Better Performance WITH RECOMPILE option

Is performace gain when executes a stored procedure using the WITH RECOMPILE option ?
Why ?
dmpcal

Is performace gain when executes a stored procedure using the WITH RECOMPILE option ?
Why ?
As with many performance question, the answer is: It depends.
If the procedure includes statements the best plan depends heavily on the input parameters, WITH RECOMPILE can help. Although, these days it is better to use the query hint OPTION (RECOMPILE) for that particular query.
But if the procedure is complex, the compilation time may overshadow execution time and WITH RECOMPILE will cost performance.
Erland Sommarskog, SQL Server MVP, [email protected]

Similar Messages

  • Get a better performance with audio interface?

    Hi all.
    I got a silly question. I run Logic 7 on a Powerbook G4 12" with 1.33 Ghz and (only) 768 Mb RAM and a (hard to built in) 160 Gb Harddrive.
    After having often the "Disk is too slow..." message I was thinking about buying a external audio interface (presonus firebox).
    My simple question is: does an external audio interface give Logic (or CoreAudio) more free DSP-Power? So that maybe I got a better performance? Or do I need to get something like a TC Powercore Firewire?
    An answer would be great ... otherwise I have to sell my nice and small 12incher and buy an Intelbook ...
    Thanks. Frank.
    Powerbook G4 12" 1.33 Ghz Mac OS X (10.4.8)

    A new audio interface won't help.
    You should start by getting an external 7200 RPM Firewire drive to put your audio on. That right there will help, especially with higher track counts.
    Secondly, get more RAM. 768 of RAM is borderline for Logic. I always recommend at least 1 gig. Go ahead and max out that Powerbook to 2 gig of RAM, if you can.
    All that to say, an Intel mac would be a huge improvement over the Powerbook. And if you do go that route, still get an external FW drive for audio, and as much RAM as you can afford.

  • Application performance with "-mt" option

    I am seeing an improved performance when my application is compiled with "-mt". But there is something strange that my program don't use thread features. The same program compiled without "-mt" slower 10%. My program generate 10 million integers and sort the values using quick sort algorithm. (The sorting algorithm borrowed from free BSD's qsort.c)
    compiler: Workshop Compiler 4.2
    OS: SunOS 5.6
    system: sun4u sparc SUNW,Ultra-1
    $ CC -O a.C # normal compilation
    $ /bin/time a.out
    real: 1:18.3
    user: 1:16.3
    sys : 0.5
    $ CC -O -mt a.C # compile with "-mt"
    $ /bin/time a.out
    real: 1:10.2
    user: 1.08.3
    sys : 0.5
    a.C program don't use system call, memory allocation, I/O operation, multi-thread. It simply compare & move integers and use recursion.
    What happen if I use "-mt" option with non-threaded application ?

    The call itself is pretty cheap. If you are calling it from a middle tier, though, it may increase the "chatter" generated by the application. If you're generating an extra packet every second to make these calls, there could be an issue.
    Justin
    Distributed Database Consulting, Inc.
    http://www.ddbcinc.com/askDDBC

  • Better performance with lower CPU_COUNT ?

    Hi
    We had big performance problems on our production database, after lots of tunings and configuration we still didn't reach anywhere (oltp takes 9 hours when it take 20 minutes on test db).
    production server Dual Xeon 2.8 2GB RAM RAID5 10k/s RPM Win 2k3.
    Test db Intel 2.8 1 GB RAM 1 IDE disk Win 2000 Pro.
    I've noticed the difference between the test db and the production db was the gets/s the production db was NOT using any CPU and I/O was almost null unlike the test db which the gauges were showing 100 % full on everything.
    so stupidly we set the CPU_COUNT to 1 and PARALLEL_THREAD_PER_CPU = 1 when it was 2 x 4.(test db 1x2 unchanged)
    surprisingly the response time went diving from 9 hours to just 7 minutes.
    we check load balancing on CPUs and HT option, everything's fine.
    now we are confused!! ANY suggestions is apreciated, has anyone else encountered such a problem???
    Oracle 9i Rel2 9.2.0.1.0 on Win 2003 Standard.
    Thanks in Advance
    Regards
    Tony G.

