64-bit VL

[bigpaws]

The question is this:  would a larger user community bring some fresh blood into the developer community?  Is that something the core developers even want?

From the growth I have seen vs the increase of developers it is a very wide margin. More developers
would be a welcome asset to Vector Linux. The problem as noted before is that there are more requests
and even less contribution.

May it be that you are overlooking the resources of Ubuntu and Fedora. These are HUGE projects, as well as Debian and Gentoo. Vector does not compare to those resources. What has put Vector on the map per se is the way things are done here. Ideas are tossed out, then the conversation starts usually to verify that the project is feasible. Then it hashed again, and when the result is a go then the person that proposed it, gets the job. No one here is limited in participation. All can join without needing to "know" someone. So for those that are making these requests take the project. The core
developers will teach you all the tricks and gotchas. The beauty of open source in its' prime.

Bigpaws

[easuter]

Another problem to consider is that stable Gambas's code isn't 64-bit clean yet.
That means that the Gambas-based tools written for VL 5.8 wouldn't work on a 64 bit build.

Here is an example from a Debian bug-report for their AMD64 build, February this year: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=411257

I think Gambas2 makes some 64bit provisions, but have no idea when it will be ready (feels like an eternity).

[The Headacher]

FWIW, the issue I personally complained about was internationalization/localization, and I promptly went to work building packages to address that.

And it is greatly appreciated. I personally don't need i18n, but I'm sure VL could use some more work there and you seem to know what you're doing in that department. All the things you've done lately for VL is a lot more than what I have, and I wish more people had the same attitude as you do.

I know I might seem opposed to changes at times, but I'm not. I'm just afraid of people who come here for a while, then their needs change and suddenly it seems to them that "VL needs to change" (accordingly of course). While the distro itself still has everything that made them come here in the first place.

I don't want VL to change that what makes it VL (great community, slack based, faster and lighter than most other distro's). But then again, it's not up to me to decide what way VL goes. And that's a good thing, I'm just another user.

[GrannyGeek]

My main desktop has a 64-bit processor. I don't have much interest in running a 64-bit OS.

As far as I can discover, I wouldn't gain anything with a 64-bit OS unless what I'm running needed more than 4 gigs of RAM. My CPU is not dual core. What I'm most interested in is compatibility with the programs I run now.

Hearing about people who won't consider VectorLinux because it doesn't have a 64-bit version reminds me of those people who won't consider VectorLinux because it doesn't meet their standards for prettiness.
--GrannyGeek

[nightflier]

There is a lot of life left in 32 bit machines. I know I have been able to get several more years out of mine since switching to Linux. Still, it is only a question of time before 64 bit becomes the standard. One can be on the bleeding edge or wait until the path is better defined and smoother.

Personally, I'm happy to wait a while.

[caitlyn]

64-bit gear and 64-bit distros have been around for a couple of years at least.  I'd hardly call that "bleeding edge" any longer.

As some of you know I wrote a very positive review of Vector Linux 5.8 SOHO yesterday for O'Reilly.  One of the questions in the comments already is asking if there is going to be a 64 bit version.  I pointed that person to this thread.  Folks, the demand is there.

[nightflier]

Just read that review. Very nice. 

If you go 64-bit, do you have to re-do all the packages? If maintaining two versions, would you need different repos?

[Toe]

In theory, all that's required is a recompile.  I'm not sure how close reality follows this, though.

[easuter]

In theory, all that's required is a recompile.

Yes, then after that comes testing and maintenace...

Honestly, it really gets to me when people won't try something because they think they know it won't work. Example, if someone is complaining about Windows and I sugest they try Linux, they may ask:

- So does Linux have MSN?

You mean MSN messanger? Yeah, you can communicate with your palls on the MSN network, but using a different app. Pidgin or AMSN are two MSN-compatible instant messangers for Linux.

-No, I mean I want "normal" MSN. If I can't use "normal" MSN, I'm not even gonna try Linux!

I understand that this is a very simplified example, but if someone is going to refuse to try a 32bit distro like VL simply because they have the all-mighty new 64bit hardware, then that person is really pretty impervious to reason...

Sorry if it sounds harsh, but unless someone actually benchmarks VL versus some other 64bit distro on 64bit hardware, and proves that VL suffers from a monumental performance loss on that hardware, then I guess its not really viable to make a 64-bit (official) port at this moment. (not my call of course )

[alec]

Yes, I admit VL is faster than any other distro out there. But knowing that, it could be even faster in 64-bit!

I benefit a lot from 64, but I do much numbercrunching scientific work. Also all kinds of compile and archiving is faster. Python too is much snappier in 64.
Other than that, its all about autosuggestion of perceivable difference, because there is none. But you never know which next program may benefit a lot from 64. And 4 gig is really becoming mainstream nowadays.

On the positive side, all 64-bit arch is new, so you don't have  to support older i486 machines and get a small boost from there too.
And new Opera 9.5 is coming with x64 linux packages! And OOo will eventually break through too.

