Socket AM2 motherboards -- best chipset for Linux, BSD?
Igor
I'm putting together an AMD Athlon 64 X2 based PC. I plan on running
(dual-booting) both Windows 2000 and one of the Unix-based OS's
(either one of the more user-friendly Linux distros or something like
PCBSD or Desktop BSD) on this machine.
I'm currently trying to select a motherboard. I've decided to go with
a socket AM2 type, since I intend to use DDR2 RAM. However, I'm having
trouble choosing a chipset, especially in relation to Linux/BSD. Are
some chipsets less likely to cause problems with Linux and BSD than
others, or are they all equally well supported?
I've heard that nVidia graphics cards are better supported under Linux
than ATI ones; is this true of nVidia chipsets as well?
Thanks for your time.
--
"Those of us whose brains did not die in college are
actually stunned by just how stupid academic ideas
are." -- Robert W. Whitaker, http://readbob.com/
R. C. White
A year ago I built my new system for Vista, starting with this mobo:
http://www.epox.com/USA/product.asp?id=EP-MF570SLI
EPoX is not well known, but this is my 3rd mobo from them and all have
worked well for me. This one cost me $138 from Newegg last December; Newegg
now says it is a "Deactivated Item". EPoX says the price now "starts from
$89.99". This one is AM2, with the NVIDIA nForce 570 SLI chipset. But you
probably should look around the EPoX website to find a newer model than my
year-old one. Mobos change fast these days!
Last December, I paid Newegg $689 for the mobo, an AMD Athlon 64 X2 5000+
CPU (boxed with HSF), and 2 GB OCZ Platinum DDR2-800 (PC6400 SDRAM). I'm
sure they would cost much less today, but I haven't priced them.
This was for upgrading my existing system (Athlon 64 3200+), so I already
had hard drives, optical drives, etc. But my ATI AIW was AGP and this mobo
has no AGP slots, so I also had to buy a PCI-e graphics card; I chose an ATI
Radeon X1600 Pro. (I'm neither a gamer nor a videophile.)
I've never run Linux/BSD or any other non-Windows OS, at least since the
1990s, so I can't comment on that aspect. But I've dual-booted many
versions of Windows and Vista, including as many as 8 at one time (on the
prior mobo) during the Vista beta period. Now I've settled down to just
Ultimate x64 and only occasionally dual-boot into x86 or WinXP.
RC
--
R. C. White, CPA
San Marcos, TX
r...@grandecom.net
Microsoft Windows MVP
(Running Vista Ultimate x64 SP1 beta on EPoX MF570sli)
"Igor" <donot...@nospam.com> wrote in message
news:svqgh353dn7gr9ijn...@4ax.com...
kony
<donot...@nospam.com> wrote:
>Howdy.
>
>I'm putting together an AMD Athlon 64 X2 based PC. I plan on running
>(dual-booting) both Windows 2000 and one of the Unix-based OS's
>(either one of the more user-friendly Linux distros or something like
>PCBSD or Desktop BSD) on this machine.
>
>I'm currently trying to select a motherboard. I've decided to go with
>a socket AM2 type, since I intend to use DDR2 RAM. However, I'm having
>trouble choosing a chipset, especially in relation to Linux/BSD. Are
>some chipsets less likely to cause problems with Linux and BSD than
>others, or are they all equally well supported?
>
>I've heard that nVidia graphics cards are better supported under Linux
>than ATI ones; is this true of nVidia chipsets as well?
>
>Thanks for your time.
Since we can assume everything is supported by windows, did
you take any time to research what is supported by 'nix?
Did you go to the respective chipset manufacturers' websites
to see if there were 'nix drivers, or what the distro you
prefer to use, *natively* supports?
Igor
>Since we can assume everything is supported by windows, did
>you take any time to research what is supported by 'nix?
Some would say posting this question to usenet is a part of that
research.
Thanks for the link.
Arno Wagner
> Howdy.
> I'm putting together an AMD Athlon 64 X2 based PC. I plan on running
> (dual-booting) both Windows 2000 and one of the Unix-based OS's
> (either one of the more user-friendly Linux distros or something like
> PCBSD or Desktop BSD) on this machine.
> I'm currently trying to select a motherboard. I've decided to go with
> a socket AM2 type, since I intend to use DDR2 RAM. However, I'm having
> trouble choosing a chipset, especially in relation to Linux/BSD. Are
> some chipsets less likely to cause problems with Linux and BSD than
> others, or are they all equally well supported?
> I've heard that nVidia graphics cards are better supported under Linux
> than ATI ones; is this true of nVidia chipsets as well?
> Thanks for your time.