    Hi
    The article on asktom.com was very usefull.
    We will 1st patch the oracle version to the certified version on Win 2K3 (9.2.0.1.3.0 or higher), and then set the nescessery parameters (PARALLEL_ADAPTIVE_MULTI_USER, PARALLEL_AUTOMATIC_TUNING, etc..).
    My question: do i have to alter the underlaying tables to parallel and gather stats on those tables in order the Cost Base Optimizer works correctly? and will this alteration affect other queries?
    Any suggestion or comment is more than welcomed, Thanks in Advance
    Regards
    Tony G.

  • 90 watt psu or 65 watt psu/ Better performance with USB devices?

    The 20Volt 65W psu that came with my 3000 N200 0769aku laptop is replaceable with
    a 20 Volt/90W Lenovo AC adapter.
    I am connecting devices to all usb ports (two hard drives, wireless mouse, ILOK software dongle) and  a two port firewire express card.
    Would having a "beefier" psu help cause when plugging all this in at once?
    Thanks

    If you need it, I don't believe so. If a 90W would hurt - no way.
    Andy  ______________________________________
    Please remember to come back and mark the post that you feel solved your question as the solution, it earns the member + points
    Did you find a post helpfull? You can thank the member by clicking on the star to the left awarding them Kudos Please add your type, model number and OS to your signature, it helps to help you. Forum Search Option T430 2347-G7U W8 x64, Yoga 10 HD+, Tablet 1838-2BG, T61p 6460-67G W7 x64, T43p 2668-G2G XP, T23 2647-9LG XP, plus a few more. FYI Unsolicited Personal Messages will be ignored.
      Deutsche Community     Comunidad en Español    English Community Русскоязычное Сообщество
    PepperonI blog 

  • What MacBook Pro will I get better performance with the 2.4ghz 256 SSD or the 2.6ghz 128 SSD these are also refurbished models

    I'll wanting to buy one of these these are refurbished and cannot choose please help me thanks :)

    Refurbished 13.3-inch MacBook Pro 2.4GHz Dual-core Intel i5 with Retina Display
    Originally released October 2013
    13.3-inch (diagonal) Retina display; 2560-by-1600 resolution at 227 pixels per inch
    8GB of 1600MHz DDR3L SDRAM
    256GB Flash Storage1
    720p FaceTime HD camera
    Intel Iris Graphics
    Apple Certified Refurbished
    OR
    Refurbished 13.3 inch MacBook Pro 2.6Ghz Dual-Core Intel i5 with Retina
    Originally released October 2013
    13.3-inch (diagonal) Retina display; 2560-by-1600 resolution at 227 pixels per inch
    8GB of 1600MHz DDR3L SDRAM
    128GB Flash Storage1
    720p FaceTime HD camera
    Intel Iris Graphics
    Apple Certified Refurbished
    Which one is the better one, they are the same  price what one will run faster and what one would you buy
    Many Thanks :)

  • How can I get better performance with OSX on my iMac G3?

    I just recently got a G3 iMac running at 400mhz with 512 mb of RAM, and it has OS 10.4.10. I am relatively new to Macs; its been a few years since I've used one.
    So, it actually runs surprisingly well, but I'm wondering if theres a way to sort of tone the visuals to make it run a little faster. I loaded windows XP pro onto an older dell that I just had sitting around, and was able to tone down the visuals enough to make it run very well on a 500 mhz machine with 256 mb of RAM, so I'm sure if I can tone down XP enough for an old computer, I can do the same with OSX.

    Welcome to Apple Discussions.
    You are somewhat limited by the 400MHz processor speed. Increasing your RAM will be helpful. Look at these links
    52 Ways to Speed Up OS X
    http://www.imafish.co.uk/articles/post/articles/130/52-ways-to-speed-up os-x/
    Tuning Mac OS X Performance
    http://www.thexlab.com/faqs/performance.html
    11 Ways to Optimize Your Mac's Performance
    http://lowendmac.com/eubanks/07/0312.html
    The Top 7 Free Utilities To Maintain A Mac.
    http://mac360.com/index.php/mac360/comments/thetop_7_free_utilities_to_maintain_amac/
    Mac OS X: System maintenance
    http://discussions.apple.com/thread.jspa?messageID=607640
     Cheers, Tom

  • Is there any way to get a better performance with this Xquery ?

