentation on the subject? I am particularly interested in documents that explain the best way to migrate database objects (tables, indexes, etc...) and data. Thanks in advance. Jorge Vergara Sispro S.A. Canoas - RS - Brasil jlvergara@sispro.com.br wrote... > Hello > > I am planning the migration from Sybase 11.5 to 12.5. Can someone point > me to some documentation on the subject? > I am particularly interested in documents that explain the b...

ASE 11.5 on same server as ASE 12.5?
Anyone have 11.5 running on same server as 12.5 on NT? Would not be concerned to do this in Solaris as I am on an NT 4.0 Box sp6. Any known gotchas or issues? Going to take toni's advise in earlier posting on doing the 11.5 install as a seperate user on the server already running 12.5 just fine. Hope this wilkl take care of env vars??? We have one app that the company who developed it will no longer support us on newer versions of Sybase as a result this one crucial app must stay up and running on 11.5 while we allow the rest of our company to progress forward to 12.5 an...

ASE 12.5 Small Business vs. ASE 12.5 Enterprise
Hi, Does anyone know the Technical offering different between these 2 products. Looking for docs on limitations of this ASE Small Business edition. Is it available on all platforms ? I have worked with the ASE Eneterprise editions but may have a new project that won't require the full blown Enterprise edition. thanks "Robert Furlong" <rfurlong@fame.com> wrote in message news:3E305851.80AA820@fame.com... > Hi, > > Does anyone know the Technical offering different between these 2 > products. > Looking for docs on limitations of this ASE Sm...

Performance Degradation after ASE 12.0 upgrade to ASE 12.5
Hello, There was a issue posted on 08/06/2002 by 'Bob' related to performance degradation after upgrade of ASE 11.9.2 to ASE 12.5. I would like to know whether anyone has found any solution/workaround to resolve those issues without making any changes in the application code. We are also facing the same problems here after upgrade of ASE 12.0 to ASE 12.5. The program which takes 10 mins. on ASE 12.0 is taking more than 3 hrs to finish on ASE 12.5 despite of using the same indexes and the program which takes 3 hrs on ASE 12.0 is taking more than 24 hours on ASE 12.5. ...

Diferrence of performance on queries betwenne a ASE 11.5.1.1 and ASE 12.0
I'm on the way to migrate from ASE 11.5.1.1 (on HP-UX 10.0) to ASE 12.0 (64= =20 bits on HP-UX 11). My tests shows good and bad performance with queries. I'm very interesting with yours. Can you tell me about them. Thank you. Fran=E7oise GOISLARD DBA=20 =20 Francoise, There are major functional changes in ASE 11.9.2 and above. I'm glad to see you're testing. How did you get the data to 12.0? Have you turned on any of the 12.0 options? HAve you attempted any tuning? What, if anything, have you done so far? If you have a specific problem query that can b...

Vulnerability of ASE 12.5.2 vs ASE 12.5.3 ESD#3
Recently, Application Security, Inc. has run a vulnerability test on our ASE12.5.2. The program gave a 'high risk' on the ASE12.5.2 and suggested we install the latest Sybase patch 12.5.3 ESD#3. I have some question on the recommendation. 1. The latest patch is definitely not 12.5.3 ESD#3. Why did the program recommend 12.5.3 ESD#3? Isn't 12.5.4 better than 12.5.3? 2. Is it practical to keep up with the latest releases or patches? Now if we upgrade to 12.5.4 and then rerun the vulnerability test, will the program suggest that newer version is better with respect to s...

ASE 12.5 to ASE 12.5.3
One of our client is having 3 no license for ASE 12.5 on Win2K (running 3 production servers) and not in the support/AMC with Sybase.. is client eligible to use ASE 12.5.3 ? Thanks KRV ...

