Page 1 of 2 12 LastLast
Results 1 to 25 of 28
  1. #1

    Mysql 64bit performance in Xeon Vs Opteron

    Hi,

    Can anyone suggest me which server performs better (XEON Vs Opteron) on SuSe 9 with mysql (64bit) and what are the pros and cons of using each. I have heard that 64bit versions of mysql do not work properly on Xeon. And that Opteron heats up too much too rapidly. Are these facts?

  2. #2
    Join Date
    Apr 2005
    Location
    Atlanta, Georgia
    Posts
    521
    .... opterons run COOL compared to xeons, xeons are ovens

  3. #3
    Any Stats ? Comparisins ? Benchmarks ?

  4. #4
    Join Date
    Aug 2002
    Location
    Stony Plain, AB
    Posts
    607
    Opteron has always beaten Opteron in Mysql.... (as far as i know).

    And 64bit OS with Opterons is great.

    Opterons are alot cooler then Xeons.

    - Eddy
    Tired of jumping web hosting providers? Tired of OVERSELLING? Tired of Poor service/quality?
    Are you Finally realizing You Get What You Pay For In Life?
    If So...Please Visit ***** eServicesUnlimited***** - We Guarantee 100% Satisfaction we promise that!

  5. #5
    Join Date
    Nov 2004
    Location
    Miami FL
    Posts
    288
    Originally posted by sirgamesalot
    Opteron has always beaten Opteron in Mysql.... (as far as i know).


    - Eddy
    Ah yes, another proven fact Opteron beats Opteron (must be a dual CPU box to beat one another...

    I take it you mean Opteron always beats Xeon... (at least for mysql)

  6. #6
    Join Date
    Jan 2004
    Location
    North Yorkshire, UK
    Posts
    4,164
    Opteron all the way, and they're cheaper than the 64bit Xeons too (the actual chips). I've also heard that the new 64bit Xeons have heat issues ...
    █ Dan Kitchen | Technical Director | Razorblue
    █ ddi: (+44) (0)1748 900 680 | e: dkitchen@razorblue.com
    █ UK Intensive Managed Hosting, Clusters and Colocation.
    █ HP Servers, Cisco/Juniper Powered BGP Network (AS15692).

  7. #7
    Join Date
    Aug 2002
    Location
    Stony Plain, AB
    Posts
    607
    hahahah

    yeah ;P wrote that one a bit early.


    - Eddy
    Tired of jumping web hosting providers? Tired of OVERSELLING? Tired of Poor service/quality?
    Are you Finally realizing You Get What You Pay For In Life?
    If So...Please Visit ***** eServicesUnlimited***** - We Guarantee 100% Satisfaction we promise that!

  8. #8
    Why are Dual Opteron dedicated servers always more expensive than dual Xeons then (if the actual chips are cheaper)?

  9. #9
    Join Date
    Aug 2004
    Posts
    96
    it depends on the cost of hardware the provider is using, and there marketing and where your trying to get an opteron server from.

    opterons are the way to go IMO, iv been running AMD servers for along time and very happy with its performace.

  10. #10
    Join Date
    Jun 2003
    Location
    San Francisco
    Posts
    623
    i have never seen any actual benchmark comparison between Xeon 64 and Opteron 64 either.

    Would someone share a link?

    Yes, Xeon (retail) costs more than Opteron. However, you can almost always get a cheaper complete dual Xeon system from Dell than building a dual Opteron yourself.

  11. #11
    Join Date
    Feb 2005
    Location
    Montreal, PQ
    Posts
    355
    Xeons are cheaper because Intel has no choice but to sell them cheaper. Since Opterons totally smokes the Xeons.

    Though the low end Opterons (142, 242) are relatively inexpensive.

  12. #12
    A lot of people posted links to benchmarks in other posts but from speculation, it seems people are still wanting 'legit' benchmarks since the testers apparently don't have a good balance between hardware specs on the opteron and xeon servers they test.

    Edit: Here is a link to a '64-Bit Battle : Intel’s Xeon 3.4 GHz vs. AMD’s Opteron 250'. The servers are pretty evenly matched hardware wise. However, this shows a comparison for Apache, in which the Opterons beat out the Xeons, but loose out to the Xeons in multimedia. I'm still checking out for MySQL tests. http://www.gamepc.com/labs/view_cont...pteron&page=13
    Last edited by CyanoX; 07-23-2005 at 02:15 PM.