    I created two tables.
    One has XBRL docoument with XMLType based on Clob and I'm going to select several columns from it and insert them into the other table.
    I used this SQL using XMLTable.
    INSERT INTO SV_XBRL_ELEMENT
    SELECT r.finance_cd,r.base_month,xt.context_id,xt.ns,xt.name,nvl(xt.lang,'na')
    as lang,xt.unit,xt.decimals,xt.value
    FROM SV_XBRL_DOC r,
    XMLTABLE(
    XMLNAMESPACES(
    'http://www.w3.org/1999/xlink' AS "xlink",
    'http://www.xbrl.org/2003/linkbase' AS "link",
    'http://www.xbrl.org/2003/instance' AS "xbrli",
    'http://www.xbrl.org/2003/iso4217' AS "iso4217",
    'http://www.xbrlkorea.com/kr/kisinfo/fr/gaap/ci/2007-02-09' AS "kisinfo-ci",
    'http://www.xbrlkorea.com/kr/kisinfo/fr/gcd/2007-02-09' AS "kisinfo-gcd",
    'http://www.xbrlkorea.com/kr/kisinfo/fr/profile/2007-02-09' AS "kisinfo-profile",
    'http://www.xbrlkorea.com/kr/kisinfo/fr/ratio/2007-02-09' AS "kisinfo-ratio",
    'http://www.xbrlkorea.com/kr/kisinfo/fr/common/scenario' AS "kisinfo-scenario",
    'http://www.xbrl.or.kr/kr/fr/gaap/ci/2006-05-31' AS "kr-gaap-ci",
    'http://www.xbrl.or.kr/kr/fr/common/pte/2006-05-31' AS "krfr-pte",
    'http://www.xbrl.or.kr/kr/fr/common/ptr/2006-05-31' AS "krfr-ptr",
    'http://www.xbrl.or.kr/2006/role/subitem-notes' AS "p0",
    'http://xmlns.oracle.com/xdb' AS "ora"),
    for $item in
    $doc/xbrli:xbrl/*[not(starts-with(name(),"xbrli:")) and
    not(starts-with(name(),"link:"))]
    where $item/@contextRef
    return <item contextRef="{$item/@contextRef}" xml:lang="{$item/@xml:lang}"
    unitRef="{$item/@unitRef}" name="{local-name($item)}" ns="{namespace-uri($item)}">{$item}</item>'
    PASSING r.xbrl as "doc"
    COLUMNS context_id varchar2(128) PATH '@contextRef',
    ns varchar2(128) PATH '@ns',
    name varchar2(128) PATH '@name',
    lang varchar2(2) PATH '@xml:lang',
    unit varchar2(16) PATH '@unitRef',
    decimals varchar2(64) PATH '@decimals',
    value varchar(256) PATH '.'
    ) xt
    SV_XBRL_DOC has 1450 records(1450 documents).
    SV_XBRL_ELEMENT has more than 110,280 record (110,280 elements)
    the sql above takes more than 6000 seconds with my machine (10g,sun5.8 4 cpus), I admit it's a big number of records, I'm looking for more efficient way to do.
    Is there any point to boost the sql ?
    here is the part of the xbrl document sample.
    <?xml version="1.0"?>
    <xbrli:xbrl xmlns:xbrli="http://www.xbrl.org/2003/instance" xmlns:link="http://www.xbrl.org/2003/linkbase" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:fines-b-ot="http://fss.xbrl.or.kr/kr/br/b/ot/2007-06-30" xmlns:fines-aa001="http://fss.xbrl.or.kr/kr/br/gaap/aa001/2007-06-30" xmlns:xbrldt="http://xbrl.org/2005/xbrldt" xmlns:ref="http://www.xbrl.org/2004/ref" xmlns:xbrldi="http://xbrl.org/2006/xbrldi" xmlns:iso4217="http://www.xbrl.org/2003/iso4217">
    <link:schemaRef xlink:type="simple" xlink:href="http://fss.xbrl.or.kr/kr/br/fines/aa001/2007-06-30/fines-aa001-2007-06-30.xsd"/>
    <xbrli:context id="ctx_AAA001C">
    <xbrli:entity>
    <xbrli:identifier scheme="http://fss.xbrl.or.kr">0010002</xbrli:identifier>
    <xbrli:segment>
    <xbrldi:explicitMember dimension="fines-b-ot:newItemDimension">fines-b-ot:AAA001C</xbrldi:explicitMember>
    </xbrli:segment>
    </xbrli:entity>
    <xbrli:period>
    <xbrli:instant>1999-03-31</xbrli:instant>