If at least compiling/packaging/building part can be automated, Vector has many testers to find bugs. So perhaps its not double amount of work.

[Joe1962]

Yes, I admit VL is faster than any other distro out there. But knowing that, it could be even faster in 64-bit!

Actually, it seems that it can't be faster:

The fact is that most operations on a 64bit OS are actually slower than a 32bit OS on the same hardware.
There are many benchmark results on the web that show this.

[Toe]

The answer to the performance question is pretty much 'depends'.  If you need to crunch a lot of big numbers, like for scientific data or some types of image processing, then yes it's probably going to be faster.  If you need more than 4GB of RAM, then obviously yes you want 64bit.  But it does have some drawbacks.  For example, moving from 32 to 64 bit, you've pretty much cut your CPU's cache size in half, which affects performance.

[SummerNight]

I am the one of the people who had asked Cait about a 64bit VL. I was getting rid of a previous distro and I had listened to her rave about VL so I asked the question. I did go with Ubuntu because I do get a noticeable improvement on my projects which are very memory intensive.

On small scale projects, there is very little difference between the two architectures but the system can most definitely handle more complex projects and that means being able to do a bigger chunk at once rather than to keep churning out render after render an then hope I can splice everything in at the end using an editor. That can be hours to days of extra work. Of course I am but a beginner right now and as my knowledge grows (hopefully along with my artistic abilities LOL), the demand that I place on a given system will grow as well.

I do agree with the notion that many people out there will not see a huge difference, if any. Depending on what they do and what they chose to do it with, they could see a decrease. However, there are many out there who will want and desire a 64bit distro because that is what they have at the moment. They view going to a 32bit environment as a step down and the consumer does not like to step down.

I did google some benchmarks and did see where 32bit ranked faster on a number of them. A couple of the reviewers knew that the codecs and drivers being used at the time did not have a 64bit equivalent and noted that.

http://en.wikipedia.org/wiki/64-bit#Pros_and_cons

Thanks to everyone for the info. If VL ever puts out a 64 distro, I will definitely give it a try.

[cintyram]

I recently bought a new computer. AMD Athlon 4400+ X2,  64bit
but the OS it came with? Microsoft Windows Vista Home Premium ( 32-bit)
[ why do they have to have such long names? compare with Leopard, Solaris10,VL5.8 ]

And The highest RAM i could get in a home machine was 2GB, shared with the graphics card,.

So for most people buying new machines, they may get processors with 64bits but the memory and OS are still not stellar.

Personally i think it would be amazing to have a 64 bit version of VL.  but really lets be pragmatic here, if i had more time to contribute, my priority is to build and test packages.  But if some one else can give me a script to run, i will gladly donate my time and my machine's time.

This is not just a statement. some time back i had written a script to build Enlightenment packages and ran into some hardware issues, so i gave it to others, and the packages i got back were good. soon the E17 team released their own script, and then others on the web released them too.

So just between the people on this thread we already have atleast 6 64 bit machines.
Here is my advice to those who want 64 bit VL.
1. start a poll and ask how many people can donate time on their machine, and donate time to test a 64 bit package.
2. if you see enough response, get the build scripts together, and share them with the volunteers. Again remember you dont have to write the scripts yourself from scratch, look hard enough and youll find them.  in most cases it is just a change of some flags.  and VL is a fairly well automated build.  you just need to get the slack dependencies right [ and the scripts for that].

3. pretty soon all the volunteers will buzz with activity and you will have a testable alpha build.  seed that to more testers and 64bit VL would have taken its baby steps .

I am happy that some one took the initiative to bring this topic to discussion.  Now one more small step needs to be taken, actually three small steps, listed above.
cheers
ram

[rbistolfi]

Even when I think the arguments against the 64bit VL are strong, I will give my time for testing and installing, and if someone guides me, for packaging. I have one 64bit box, VL runs great on it. I am planning to install some 64bits sys to test, since I do some graphics and I think the 64bit will improve performance, almost in rendering video. I think a 64bit box is really needed by just very few people. When I work on my graphics seriously, I use a render farm made with several pentium III, in the house of a friend. I save my files, and he renders. A professional designer who have a 8gig of ram machine, and has the needs of use them would be a target for the 64bit VL. I think they are just a few. A server edition of VL is not developed atm, so the possible users are even lesser.
I agree with cintyram about would be a great system, but I agree with the others about the 32bit version is the priority.
The needs of 64bits cpu in the low end market is just generated by the publicity, and is not a real need. Most of the people dont know what a hell that means, and they are selling it because the price difference is little enough.
The argument "is what people wants" is not good enough, imho. I dont know if we want to be considered a major distro, but yes I know we want to be considered the ones who do the rigth thing, and just because is the rigth thing. Of course, which one is the rigth thing will come out, probably, from topics like this one, so, I am happy that someone brings this to discussion too...