  13. #13
    Join Date
    Feb 2005
    Location
    Montreal, PQ
    Posts
    355
    AMD always seem to lose to Intel in multimedia, such as encoding.

    However gaming and applications seems to be a different story.

  14. #14
    Eh, I've given up the search for a benchmark comparing both 64 bit models of Xeon and Opterons in MySQL. Perhaps someone has already stumbled upon a nice demonstation and is willing to poast their link. The fact that Opteron beats out the Xeon (both 64 bit) in everything (in the tests performed), except multimedia, still stands. Can't seem to grasp a hold of any place that shows a comparison using MySQL and having comparable hardware specs for both servers...

  15. #15
    Join Date
    Aug 2004
    Posts
    96
    Originally posted by XF-Chan
    AMD always seem to lose to Intel in multimedia, such as encoding.
    well where talkin a small loss on this

    It takes me on my system
    AMD 64bit 3200 w/1MB Cache Socket 754
    2gigs DDR 400 Ram
    2 200gig Sata drives Raid 0

    2and half hours to encode a DVD movie to DivX

    Can intel do this in 45mins? no..lol
    i personaly wouldnt go by and intel just to encode
    a video in 2 hours

    also you need to have a pretty fast hard drive as well
    its not completly directed to CPU and Ram


    but none the less intel and amd both work good
    its just a matter of how far you are willing to go to tweak the heck out of your server / system to gain that extra second out of it..

  16. #16
    Join Date
    Feb 2005
    Location
    Montreal, PQ
    Posts
    355
    Tweaking won't help much.

    Encoding just works better on Intel's architechure.

  17. #17
    Join Date
    Feb 2001
    Location
    Solar Star system
    Posts
    183
    Originally posted by XF-Chan
    Tweaking won't help much.

    Encoding just works better on Intel's architechure.

    Yes and no.

    The software that is doin the "encoding" is better optimized to be used under INTEL.

    But since nobody here really need "encoding" - no point to go there anyway.

    N.

  18. #18
    Join Date
    Feb 2001
    Location
    Solar Star system
    Posts
    183
    Ok here is the benchmarks.

    Was very easy to find - there is ALOT of test's on http://anandtech.com

    Here is the test configuration:

    Processor(s):

    AMD Opteron 150 (130nm, 2.4GHz, 1MB L2 Cache)
    VS
    Intel Xeon 3.6GHz (90nm, 1MB L2 Cache)

    RAM:
    2 x 512MB PC-3200 CL2 (400MHz) Registered
    2 x 512MB PC2-3200 CL3 (400MHz) Registered

    Memory Timings: Default

    Operating System(s):
    SuSE 9.1 Professional (64 bit)
    Linux 2.6.4-52-default
    Linux 2.6.4-52-smp

    The link u wall wana see is here

    http://www.anandtech.com/linux/showdoc.aspx?i=2163&p=2

    For all people that dont belive this results - well setup a hardware try it your self.

    We did. And under BSD we even got better results. (also 64 BIT).

    Opterons are really BEST CPU's out there at the moment.

    Except for encoding ofcourse

    N.

  19. #19
    Join Date
    Feb 2004
    Posts
    633
    To followup NEMON's post, AnandTech very recently did another article comparing 32 & 64 bit versions of MySQL & DB2 on both Xeon & Opteron:

    http://www.anandtech.com/IT/showdoc.aspx?i=2447&p=5

    Regarding MySQL, here are a few quotes that might interest you:

    This is really remarkable, as the Xeon does not benefit from 64 bit at all. Worse, a 10% performance penalty is paid for moving over to 64 bit. The Opteron, however, thrives on 64 bit and gets a 30% boost from 64 bit.
    Interestingly once again, the dual core CPU is quite a bit faster than our Dual CPU (single core) machine. A 10% bonus is nothing to sneeze at, especially when you consider that server boards with only one socket are quite a bit cheaper. It seems that one dual core Opteron is an ideal solution for a rather powerful MySQL database server.
    Their methodology seems pretty solid, though I would have liked to have seen them also test using Intel's C compiler for Linux instead of just gcc. It seems to me that there historically has been a cycle between gcc and the Intel compiler, where the newest Intel features are better optimized in the Intel compiler, and it takes gcc some time to close the gap, but by then there are some newer Intel features and it starts over again. So it's possible that EMT64 support might be much better optimized in the Intel compiler right now. I know in the past the Intel compiler has been around 20-30% faster on many tests, so it would have been interesting to see it used.