    </xbrli:period>
    </xbrli:context>
    <xbrli:context id="ctx_BAA001C">
    <xbrli:entity>
    <xbrli:identifier scheme="http://fss.xbrl.or.kr">0010002</xbrli:identifier>
    <xbrli:segment>
    <xbrldi:explicitMember dimension="fines-b-ot:newItemDimension">fines-b-ot:BAA001C</xbrldi:explicitMember>
    </xbrli:segment>
    </xbrli:entity>
    <xbrli:period>
    <xbrli:instant>1999-03-31</xbrli:instant>
    </xbrli:period>
    </xbrli:context>
    <xbrli:context id="ctx_CAA001C">
    <xbrli:entity>
    <xbrli:identifier scheme="http://fss.xbrl.or.kr">0010002</xbrli:identifier>
    <xbrli:segment>
    <xbrldi:explicitMember dimension="fines-b-ot:newItemDimension">fines-b-ot:CAA001C</xbrldi:explicitMember>
    </xbrli:segment>
    </xbrli:entity>
    <xbrli:period>
    <xbrli:instant>1999-03-31</xbrli:instant>
    </xbrli:period>
    </xbrli:context>
    <xbrli:context id="ctx_DAA001C">
    <xbrli:entity>
    <xbrli:identifier scheme="http://fss.xbrl.or.kr">0010002</xbrli:identifier>
    <xbrli:segment>
    <xbrldi:explicitMember dimension="fines-b-ot:newItemDimension">fines-b-ot:DAA001C</xbrldi:explicitMember>
    </xbrli:segment>
    </xbrli:entity>
    <xbrli:period>
    <xbrli:instant>1999-03-31</xbrli:instant>
    </xbrli:period>
    </xbrli:context>
    <xbrli:unit id="KRW">
    <xbrli:measure>iso4217:KRW</xbrli:measure>
    </xbrli:unit>
    <fines-b-ot:AAA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">14</fines-b-ot:AAA001R>
    <fines-b-ot:AAA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">0</fines-b-ot:AAA001R>
    <fines-b-ot:AAA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">14</fines-b-ot:AAA001R>
    <fines-b-ot:AAA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:AAA001R>
    <fines-b-ot:A1AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">7</fines-b-ot:A1AA001R>
    <fines-b-ot:A1AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">0</fines-b-ot:A1AA001R>
    <fines-b-ot:A1AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">7</fines-b-ot:A1AA001R>
    <fines-b-ot:A1AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:A1AA001R>
    <fines-b-ot:A2AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">7</fines-b-ot:A2AA001R>
    <fines-b-ot:A2AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">0</fines-b-ot:A2AA001R>
    <fines-b-ot:A2AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">7</fines-b-ot:A2AA001R>
    <fines-b-ot:A2AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:A2AA001R>
    <fines-b-ot:BAA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">4788</fines-b-ot:BAA001R>
    <fines-b-ot:BAA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">49</fines-b-ot:BAA001R>
    <fines-b-ot:BAA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">4837</fines-b-ot:BAA001R>
    <fines-b-ot:BAA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">30</fines-b-ot:BAA001R>
    <fines-b-ot:B1AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">4788</fines-b-ot:B1AA001R>
    <fines-b-ot:B1AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">48</fines-b-ot:B1AA001R>
    <fines-b-ot:B1AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">4836</fines-b-ot:B1AA001R>
    <fines-b-ot:B1AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">29</fines-b-ot:B1AA001R>
    <fines-b-ot:B11AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">2317</fines-b-ot:B11AA001R>
