Ok, I got it. Thanks for all.

regards, Jean-Lin

Selon Andi Kleen ak@suse.de:

On Wednesday 28 June 2006 13:08, Jean-Lin Pacherie wrote:

...

Is it politycaly correct to ask here if the new "Core 2 Duo" processors

will

...

be suported by SuSE 64bit ?

I planed to build a server with AMD 64 x2 processor and / or Opteron 275

but

...

the announces of the "Core 2 Duo" chnaged my mind. In any cases, I will

keep

...

SuSE on this new workstation but I need the best processor power.

Any clue ?

x86 CPUs are pretty good at compatibility so there is very rarely trouble with new CPUs with Linux at least in basic operation. All the x86 CPU vendors have Linux in their regression tests so if there was a problem they will fix it in the hardware long before the CPUs are released.

Some advanced hardware specific functions like performance counter profiling support or machine check decoding might need special support, but the machine is perfectly usable even without that.

If anything you might have trouble with whatever new motherboard you're using with it.

But Core 2 plugs into already shipping board, so if you chose a board that already works under Linux you should be on the safe side.

That said nearly all (even not yet shipping) boards work, although you could sometimes get a different impression from this list. That's mostly because people who don't have any problems rarely post here.

-Andi

-- Ce message ne contient pas de virus connu. neoDomaine Postmaster - http://www.neodomaine.com/

On Thursday 29 June 2006 00:59, Hans du Plooy wrote:

...

Is it politycaly correct to ask here if the new "Core 2 Duo" processors will be suported by SuSE 64bit ?

Not sure about politically correct or not, but I can give you the answer to the question :-)

No, because the Core Duo is not a 64bit processor.

He has asked about Core 2 Duo processors, which will support x86_64 (aka. EM64T) standard.

Selon Hans du Plooy koffiejunkielistlurker@koffiejunkie.za.net:

On Thu, 2006-06-29 at 09:05 +0200, Michal Schulz wrote:

...

He has asked about Core 2 Duo processors, which will support x86_64 (aka. EM64T) standard.

My bad, my apologies. I didn't notice the "2" - didn't even know about it. Looks good, but it doesn't look much different from the Core Duo, EMT64 and the shared L2 cache aside.

Hans

As far as I read on web, the new "Core 2 Duo" introduces tremendous changes related to performance, heating and price. That's the reasons why I'm still waiting for the release.

And if the promises and pre-benchmarks were too optimistic, then at least I will benefit from the new prices of the AMD processors :)

I must say that I didn't see any comparison of the new Core 2 Duo with the Opteron dual core. Only x2 (including the FX-60) where used in the comparative benchmarks.

What I heard about the new AM2 solution from AMD didn't need much attention i far as i understood.

Again, I'm not hardware expert and all I know on the subject is from googleling around. So I might be wrong.

Regards, Jean-Lin

On Thu, 2006-06-29 at 11:22 +0200, Jean-Lin Pacherie wrote:

As far as I read on web, the new "Core 2 Duo" introduces tremendous changes related to performance, heating and price. That's the reasons why I'm still waiting for the release.

Well, I suppose if you compare Core 2 Duo (Pentium-M derived, from the looks of it) with the current Xeon (Pentium-4 based), the improvements are certainly significant. I think the Pentium-M is the best Intel has, far far better than anything based on the Pentium-4 core. Sure, the huge clock speeds do help for some things, but for most uses, and especially servers that do many different things, I find even the entry level AMD chips to perform far better. Two examples:

I set up ISPconfig on two machines. One a dual Xeon 3ghz, 1mb cache each server, proper Intel server board, 2gb ram, the other a desktop class box, VIA based mini-atx board with a 1.6ghz Sempron (the 64bit version, although I didn't know this when I installed so I loaded 32bit), with 1gb RAM. Guess what? The Sempron was a good 5 minutes quicker (both machines had websites and mail on, roughly the same load).

The other example to me proves that the Pentium-3 really is a better multitasking chip than anything based on Pentium-4. A client has a Windows terminal server, and around 20 linux destops that are really just there for rdesktop. The terminal server is a dual 2ghz HT Xeon, 1mb cache per CPU, 2gb ram. The box didn't cope, load was almost permanently at 100%. They thought there was something wrong with the box, so they replaced it with another server that had completely different, but similar spec hardware. No joy.

Someone sold them two secondhand dell servers, dual 1.1ghz P-3 with 1gb ram each. They replaced the terminal server with these two Dells, set them up identically, and put half of the clients on each. This worked great, with around 20% load on each. After a few months, the one server's motherboard failed, so the users on it were moved to the other server. So now we have all the users on *one* dual P-3. Guess what? It's coping marvellously. It shows around 60% CPU usage at the heaviest of times. The stuff that the clients use are MS Office, MS Great Plains, and two locally designed accounting packages, one of which tends to make the most modern desktops feel sluggish.

Anyway, to end my ramble, I think that the P-4 based CPUs have never been the best option on the mark for any server work. I still continue to see Pentium-3 based boxes run circles around P-4 ones when the load gets heavy, not to mention the AMDs. I am really glad that Intel is finally moving the server chips over to the Pentium-M based design. It's far more powerful, and far more cool and quiet.

And if the promises and pre-benchmarks were too optimistic, then at least I will benefit from the new prices of the AMD processors :)

The only ones I've seen use Intel supplied boxes, using pre-release chips, and the testers couldn't really fiddle with the box. I'm interested to see how they compare once the hardware sites can test with off-the-shelf kit.

Hans

On Mon, 2006-07-03 at 08:03 -0400, Curt Purdy wrote:

I was hoping 10.1 32-bit would run on my new Mac OSX 17" Duo Core laptop. After running BootCamp to install XP (only to repartition so I could install SuSE), I tried to boot from the live-cd version to test before installing and it would load to the boot choice screen, sit for a while, and then die.

Is anyone able to make it run on Duo Core?

In Issue 63 of Linux User&Developer magazine (www.linuxuser.co.uk) there is an article explaining why it does not work (yet). I only skimmed over it, so my understanding is sketchy at best, but it basically comes down to this.

The new Apples don't have a traditional PC bios, but instead uses EFI. Windows (and most, if not all operating systems built for PC hardware) don't support EFI, and expect to be able to make calls to the BIOS. With Bootcamp you also get a firmware update for your Mac which adds a classic BIOS interface of sorts to the Mac. Also, drivers to deal with the slightly funky hardware.

Now I assume ordinary SUSE also expects to see a BIOS on your PC. There's also an issue with the hard discs, as the Macs don't use an ordinary mbr style partition table like we're used to.

I think there are only a few things to fix for normal linux distros to cope with the new Macs, but the current ones won't work.

Hans