  20. #20
    Originally posted by lockbull
    Their methodology seems pretty solid, though I would have liked to have seen them also test using Intel's C compiler for Linux instead of just gcc. It seems to me that there historically has been a cycle between gcc and the Intel compiler, where the newest Intel features are better optimized in the Intel compiler, and it takes gcc some time to close the gap, but by then there are some newer Intel features and it starts over again. So it's possible that EMT64 support might be much better optimized in the Intel compiler right now. I know in the past the Intel compiler has been around 20-30% faster on many tests, so it would have been interesting to see it used. [/B]
    AMD claims that the Intel compiler deliberately decreases performance if used with an AMD chip. http://www.amd.com/breakfree
    but I have no clue about compilers and I am not even sure if the one your talking about is the same one...

  21. #21
    Join Date
    Jun 2003
    Location
    San Francisco
    Posts
    623
    Originally posted by NEMON
    Ok here is the benchmarks.

    Was very easy to find - there is ALOT of test's on http://anandtech.com

    Here is the test configuration:

    Processor(s):

    AMD Opteron 150 (130nm, 2.4GHz, 1MB L2 Cache)
    VS
    Intel Xeon 3.6GHz (90nm, 1MB L2 Cache)

    RAM:
    2 x 512MB PC-3200 CL2 (400MHz) Registered
    2 x 512MB PC2-3200 CL3 (400MHz) Registered

    Memory Timings: Default

    Operating System(s):
    SuSE 9.1 Professional (64 bit)
    Linux 2.6.4-52-default
    Linux 2.6.4-52-smp

    The link u wall wana see is here

    http://www.anandtech.com/linux/showdoc.aspx?i=2163&p=2

    For all people that dont belive this results - well setup a hardware try it your self.

    We did. And under BSD we even got better results. (also 64 BIT).

    Opterons are really BEST CPU's out there at the moment.

    Except for encoding ofcourse

    N.
    Any good benchmark would include the complete system configuration. Curiously enough, this particular one doesn't include the HD, which is critical for database. If the AMD machine has SCSI and the Xeon one has SATA, it would have been a non-brainer than the AMD would outperform.

    Opteron is quite new. I know the reference motherboard has SCSI RAID. Xeon has been on the market for a long time. SATA motherboards are quite readily available. So the HD is a piece of information which should not have been omitted.

  22. #22
    Join Date
    Feb 2001
    Location
    Solar Star system
    Posts
    183
    Originally posted by riverpast
    Any good benchmark would include the complete system configuration. Curiously enough, this particular one doesn't include the HD, which is critical for database. If the AMD machine has SCSI and the Xeon one has SATA, it would have been a non-brainer than the AMD would outperform.

    Opteron is quite new. I know the reference motherboard has SCSI RAID. Xeon has been on the market for a long time. SATA motherboards are quite readily available. So the HD is a piece of information which should not have been omitted.
    Trust me it was not done like that.

    That would be totaly unfair. What u can do is next. Send a email to Kristopher Kubicki at kristopher@anandtech.com - and he will for sure get back to u with information.

    He is cool guy and usual replay on all emails

    Anyway, people at anandtech dont favor any CPU.

    Btw - OPTERON is not new CPU ok? That benchmark is done almost YEAR AGO! (August 12th, 2004)

    It's maybe new for u - but for me it's old thing already.

    I am sure u use only INTEL that's why u maybe dont have feeling what is "new" or "old".

    I also dont favor any CPU - I dont care for AMD or INTEL. I just wana best for my money and my clients.

    When my client ask me - "Ok we need FAST DUAL CPU SERVER FOR DATABSE WORK!" - I will not tell him go "DUAL INTEL XEON!" - cause that simple is not true. It's not fast as same Opteron server.

    Tomorow if thing change in favor of INTEL - will do the same thing.

    My people and me also really did ALOT of research on this subject - cause we wana to save money - since we have many clients that have sites that grow with time (dynamic content - forums etc) - we need best possible solution to save money. To get more for less if possible.

    And Opteron's are just that.

    N.

  23. #23
    Join Date
    Jun 2003
    Location
    San Francisco
    Posts
    623
    Originally posted by NEMON
    Trust me it was not done like that.

    That would be totaly unfair. What u can do is next. Send a email to Kristopher Kubicki at kristopher@anandtech.com - and he will for sure get back to u with information.