    <fines-b-ot:B11AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">21</fines-b-ot:B11AA001R>
    <fines-b-ot:B11AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">2338</fines-b-ot:B11AA001R>
    <fines-b-ot:B11AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">2</fines-b-ot:B11AA001R>
    <fines-b-ot:B111AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">0</fines-b-ot:B111AA001R>
    <fines-b-ot:B111AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">0</fines-b-ot:B111AA001R>
    <fines-b-ot:B111AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">0</fines-b-ot:B111AA001R>
    <fines-b-ot:B111AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:B111AA001R>
    <fines-b-ot:B112AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">60</fines-b-ot:B112AA001R>
    <fines-b-ot:B112AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">2</fines-b-ot:B112AA001R>
    <fines-b-ot:B112AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">62</fines-b-ot:B112AA001R>
    <fines-b-ot:B112AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:B112AA001R>
    <fines-b-ot:B113AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">185</fines-b-ot:B113AA001R>
    <fines-b-ot:B113AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">1</fines-b-ot:B113AA001R>
    <fines-b-ot:B113AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">186</fines-b-ot:B113AA001R>
    <fines-b-ot:B113AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:B113AA001R>
    <fines-b-ot:B114AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">408</fines-b-ot:B114AA001R>
    <fines-b-ot:B114AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">5</fines-b-ot:B114AA001R>
    <fines-b-ot:B114AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">413</fines-b-ot:B114AA001R>
    <fines-b-ot:B114AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:B114AA001R>
    <fines-b-ot:B115AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">1664</fines-b-ot:B115AA001R>
    <fines-b-ot:B115AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">13</fines-b-ot:B115AA001R>
    <fines-b-ot:B115AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">1677</fines-b-ot:B115AA001R>
    <fines-b-ot:B115AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">2</fines-b-ot:B115AA001R>
    <fines-b-ot:B12AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">2471</fines-b-ot:B12AA001R>
    <fines-b-ot:B12AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">27</fines-b-ot:B12AA001R>
    <fines-b-ot:B12AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">2498</fines-b-ot:B12AA001R>
    <fines-b-ot:B12AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">27</fines-b-ot:B12AA001R>
    <fines-b-ot:B2AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">0</fines-b-ot:B2AA001R>
    <fines-b-ot:B2AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">0</fines-b-ot:B2AA001R>
    <fines-b-ot:B2AA001R decimals="0" contextRef="ctx_CAA001C" unitRef="KRW">0</fines-b-ot:B2AA001R>
    <fines-b-ot:B2AA001R decimals="0" contextRef="ctx_DAA001C" unitRef="KRW">0</fines-b-ot:B2AA001R>
    <fines-b-ot:B3AA001R decimals="0" contextRef="ctx_AAA001C" unitRef="KRW">0</fines-b-ot:B3AA001R>
    <fines-b-ot:B3AA001R decimals="0" contextRef="ctx_BAA001C" unitRef="KRW">1</fines-b-ot:B3AA001R>
    </xbrli:xbrl>