Converting NT 11.5 ASE to a Solaris 12.5 ASE
We are trying to migrate a server running on Windows NT to a Solaris machine. The NT box is running ASE 11.5 and the Solaris box is running 12.5. So far, the only way we have came up with to migrate the data is bcp. The problem is we estimate this to be probably about a 6-7 hour process. Our shop runs 24*7*365 and unfortunately downtime is not something we get. In the past when we have migrated servers, we have used warm-standby which has worked really nicely. But in this case we can't run dump and load to populate the database. Has anyone been able to migrate an NT box to a...

Same machine: ASE 12.5.3 Linux vs ASE 12.5.3 Windows
We had a report from one of our clients recently that our Sybase ASE-based app was running slowly on their new Linux box. As we develop internally against ASE on Windows we thought it would be an interesting test to install both the Windows and Linux flavors of ASE 12.5.3 onto the same server and look at the relative performance. We used a Dell PowerEdge 1950 with 2 x Intel Xeon 3.73GHz dual core processors. On one drive we put SuSE Enterprise Linux Server 10 x86-64, and on the other we put Windows Server 2003 Std R2 64-bit edition. Both were fresh-off-the-CD OS installs. ...

ASE 11.5.1 vs 12 query performance
The following SQL runs in < 1 second on 11.5.1 versus 43 seconds on ASE12. use advdb go select distinct pledge_pledge_number into #P2 from pledge where (pledge_donor_id = '0000060648' and pledge_annual_sw != "Y" and (pledge_amount > 0 or pledge_associated_credit_amt > 0)) go create unique index pn on #P2(pledge_pledge_number) go set statistics time on go set statistics io on go set showplan on go select sum(isnull(tt.gift_associated_amount,0)) from #P2, primary_pledge pp, primary_gift pg, gift tt ...

Web resources about - ASE 12.5 vs ASE 11.5 performance issues - sybase.ase.performance+tuning

Performance - Wikipedia, the free encyclopedia
A performance , in performing arts , generally comprises an event in which a performer or group of performers behave in a particular way for ...

Andrew Bogut puts back issues behind him in return performance against Utah Jazz
Andrew Bogut's third double-double of the NBA season shows he is ready for the Christmas Day clash with the Cleveland Cavaliers.

Ars Technica’s (virtual) Performance Car of the Year
... coming out of the very slow turns 11 and 12. Brakes could be better. 18 more images in gallery 2015 has been the year of the performance car ...

Cam Newton Comments on Panthers' Performance in Loss to Falcons - Bleacher Report
The 1972 Miami Dolphins can rest easy, but the rest of the NFL will have to deal with an even more determined Cam Newton after the Carolina Panthers ...

How Twitter used Doze in Android 6.0 Marshmallow to Improve Notification Performance
... Advocate In October, we published a blog post about how Google Cloud Messaging (GCM) works with Doze in Android 6.0 Marshmallow to bring performance ...

Consumers value this in-car technology more than driving performance
... by the infotainment screen has been growing over the past year. In fact, consumers value in-car technology more than the car's driving performance, ...

How to make your performance review meaningful
Make your performance review meaningful Image by Thinkstock There's a prevailing opinion that performance reviews aren't the best way to evaluate ...

Wingels evaluates personal performance vs Hawks, reflects on year
... 58° Navigation Home Giants A's Sharks Warriors Kings 49ers Raiders Quakes Insiders More Tickets Shop Watch Wingels evaluates personal performance ...

Deadline’s The Contenders: How The Music In ‘Youth’ Became Such A Vital Part Of Michael Caine’s Performance ...
At Deadline’s big awards season day long event last month , The Contenders Presented By Deadline, Youth composer David Lang explained his process ...

Peyton Manning HGH Allegations - Peyton Manning Accused of Using Performance-Enhancing Drugs to Recover ...
... from a severe neck injury. An explosive report by Al Jazeera alleges that an Indianapolis clinic sent human growth hormone and other performance-enhancing ...

Resources last updated: 12/29/2015 2:09:15 PM