I have an Asus M2N32 WS Professional. Drivers and such
are ok, but ithere are some issues. One is that the
chipset gets far, far too hot IMO. And the sensor chip
lies about the temperature and pretends it is about 15C
colder. I find that highly suspicuous. The other is that
the external SATA connector seems not to have hot-plug.
Oh, and it is unreliable, except with the newest BIOS.
I would advise against getting this board. Not Asus-like
quality. Pretty bold to label this "professional".
Seems to be marketing lies at work.
Arno
Igor
<snip>
>
>I have an Asus M2N32 WS Professional. Drivers and such
>are ok, but ithere are some issues. One is that the
>chipset gets far, far too hot IMO. And the sensor chip
>lies about the temperature and pretends it is about 15C
>colder. I find that highly suspicuous. <snip>
If that's true, that does indeed sound very underhanded.
Igor
wrote:
>Hi, Igor.
>
>A year ago I built my new system for Vista, starting with this mobo:
>http://www.epox.com/USA/product.asp?id=EP-MF570SLI
>
>EPoX is not well known, but this is my 3rd mobo from them and all have
>worked well for me. <snip>
Honestly, this is the first time I've heard of them. If I see their
stuff listed in any of the catalogs that I'll be ordering from, I'll
remember the good word you put in for them.
At this point, however, I've decided to select a board more by
features, price, and availability (and compatibility with Linux/BSD)
than by brand. I was initially thinking in terms of brand, but after
going through the user reviews at newegg.com, I realized there were an
equal number of horror stories about *every* manufacturer and gave up
on that approach.
Over the years, I've owned PCs with motherboards by MSI, ECS, and
AOpen and never had problems with any of them. A relative of mine
recently replaced a computer she had gotten in '99. There was nothing
wrong with it, she just wanted something more powerful. That one had
an Asus motherboard in it.
So maybe I'm wrong, but I'm thinking that all manufacturers must
produce their share of defective or poorly designed boards, and that
they're probably all fairly equal quality-wise (as long as you're
comparing apples with apples, that is).
Arno Wagner
> <snip>
>>
>>I have an Asus M2N32 WS Professional. Drivers and such
>>are ok, but ithere are some issues. One is that the
>>chipset gets far, far too hot IMO. And the sensor chip
>>lies about the temperature and pretends it is about 15C
>>colder. I find that highly suspicuous. <snip>
> If that's true, that does indeed sound very underhanded.
Well, it is displayed temperature in the BIOS vs. an infrared
thermometer directly on top of the chipset cooler. I am fairly
sure my measurements are correct.
Arno
Arno Wagner
I think you are right on this one. I have made
good experiences with Epox, until I needed to repair 4
boards and had problems with BIOS settings vanishing on
occasion with 22 others. A real bother and I will not get
Epox again. Some people believe Asus can do no wrong, but
I have observed otherwise. Server-boards are not that much
better: I have had to send an MSI board back, because of poorly
(visilby !) mounted components. I used to like Tyan, but
not anymore for much the same reason. Best choice is probably
to get a board a bit older, say on the market for a year or so.
Then you should at least be able to google its faults and can
decide whether they are still acceptable to you.
Arno
kony
wrote:
>In comp.os.linux.hardware Igor <donot...@nospam.com> wrote:
>> On 20 Oct 2007 04:28:48 GMT, Arno Wagner <m...@privacy.net> wrote:
>
>> <snip>
>
>>>
>>>I have an Asus M2N32 WS Professional. Drivers and such
>>>are ok, but ithere are some issues. One is that the
>>>chipset gets far, far too hot IMO. And the sensor chip
>>>lies about the temperature and pretends it is about 15C
>>>colder. I find that highly suspicuous. <snip>
>
>> If that's true, that does indeed sound very underhanded.
>
More likely it's just measuring the wrong (a different)
chip, or the bios has a bug as many do. It's not likely to
be "underhanded" in any way. I'd imagine Asus would rather
the system integrator can tell what's wrong instead of
causing addt'l RMAs due to the board seemingly
malfunctioning.
>Well, it is displayed temperature in the BIOS vs. an infrared
>thermometer directly on top of the chipset cooler. I am fairly
>sure my measurements are correct.
>
>Arno
The measurement may be correct, but the top of a chipset
cooler doesn't necessarily give usable data unless you
happened to know the exact C/W rating of the heatsink, and
that as-installed including the thermal interface
effectiveness.
More importantly, is it instable as a result? We don't have
a number from you, but a lot of people do jump to
conclusions about something being too hot when too hot
usually means above 70C (unless you're overclocking, making
peak stable operation more dependant on lower temp).
Igor
I trust your measurement. That's quite the discrepancy alright.