    Using 11g along XML indexes can greatly improve this XQuery. We can discuss further offline.
    Regards,
    Geoff

  • Better performance with higher resolution with ultra..why ?

    how is this possible ? I got a q6600, 8gb ram and a msi ultra8800.Running crysis with everything to max on 1024X768 at 32 fps, and 1280X1024 at 40 ?
    weird, any ideas as to why thats happening ?
    cheers

    me again...sorry for bothering
    got another question concerning my ultra, and might as well post it here....Is this h2o cooling compatible with my msi 8800 ultra ?
    http://www.virtual-hideout.net/reviews/Mega_Watercooling_Roundup/Zalman_Reserator_2/index.shtml
    sorry if its a stupid question, but im really really REALLY a noob here
    cheers for any info

  • Best performance with forms 4.5 than 6i ?!!!

    Hi,
    Since i'd upgrade my app with forms 6i, i find that when i make a query, this is more slow with 6i than 4.5, with the same db(8i).
    Does anyone else had the same problem?
    What could i do to resolve the problem?
    Thanks.
    null

    See my earlier posting 'Speeding load time of WEB Forms50'. I
    have had no takers yet on a solution to this. There may be some
    hints for you there.
    Good luck on this one - Forms over the WEB seems to be very
    processor and memory hungry. We would like to support about 200
    users on a VERY simple search screen, but we need about 2Gbyte
    of physical memory on the server to get any amount of useful
    speed out of it. Each Forms session takes up about 10Mbyte on
    the server - as soon as this starts swapping out to disk then
    your run-time performance goes right downhill. The earlier-
    mentioned posting gives some hints on speeding up the startup
    time, and it does work.
    Regards,
    Jason Judge
    Natasha Murillo (guest) wrote:
    : How I can Get better performance with Forms 4.5 and 6.0 on the
    : WEB???
    : It is very slow and my customers are not very happy with
    that...
    : Please help me...
    null

  • Bottom line, does 8800 ultra sli setup ensure better performance ?

    greetings all, first time poster here !
    anyways, just got my new rig in town the other day, quad core q6600, 4 mg ram, msi 8800 ultra OC, vista 32bit..thing is mad.Have a question tho, putting the already insane performance bar aside, if money isnt an object, would another ultra result in an obvious performance boost, or are we speaking only marginal advantages here, like early slis 6600gt for example, sli setup was only 25/30 percent faster than one card performance.
     thanks for every answer in advance
    cheers

    Quote
    so you say, the bigger the resolution the better performance with ultra ?
    Well the performance is not better the higher the resolution. The ADVANTAGE compared to a slower card or single card to SLi is higher.
    Quote
    How far can you reckon can this be pushed
    Very doubtful it will do 700mhz gpu. At least nobody can tell you what one of those vgas does. It's a matter of the individual sample. You will only find out by testing. But the possibility you kill it that way is very little. If you keep an eye out for the temps and those don't get past 100°C you're fine.
    If you set gpu clock too high it will show with artifacts or bluescreens in 3d. So if you lower the clock when experiencing such behavior you won't damage your card. Of course you shouldn't set way too high clocks. I'd recommend not to set any higher than 700mhz unless this proves to be rockstable. Mem clock shouldn't set higher than 2300 but doubtful it will make it.
    DON'T TRUST MSI UTILITY! Set clocks manually and check after any step. Many of those GeForce 8 especially oc versions or high end models don't even do stock settings. So best would be checking the card first at stock settings and then starting to oc.

  • GroupingCollection2 with better performance

    I created a GroupingCollection2 with better performance (in
    comparision with GroupingCollection) by tweaking the
    implementation. You can find a sample/swc here
    http://flexpearls.blogspot.com/2008/06/groupingcollection-with-some-better.html
    If you have some free time please test it and let me know if
    you find any bugs !
    Of-course if you like the performance and would like to use
    it in your app, feel free :)

    Thanks for your input Grant; it's very helpful
    Just to summarize my next steps.
    - since a MAC Pro can't handle the new 6G SSDs with SATAIII, I'd have to go with the SSD 3G.
    I'll proably get the 115 GB or 240GB model – any thoughts which one I should get
    - I'll get an additional hard drive; perhaps: 2.0TB Western Digital 'Caviar BLACK' 7200RPM 6Gb/s HDD 64MB Cache  – is this a model you'd recommend
    Which one of these should be my boot drive? On the other drive I'll have all my big Photoshop and InDesign job files; applications should be on the boot drive; correct?
    - I'll get more RAM. I could return all of my 8 1GB pieces and get 6 x 4GB = 24 GB RAM. Or I can keep 4 of my old and get 4 x 4 GB RAM = 20 GB RAM
    I never seen posts were 20 GB RAM was tested. I wonder if it makes a huge difference if I get 20 or 24, since in tests the performance difference between 16 and 24 doesn't seem to be very much. Any thoughts?

  • Better Performance on Win7 64-bit with Crysis, GTA IV?

    Hi. Has anyone tried the Boot Camp 3.2 on Windows 7 64-bit? How is the performance with Crysis/Warhead, GTA IV and similar taxing games? Any fps improvements? Thank you in advance. Gbu.

    If you want to reduce memory(60%) '''when firefox minimize[only in windows os] '''there is an easy old good trick:
    type in the address bar '''about:config'''
    then press the button '''i'll be careful i promise'''
    then '''right-click''' anywhere in the list of preferences and choose '''new''' > '''boolean'''
    then type in the open window '''config.trim_on_minimize'''
    then click''' OK''', and in the next window select '''TRUE''' and click '''OK'''.
    exit firefox and restart-it.
    '''works only in windows os'''
    see for more info :
    [http://kb.mozillazine.org/About:config About:config]
    [http://kb.mozillazine.org/Config.trim_on_minimize Config.trim on minimize]
    thank you
    Please mark "Solved" the answer that really solve the problem, to help others with a similar problem.

  • Performance with LinkedList in java

    Hello All,
    Please suggest me any solution to further improve the performance with List.
    The problem description is as follows:
    I have a huge number of objects, lets say 10,000 , and need to store the objects in such a collection so that if i want to store an object at some particular index i , I get the best performance.
    I suppose as I need indexed based access, using List makes the best sense as Lists are ordered.
    Using LinkedList over ArrayList gives the better performance in the aforementioned context.
    Is there are way I can further improve the performance of LinkedList while solving this particular problem
    OR
    Is there any other index based collection using that i get better performance than LinkedList?
    Thanks in advance

    The trouble with a LinkedList as implemented in the Java libraries is that if you want to insert at index 100, it has no option but to step through the first 100 links of the list to find the insert point. Likewise is you retrieve by index. The strength of the linked list approach is lost if you work by index and the List interface gives no other way to insert in the middle of the list. The natural interface for a linked list would include an extended Iterator with methods for insert and replace. Of course LinkedLists are fine when you insert first or last.
    My guess would be that if your habitual insertion point was half way or more through the list then ArrayList would serve you better especially if you leave ample room for growth. Moving array elements up is probably not much more expensive, per element, than walking the linked list. Maybe 150% or thereabouts.
    Much depends on how you retrieve, and how volatile the list is. Certainly if you are a read-mostly situation and cannot use an iterator then a LinkedList won't suit.

  • In PI 7.1 better performance is reached using RFC or Proxy?

    Hello Experts,
    As with PI 7.1 which one would be better option to have better performance?
    1)Proxy which goes through the Integration Engine by omiting Advance adaptor Engine
    2)RFC which goes through the AAE by omiting Integration Engine
    As we know there are alot of advantages of Proxies over RFC:
    1. Proxy communication always by passes the Adapter Engine and will directly interact with the application system and Integration engine. So it will give us better performance.
    2. Proxies communicate with the XI server by means of native SOAP calls over HTTP.
    3. Easy to handle messages with ABAP programming if it is ABAP Proxy .
    4. Proxy is good for large volumes of data. we can catch and persist the errors ( both system & application fault ) which was generated by Proxy setting.
    Thanks in Advance
    Rajeev

    Hey
    More than the performance,its a question of requirement.
    There are several restrictions which you need to consider before using AAE.To name a few
    IDOC,HHTP adapter wont be available
    No support for ABAP mapping
    No support for BPM
    No support for proxy
    No support for Multimapping,content based routing ( in first release)
    So if you want to use any of the above you cant use AAE in first place.but performance is significantly improved,upto 4 times better that simple AE-IE
    /people/william.li/blog/2008/01/10/advanced-adapter-engine-configuration-in-pi-71
    check the above blog and the article mentioned in it.
    Now coming to proxy,it supports all the above and performance is not that bad either.
    so it all boils down to what your requirements are:)
    Thanks
    Aamir

Maybe you are looking for