Hi Jani, Thanks for your answer.
That is true that you guys are making some available. I sure don't deny that at all and it is very generous of you guys to spend the time doing it as well! One reason is that I wanted Apache with PHP and Static MySQL built in. So, I have to compile it anyway. So, may as well try to do the best I can! (:> But, also it is for some issue with OpenBSD that I saw in my research so far. Some posting in that regards point to some problem sometime with it. I wanted to check this. Here is an example of it on OpenBSD http://marc.theaimsgroup.com/?l=mysql&m=99970869525188&w=2 Also, I did read the doc here: http://www.mysql.com/documentation/mysql/bychapter/manual_Installing.html#co nfigure_options for the options you guys are using and base my stuff on it. But, as I am doing it on OpenBSD, you do not have a packages for it, but even if you did, I still would like to do the static in PHP, so I still would need to do part of it. Never the less OpenBSD have a great ports tree via CVS that make the process very easy to do. I wanted to find out why I am not getting ( that can be just a perception on my part as well) as big of a difference between two hardware platform. I also find specific information from MySQL website on OpenBSD here: http://www.mysql.com/documentation/mysql/bychapter/manual_Installing.html#Op enBSD_2.8 But, I did go over it quickly as it was for 2.8. So, to see it that was my imaginations, or something, I started to do some test. I tried various compiling version with native treads and mit-pthreads. Customer kernel, etc. I am not saying there is a bug or anything. I wouldn't do that. But I wanted to see if I could do better. One reason was that I didn't see a major difference in my test using a Cobalt RaQ2 server and a new VALinux1000 with 2Gb of ram in my test. Yes, it is better, but I wanted to find out more as I was expecting a bigger difference I guess. So, I really do appreciate your answer on this, but I was and still try to find why this is so. I don't think I will be successful, but I thought to guy it a shot. Final point in here is that, there is a remark in the CVS tree for OpenBSD, where the optimization for I686 and I586 for the OS and kernel was disable because of bug in GCC 2.95.3. I don't know more then that, but if so, I was thinking of trying the GCC 3.01 and see if that give different result as well. That was my reason for asking all that. May be it also could affect MySQL if compile with that optimization and running on OpenBSD. I Don't know they answer to that. Anyway, I still haven't found all the answer yet. But, I haven't give up on it. Now, just in case you are thinking why I am doing this on OpenBSD. It is for a very simple reason. I really wanted to get over the security issue and make my server setup once, spend the time I need and then be able to forget about it and only do patch when ever I have time instead of needing to do it because it is required. OpenBSD is the safest I found so far. I am welling to give up plenty of applications, or feature if need be to sleep more (:> But, that doesn't mean I shouldn't try to get the maximum of the box if you follow my thoughts. Many thanks for your feedback. Daniel PS: I you want to do test. I am sure you have hardware, but I just got 9 new boxes that I am playing with right now and that I will need to setup in the next few weeks. But I can make one available to you on the Internet if you want to do test as well. That's not a problem for me. But I am sure your time is very short so, I won't ask. This is just something that is bugging me and I need to find an answer for myself. Doesn't mean it will work, but I just need to understand it for myself. The only explication I can come up with so far, is that MySQL is heavy on treads and OpenBSD doesn't support it to well. But, please, don't flame me! I can't say it is the case. Just what I got so far as a logical explication of may be why. Key words here is MAY BE! > Hi Daniel, > > > I am looking into using optimization of the GCC 2.95.3 for I686 as > > > > -O3 -mpentiumpro > > > > to compile an optimize version of MySQL. > > > > But as I look around, I see that some user warned about using > this as they > > said there is bug in GCC for I686 optimization. > > > > Any information you can pass to me on that? Is it still true > for the 2.95.3. > > I know that the 2.95.2 had issues, but I want to check for the this. > > We've been using -O3 -mpentiumpro for Pentium processors for a long > time and have seen no problems with it. It should be safe on 2.95.2 > and 2.95.3. > > > Also, as I am on the subject, what about using the GCC 3.01. > Any better with > > good result so far. > > GCC 3.x has been working also without problems, but we don't have that > much experience of it yet. Currently we are using 2.95.3 for most things. > > > Any one? > > > > Thanks > > > > Daniel > > Regards, > > - Jani > > Ps. Why are you compiling MySQL on your own (just curious) ? If you > are only looking for optimization, I would strongly recommend you to > download one of our binaries, because they are very highly optimized, > and the risk of getting an unfunctioning binary (in high loads for > example) is low. If you need to compile one by yourself, please check > the MySQL manual for hints about compiling options for your platform. > > For technical support contracts, visit https://order.mysql.com/ > __ ___ ___ ____ __ > / |/ /_ __/ __/ __ \/ / Mr. Jani Tolonen <[EMAIL PROTECTED]> > / /|_/ / // /\ \/ /_/ / /__ MySQL AB, Full-Time Developer > /_/ /_/\_, /___/\___\_\___/ Helsinki, Finland > <___/ www.mysql.com > --------------------------------------------------------------------- Before posting, please check: http://www.mysql.com/manual.php (the manual) http://lists.mysql.com/ (the list archive) To request this thread, e-mail <[EMAIL PROTECTED]> To unsubscribe, e-mail <[EMAIL PROTECTED]> Trouble unsubscribing? Try: http://lists.mysql.com/php/unsubscribe.php