Faulty sensor? Software bug? Deliberate deception? I wonder which it
is.
kony
<donot...@nospam.com> wrote:
>On 21 Oct 2007 02:04:53 GMT, Arno Wagner <m...@privacy.net> wrote:
>
>>In comp.os.linux.hardware Igor <donot...@nospam.com> wrote:
>>> On 20 Oct 2007 04:28:48 GMT, Arno Wagner <m...@privacy.net> wrote:
>>
>>> <snip>
>>
>>>>
>>>>I have an Asus M2N32 WS Professional. Drivers and such
>>>>are ok, but ithere are some issues. One is that the
>>>>chipset gets far, far too hot IMO. And the sensor chip
>>>>lies about the temperature and pretends it is about 15C
>>>>colder. I find that highly suspicuous. <snip>
>>
>>> If that's true, that does indeed sound very underhanded.
>>
>>Well, it is displayed temperature in the BIOS vs. an infrared
>>thermometer directly on top of the chipset cooler. I am fairly
>>sure my measurements are correct.
>>
>
>I trust your measurement. That's quite the discrepancy alright.
>
>Faulty sensor? Software bug? Deliberate deception? I wonder which it
>is.
Who cares? you're acting as if this is important.
The only real factor is whether the chipset stays cool
enough to be stable, and if it does not, whether Asus would
rectify the situation or the owner would have to seek an
alternate chipset heatsink instead.
J.O. Aho
> I'm currently trying to select a motherboard. I've decided to go with
> a socket AM2 type, since I intend to use DDR2 RAM. However, I'm having
> trouble choosing a chipset, especially in relation to Linux/BSD.
No hard evidence, but I guess nVidia chipsets are those "best" supported
nowadays, but of course VIA and ATi/AMD chipsets works fine too, most of SiS
chipsets too, but there may be some that don't.
> I've heard that nVidia graphics cards are better supported under Linux
> than ATI ones; is this true of nVidia chipsets as well?
IMHO it depends on which way you go with the graphics card driver, open source
or closed source. In the war between AMD and nVidia the winner is
Open source class..: ATi/AMD
Closed source class: nVidia
When it comes to chipset, nVidia has got better on supporting the open source
solution (they dropped their closed source drivers).
I think I will go for a nVidia based card on my next 64bit machine (have that
on my laptop already), the only thing that can be a bit of a trouble is the
built in audio, I do suggest you get a SB Live or SB Audigy for audio and you
will be a lot happier.
--
//Aho
J.O. Aho
But thats the cooler temperature and not the chip temperature, those will be
different, the first one having a lower temperature.
--
//Aho
Igor
Well, if Asus are the type of company who would knowingly put a bad
product on the market and try to cover it up by rigging the software
(not that I'm saying that's the case), that would certainly affect my
decision to buy one of their boards.
kony
<donot...@nospam.com> wrote:
>>The only real factor is whether the chipset stays cool
>>enough to be stable, and if it does not, whether Asus would
>>rectify the situation or the owner would have to seek an
>>alternate chipset heatsink instead.
>
>Well, if Asus are the type of company who would knowingly put a bad
>product on the market and try to cover it up by rigging the software
>(not that I'm saying that's the case), that would certainly affect my
>decision to buy one of their boards.
You're prematurely jumping to conclusions.
The chipset will have same cooling requirement regardless of
manufacturer, judge only on whether their 'sink was
especially worse than anybody else's.
Arno Wagner
I think the heatsink is pretty good. It seems this chipset is just way
over the top in its power consumption.
Arno
Arno Wagner
Oh, you are right. The chip will be warmer than the cooler,
making the gap between displayed temperature and actual
chip temerature even larger.
As to instabilities, I don't know. I had several crashes
before I specifically tunneled outside air on the cooler
by the CPU. This reduced the temperature by about 10C,
something which the pretty good CPU cooler could not do.
Since then, no crashes.
Arno
Arno Wagner
Yes indeed. That makes the observed gap between displayed
chip temerature and actual chip temperature even larger.
After all the cooler is already 15C above the reported chip
temperature.
Arno
kony
wrote:
>I think the heatsink is pretty good. It seems this chipset is just way
>over the top in its power consumption.
I definitely wouldn't call it a cool running chipset but
there are quite a few boards w/same out there which don't
have this problem. I would suspect one of three things:
1) Poorer than average case ventilation
2) Specific model of board has inferior heatsink
3) Chipset voltage regulation subcircuit has a feedback
resistor that's a little out of spec, raising the voltage a
bit beyond what the target was.
I'd tend to suspect #1 the most since a duct helped.
Arno Wagner
I am pretty sure 2) and 3) are not the case. 1) might be,
since it is a pretty old case from an area where
ventilation was not that critical. It is also
isnulated to dampen noise. That may result in much worse
cooling than ist standard today.
Arno