    He is cool guy and usual replay on all emails

    Anyway, people at anandtech dont favor any CPU.

    Btw - OPTERON is not new CPU ok? That benchmark is done almost YEAR AGO! (August 12th, 2004)

    It's maybe new for u - but for me it's old thing already.

    I am sure u use only INTEL that's why u maybe dont have feeling what is "new" or "old".

    I also dont favor any CPU - I dont care for AMD or INTEL. I just wana best for my money and my clients.

    When my client ask me - "Ok we need FAST DUAL CPU SERVER FOR DATABSE WORK!" - I will not tell him go "DUAL INTEL XEON!" - cause that simple is not true. It's not fast as same Opteron server.

    Tomorow if thing change in favor of INTEL - will do the same thing.

    My people and me also really did ALOT of research on this subject - cause we wana to save money - since we have many clients that have sites that grow with time (dynamic content - forums etc) - we need best possible solution to save money. To get more for less if possible.

    And Opteron's are just that.

    N.
    I am not a hosting company. I only have two servers, and I obsolutely favor Xeon, but only if it is a good deal from Dell. If you ask me which one is better given the same cost, I would say "I don't know".

    1 year is still "new", and I can bet their testing Opteron system has SCSI.

    Note, I ask this because another review (tomshardware, in another thread) has obvious problem with configuration - they used a testing system with the reference Opteron server board with SCSI RAID, and then IDE (not even SATA) with 2MB cache (not even 8MB cache) for everything else (P4, Xeon, Athleon...), and you have to read carefully into the story to notice that.

    I am not going to get another server, so I really don't care which one comes out ahead. However, any testing should have all configuration revealed (without asking), otherwise, I will assume it is biased.

    I am sure you do a lot of testing yourself, so if you don't mind, please share your result (with the configuration data). Meanwhile, we know there are others who swear their Xeon is faster than Opteron (especially on games, it seems). That shows nobody should only take your (or others') words for it. Do share your information, and complete information.

  24. #24
    Join Date
    Feb 2001
    Location
    Solar Star system
    Posts
    183
    Originally posted by riverpast
    I am not a hosting company. I only have two servers, and I obsolutely favor Xeon, but only if it is a good deal from Dell. If you ask me which one is better given the same cost, I would say "I don't know".

    1 year is still "new", and I can bet their testing Opteron system has SCSI.

    Note, I ask this because another review (tomshardware, in another thread) has obvious problem with configuration - they used a testing system with the reference Opteron server board with SCSI RAID, and then IDE (not even SATA) with 2MB cache (not even 8MB cache) for everything else (P4, Xeon, Athleon...), and you have to read carefully into the story to notice that.

    I am not going to get another server, so I really don't care which one comes out ahead. However, any testing should have all configuration revealed (without asking), otherwise, I will assume it is biased.

    I am sure you do a lot of testing yourself, so if you don't mind, please share your result (with the configuration data). Meanwhile, we know there are others who swear their Xeon is faster than Opteron (especially on games, it seems). That shows nobody should only take your (or others') words for it. Do share your information, and complete information.

    I agree with u that data regarding FULL list hardware should be posted.

    Also anandtech and tomshardware is not the only place on internet where u can check out this results. There is another sites as well.

    XEONS would be faster for encoding and game servers. That's all.

    For database and web server - opteron will be faster.

    Opteron is not around 1 year or so. Opteron is same as XEON for Intel.

    This are actually single CPU servers today called AMD64 and P4 - just with more L2 Cache on it - that's about only difference (have few others also).

    I am sure your XEONS works great! My XEONS servers also works great!

    At the moment we have more Intel based servers then AMD - but the thing is Opteron servers are also working GREAT! - but at the end of the day they also do more work then XEONS.

    I am not saying XEON Is rubish or whatever, but at the moment every single server that we build for our self is AMD based - and we recommend that to all our clients. We do have orders for XEONS or Intel based serves - and ofcourse we build them and whatever is need it around them, after all we are in this business cause of the money - not case we like some CPU more then other

    N.

  25. #25
    Join Date
    Jun 2003
    Location
    San Francisco
    Posts
    623
    Also anandtech and tomshardware is not the only place on internet where u can check out this results. There is another sites as well.
    The problem for me is nobody does a WIMP (Windows, IIS, MySQL, PHP) benchmark testing between Xeon and Opteron. I am not interested in LAMP (Linux, Apache, MySQL, PHP).

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •