From dmccormick at wvmcc.com Wed Jan 4 12:46:14 2006 From: dmccormick at wvmcc.com (David McCormick) Date: Wed, 04 Jan 2006 07:46:14 -0500 Subject: Lockup with modprobe Message-ID: <43BBC396.6060401@wvmcc.com> I think I have seen an answer to this in the past but can't find it in the archives. Installing ndiswrapper 1.7-0 on a Fedora Core 4 64 bit box all goes well until I run modeprobe ndiswrapper at which point the machine locks up and I have to reboot. I have it running on a 32 bit laptop with no problems also FC-4. Both machines are ADM based. Can someone point me to the thread about this or give me a pointer on avoiding the lockup? Thanks David From maurice at harddata.com Thu Jan 5 17:49:58 2006 From: maurice at harddata.com (Maurice Hilarius) Date: Thu, 05 Jan 2006 10:49:58 -0700 Subject: SuperMicro H8SSL-i (ServerWorks HT1000) Competitor to Sun offering In-Reply-To: <200512071656.jB7Gupoi003448@bach.leonora.org> References: <200512071656.jB7Gupoi003448@bach.leonora.org> Message-ID: <43BD5C46.2090803@harddata.com> Good day all! Recent discussions have been seen here regarding the SuperMicro H8SSL-i board, and also the new Sunfire 2100 I would like to point out the 1U server offered by Hard Data Ltd. using the new SuperMicro Serverworks board. All aluminum chassis, with optional 4 x S-ATA2 hotswap trays. Complete equipped and built system weighs under 20 lbs (9Kg) Airflow and cooling is significantly better than competing offerings. With AMD Opteron S939 dual core 2.4GHz CPU the CPU temperature stays below 45C at room ambient temperature of 21C. All components are 3 year warranty, both from Hard Data Ltd. AND from the manufacturers. Contrast that with the 1 year, VENDOR ONLY warranty from competitors. Also, add up the prices with such options as rack rails, faster CPUs, IPMI 2 option and the price comparison is compelling, to say the least! http://www.harddata.com/icefire Thanks! -- With our best regards, Maurice W. Hilarius Telephone: 01-780-456-9771 Hard Data Ltd. FAX: 01-780-456-9772 11060 - 166 Avenue email:maurice at harddata.com Edmonton, AB, Canada http://www.harddata.com/ T5X 1Y3 From Philip.R.Schaffner at nasa.gov Fri Jan 6 13:50:14 2006 From: Philip.R.Schaffner at nasa.gov (Phil Schaffner) Date: Fri, 06 Jan 2006 08:50:14 -0500 Subject: SuperMicro H8SSL-i (ServerWorks HT1000) Competitor to Sun offering In-Reply-To: <43BD5C46.2090803@harddata.com> References: <200512071656.jB7Gupoi003448@bach.leonora.org> <43BD5C46.2090803@harddata.com> Message-ID: <1136555414.15300.120.camel@wx1.larc.nasa.gov> On Thu, 2006-01-05 at 10:49 -0700, Maurice Hilarius wrote: > Good day all! > > Recent discussions have been seen here regarding the SuperMicro H8SSL-i > board, and also the new Sunfire 2100 > > I would like to point out the 1U server offered by Hard Data Ltd. using Is is safe to assume that spammers will be unsubscribed? From b.j.smith at ieee.org Fri Jan 6 14:15:35 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Fri, 06 Jan 2006 09:15:35 -0500 Subject: SuperMicro H8SSL-i (ServerWorks HT1000) Competitor to Sun offering In-Reply-To: <1136555414.15300.120.camel@wx1.larc.nasa.gov> References: <200512071656.jB7Gupoi003448@bach.leonora.org> <43BD5C46.2090803@harddata.com> <1136555414.15300.120.camel@wx1.larc.nasa.gov> Message-ID: <1136556935.4705.0.camel@bert64.oviedo.smithconcepts.com> On Fri, 2006-01-06 at 08:50 -0500, Phil Schaffner wrote: > Is is safe to assume that spammers will be unsubscribed? Not a spammer, but an active member of this list just giving notice that their H8SSL-i product has become available. Many people, including myself, have been awaiting the arrival of this product. -- Bryan J. Smith mailto:b.j.smith at ieee.org http://thebs413.blogspot.com ------------------------------------------ Some things (or athletes) money can't buy. For everything else there's "ManningCard." From andy at andydeano.co.uk Fri Jan 6 22:04:08 2006 From: andy at andydeano.co.uk (Andy Dean) Date: Fri, 6 Jan 2006 22:04:08 +0000 Subject: opteron 146 thoughts on motherboards Message-ID: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> Hi am looking at getting an opteron 146, and would like some input with regards to motherboards, whether it's worthwhile going for a tyan tomcat k8e or just for a asus nvidia sli board of some description, or something else if any one has some recomedations. Mainly to be used as a desktop system really. Just wanting to get the best that i can with the budget i have and would value your input. TIA Andy Dean From cochranb at speakeasy.net Sat Jan 7 02:59:55 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Fri, 06 Jan 2006 21:59:55 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> Message-ID: <43BF2EAB.2000100@speakeasy.net> The archives of this list has a considerable amount of discussion about motherboards and Opteron processors. Spend a little time reading them. I wanted to upgrade to a dual core processor myself but had cost concerns. And it was a bit confusing reviewing the available motherboards. Someone suggested an Asrock 939Dual-SATA2 motherboard, and with the v1.5 BIOS update it is quite nice, I like the board. You might want to look at it. A big advantage is that you can recycle your existing memory sticks and video card to the Asrock. Bob Cochran Andy Dean wrote: > Hi > > am looking at getting an opteron 146, and would like some input with > regards to motherboards, whether it's worthwhile going for a tyan > tomcat k8e or just for a asus nvidia sli board of some description, or > something else if any one has some recomedations. Mainly to be used as > a desktop system really. Just wanting to get the best that i can with > the budget i have and would value your input. > > TIA > > Andy Dean > From cochranb at speakeasy.net Sat Jan 7 03:03:10 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Fri, 06 Jan 2006 22:03:10 -0500 Subject: Does NASM Do 64 Bit Assemblies? Message-ID: <43BF2F6E.5090106@speakeasy.net> I have nasm installed on my Fedora Core 4/AMD Athlon 64 X2 4400 box. I'd like to play with it, although I suppose I might be better off playing with C. Will nasm automatically default to 64 bit addressing? I was given Randall Hyde's "The Art of Assembly Language" book and finally noticed it is a 32 bit version. Is there a better (open source) assembler for the Athlon 64 X2? Thanks Bob Cochran From b.j.smith at ieee.org Sat Jan 7 05:35:17 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Sat, 07 Jan 2006 00:35:17 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <43BF2EAB.2000100@speakeasy.net> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> Message-ID: <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> On Fri, 2006-01-06 at 21:59 -0500, Robert L Cochran wrote: > A big advantage is that you can recycle your existing memory sticks You can do that on _all_ Socket-754 and 939 mainboards. The 184 and 368 traces, 754 and 939 respectively, run directly to the CPU, and devide their clock from it. This is not front side bottleneck (FSB) Intel technology where the memory clock affects the rest of the system. So you can use DDR200 (PC1600), DDR266 (PC2100), DDR333 (PC2700) and DDR400 (PC3200). The only issue I've seen is maybe the number of DIMMs you can use. JEDEC specs specify 3 DIMMs per DDR200 channel, 2 DIMMs per DDR266/333 channel and 1 DIMM per DDR400 channel. > and video card to the Asrock. Now that's one of the neat designs of the ULI M1695/1567. They have a _true_, _segmented_ AGP 3.0** slot, whereas some of the other chipset/cheap designs just connect the AGP to the shared, 32-bit PCI bus. That's in addition to the PCIe x16 option. [ **NOTE: It must be newer 1.5/0.8V AGP x4/8 card. You can't use older 3.3V AGP x1/2/4 cards. ] Of course, I don't know how well the peripherals of the ULI M1567 are, so be _wary_ of possible Linux compatibility issues. -- Bryan J. Smith mailto:b.j.smith at ieee.org http://thebs413.blogspot.com ------------------------------------------ Some things (or athletes) money can't buy. For everything else there's "ManningCard." From loony at loonybin.org Sat Jan 7 05:34:24 2006 From: loony at loonybin.org (Peter Arremann) Date: Sat, 7 Jan 2006 00:34:24 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> Message-ID: <200601070034.24798.loony@loonybin.org> On Saturday 07 January 2006 00:35, Bryan J. Smith wrote: > Of course, I don't know how well the peripherals of the ULI M1567 are, > so be _wary_ of possible Linux compatibility issues. I have that board and the only thing that does not work in the current FC4 install is the lm sensors... that requires an additional patch that's floating around out there somewhere... Peter. From b.j.smith at ieee.org Sat Jan 7 05:53:46 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Sat, 07 Jan 2006 00:53:46 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <200601070034.24798.loony@loonybin.org> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> <200601070034.24798.loony@loonybin.org> Message-ID: <1136613226.4774.30.camel@bert64.oviedo.smithconcepts.com> On Sat, 2006-01-07 at 00:34 -0500, Peter Arremann wrote: > I have that board and the only thing that does not work in the current FC4 > install is the lm sensors... that requires an additional patch that's > floating around out there somewhere... Okay, cool. I've been so locked into nVidia (with the occasional AMD8131/8132 mixed in) as of late that I haven't looked at much else. -- Bryan J. Smith mailto:b.j.smith at ieee.org http://thebs413.blogspot.com ------------------------------------------ Some things (or athletes) money can't buy. For everything else there's "ManningCard." From loony at loonybin.org Sat Jan 7 05:59:05 2006 From: loony at loonybin.org (Peter Arremann) Date: Sat, 7 Jan 2006 00:59:05 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <1136613226.4774.30.camel@bert64.oviedo.smithconcepts.com> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601070034.24798.loony@loonybin.org> <1136613226.4774.30.camel@bert64.oviedo.smithconcepts.com> Message-ID: <200601070059.05759.loony@loonybin.org> On Saturday 07 January 2006 00:53, Bryan J. Smith wrote: > On Sat, 2006-01-07 at 00:34 -0500, Peter Arremann wrote: > > I have that board and the only thing that does not work in the current > > FC4 install is the lm sensors... that requires an additional patch that's > > floating around out there somewhere... > > Okay, cool. I've been so locked into nVidia (with the occasional > AMD8131/8132 mixed in) as of late that I haven't looked at much else. I can't recommend this board for a server use or a high end setup. But if you are looking for a cheap way to upgrade this board is unbeatable. At about $70 shipped you get something that will allow you to reuse your AGP card and then allow you to go PCI-E whenever you feel like it... Peter. From cochranb at speakeasy.net Mon Jan 9 01:23:45 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Sun, 08 Jan 2006 20:23:45 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> Message-ID: <43C1BB21.3000501@speakeasy.net> Bryan J. Smith wrote: >On Fri, 2006-01-06 at 21:59 -0500, Robert L Cochran wrote: > > >>A big advantage is that you can recycle your existing memory sticks >> >> > >You can do that on _all_ Socket-754 and 939 mainboards. The 184 and 368 >traces, 754 and 939 respectively, run directly to the CPU, and devide >their clock from it. This is not front side bottleneck (FSB) Intel >technology where the memory clock affects the rest of the system. > >So you can use DDR200 (PC1600), DDR266 (PC2100), DDR333 (PC2700) and >DDR400 (PC3200). The only issue I've seen is maybe the number of DIMMs >you can use. JEDEC specs specify 3 DIMMs per DDR200 channel, 2 DIMMs >per DDR266/333 channel and 1 DIMM per DDR400 channel. > > > >>and video card to the Asrock. >> >> > >Now that's one of the neat designs of the ULI M1695/1567. They have a >_true_, _segmented_ AGP 3.0** slot, whereas some of the other >chipset/cheap designs just connect the AGP to the shared, 32-bit PCI >bus. That's in addition to the PCIe x16 option. > >[ **NOTE: It must be newer 1.5/0.8V AGP x4/8 card. > You can't use older 3.3V AGP x1/2/4 cards. ] > >Of course, I don't know how well the peripherals of the ULI M1567 are, >so be _wary_ of possible Linux compatibility issues. > > > I stand corrected over the memory, then. I've also changed my 3-pin PLED connector to a 2 pin configuration as suggested earlier simply by pulling one pin out and putting it in the middle position of the connector shell. I hadn't realized one can do that. Now my power LED works. I'm starting to wonder how to crimp those little connectors for other electronics projects. Thanks Bob From cochranb at speakeasy.net Mon Jan 9 01:34:15 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Sun, 08 Jan 2006 20:34:15 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <200601070059.05759.loony@loonybin.org> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601070034.24798.loony@loonybin.org> <1136613226.4774.30.camel@bert64.oviedo.smithconcepts.com> <200601070059.05759.loony@loonybin.org> Message-ID: <43C1BD97.1000804@speakeasy.net> Peter Arremann wrote: >On Saturday 07 January 2006 00:53, Bryan J. Smith wrote: > > >>On Sat, 2006-01-07 at 00:34 -0500, Peter Arremann wrote: >> >> >>>I have that board and the only thing that does not work in the current >>>FC4 install is the lm sensors... that requires an additional patch that's >>>floating around out there somewhere... >>> >>> >>Okay, cool. I've been so locked into nVidia (with the occasional >>AMD8131/8132 mixed in) as of late that I haven't looked at much else. >> >> >I can't recommend this board for a server use or a high end setup. But if you >are looking for a cheap way to upgrade this board is unbeatable. At about $70 >shipped you get something that will allow you to reuse your AGP card and then >allow you to go PCI-E whenever you feel like it... > >Peter. > > > Why isn't it a good server board? I'm not trained in how to evaluate a board that can be used for server applications. I installed an Athlon 64 X2 4400 "processor in a box" on this board. AMD puts in a heat sink that has nice looking copper heat pipes on one side. The heat sink can take a little try, try again to fit onto the mating black hold-down assembly that surrounds the CPU socket, and you want the pipes to be on the side where the ATX 12v connector is, not on the northbridge side. Now I just need to figure out what to do with my old mainboard....it is still perfectly good. It's sitting in a nice antistatic bag for the time being. Bob Cochran From loony at loonybin.org Mon Jan 9 01:52:58 2006 From: loony at loonybin.org (Peter Arremann) Date: Sun, 8 Jan 2006 20:52:58 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <43C1BD97.1000804@speakeasy.net> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601070059.05759.loony@loonybin.org> <43C1BD97.1000804@speakeasy.net> Message-ID: <200601082052.58788.loony@loonybin.org> On Sunday 08 January 2006 20:34, Robert L Cochran wrote: > Peter Arremann wrote: > >I can't recommend this board for a server use or a high end setup. But if > > you are looking for a cheap way to upgrade this board is unbeatable. At > > about $70 shipped you get something that will allow you to reuse your AGP > > card and then allow you to go PCI-E whenever you feel like it... > > > >Peter. > > Why isn't it a good server board? I'm not trained in how to evaluate a > board that can be used for server applications. You'll probably get two very different responses back from me and Bryan because we interpret the term "server" very different. For me, a server doesn't have to have redundancy, expensive raid or anything - it is simply something that is used to run server software. While many people that run a PC as a server simply don't have the money or the server task is not that important, you still want to get maximum stability possible. And that is unfortunately where the AsRock board falls short. For one, Asrock does not cross ship. If you have an RMA request, you're looking at a week. Other vendors will cross ship - you give them your credit card and as long as you return the borken board within 2 weeks they won't charge you. Another issue is the lm_sensor issue. For a server you want at least some monitoring. If you install unpatched FC4, you won't get any sensor information. The thrid and for me biggest issue is the driver support. This board is exotic and that is fine for a desktop board. But if you're somewhere at 2am trying to reload a box, you surely don't want to hunt around for some weird bug you might encounter. I have not found anything other than the lm_sensors yet, but still - just because less people use it, its less well tested. In the end, as I said (after all I pointed you to this board), I think its a great board for the price. Especially the dual agp/pci-e slots is cool. But if I had to run it as a server, I would spend the extra $30, get a NForce 4 board and have something that 20% of the comunity runs... Peter. From cochranb at speakeasy.net Mon Jan 9 02:22:29 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Sun, 08 Jan 2006 21:22:29 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <200601082052.58788.loony@loonybin.org> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601070059.05759.loony@loonybin.org> <43C1BD97.1000804@speakeasy.net> <200601082052.58788.loony@loonybin.org> Message-ID: <43C1C8E5.5050202@speakeasy.net> Peter Arremann wrote: >On Sunday 08 January 2006 20:34, Robert L Cochran wrote: > > >>Peter Arremann wrote: >> >> >>>I can't recommend this board for a server use or a high end setup. But if >>>you are looking for a cheap way to upgrade this board is unbeatable. At >>>about $70 shipped you get something that will allow you to reuse your AGP >>>card and then allow you to go PCI-E whenever you feel like it... >>> >>>Peter. >>> >>> >>Why isn't it a good server board? I'm not trained in how to evaluate a >>board that can be used for server applications. >> >> > >You'll probably get two very different responses back from me and Bryan >because we interpret the term "server" very different. For me, a server >doesn't have to have redundancy, expensive raid or anything - it is simply >something that is used to run server software. > >While many people that run a PC as a server simply don't have the money or the >server task is not that important, you still want to get maximum stability >possible. And that is unfortunately where the AsRock board falls short. For >one, Asrock does not cross ship. If you have an RMA request, you're looking >at a week. Other vendors will cross ship - you give them your credit card and >as long as you return the borken board within 2 weeks they won't charge you. > >Another issue is the lm_sensor issue. For a server you want at least some >monitoring. If you install unpatched FC4, you won't get any sensor >information. > >The thrid and for me biggest issue is the driver support. This board is exotic >and that is fine for a desktop board. But if you're somewhere at 2am trying >to reload a box, you surely don't want to hunt around for some weird bug you >might encounter. I have not found anything other than the lm_sensors yet, but >still - just because less people use it, its less well tested. > >In the end, as I said (after all I pointed you to this board), I think its a >great board for the price. Especially the dual agp/pci-e slots is cool. But >if I had to run it as a server, I would spend the extra $30, get a NForce 4 >board and have something that 20% of the comunity runs... > >Peter. > > > Thanks Peter. I was just wondering, and you make some good points here that I haven't thought of. I'm not sure I fully understand what lm_sensors will give you -- is this the ability to remotely check things like the CPU temperature? Using SNMP? And yes, the driver issue is an interesting one. I had a tough time with the driver for the NIC until I installed the latest FC4 kernel. I really need to learn how to compile a Fedora kernel, even though in this case there wasn't the need -- it would have been more work. I really I don't have any complaints about the board and I like it; I'm glad to have saved some money while I'm still focused on applications programming. I'm using it as a desktop board obviously. I am more than pleased with the CPU speed. I recently did an 'rpmbuild -ba' of Firefox 1.5 from source code that completed so quickly I didn't quite realize it. Bob From loony at loonybin.org Mon Jan 9 03:03:30 2006 From: loony at loonybin.org (Peter Arremann) Date: Sun, 8 Jan 2006 22:03:30 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <43C1C8E5.5050202@speakeasy.net> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601082052.58788.loony@loonybin.org> <43C1C8E5.5050202@speakeasy.net> Message-ID: <200601082203.31132.loony@loonybin.org> On Sunday 08 January 2006 21:22, Robert L Cochran wrote: > Thanks Peter. I was just wondering, and you make some good points here > that I haven't thought of. I'm not sure I fully understand what > lm_sensors will give you -- is this the ability to remotely check things > like the CPU temperature? Using SNMP? lm_sensors is the underlaying code for accessing SMBus implementations. SMBus (System Management Bus) is a very simple serial bus based on I2C that is used to monitor the system. What can be monitored depends on the board but usually its things like temperatures, voltages and fan speeds. I've seen PCI bus usage implemented once but unusual stuff like that is rare. lm_sensors implements drivers for the various SMBus controllers used on various boards. Ontop of that you have many different programs like gkrellm to do the actual monitoring for you. > And yes, the driver issue is an > interesting one. I had a tough time with the driver for the NIC until I > installed the latest FC4 kernel. I really need to learn how to compile a > Fedora kernel, even though in this case there wasn't the need -- it > would have been more work. Yeah - I warned you about that issue :-) I have a bunch of really cheap network cards so its no problem but it shows the issue I meant. Imagine its 2am, you only got FC4 cd's and no spare network card... enjoy :-) As for the kernel rebuild - Bryan posted a link to a kernel rebuild document recently he has written on one of the fedora mailing lists recently. Look for it - it has all you need. > I really I don't have any complaints about the board and I like it; I'm > glad to have saved some money while I'm still focused on applications > programming. I'm using it as a desktop board obviously. I am more than > pleased with the CPU speed. I recently did an 'rpmbuild -ba' of Firefox > 1.5 from source code that completed so quickly I didn't quite realize it. You're so lucky. I'm stuck doing most of my work on a P4 2.4Ghz laptop with the slowest drive I've ever seen... Compile for my current project (small, just about 15000 C++ lines I've written over the past few weeks) takes almsot 3 minutes - on my desktop FX-55 it takes about 45 seconds... Ordered a Dell with a Centrino Duo - but it won't be shipped till mid febrary *grrrr* Peter. From cochranb at speakeasy.net Mon Jan 9 04:05:43 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Sun, 08 Jan 2006 23:05:43 -0500 Subject: opteron 146 thoughts on motherboards In-Reply-To: <200601082203.31132.loony@loonybin.org> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <200601082052.58788.loony@loonybin.org> <43C1C8E5.5050202@speakeasy.net> <200601082203.31132.loony@loonybin.org> Message-ID: <43C1E117.9060606@speakeasy.net> Peter Arremann wrote: >On Sunday 08 January 2006 21:22, Robert L Cochran wrote: > > >>Thanks Peter. I was just wondering, and you make some good points here >>that I haven't thought of. I'm not sure I fully understand what >>lm_sensors will give you -- is this the ability to remotely check things >>like the CPU temperature? Using SNMP? >> >> >lm_sensors is the underlaying code for accessing SMBus implementations. SMBus >(System Management Bus) is a very simple serial bus based on I2C that is used >to monitor the system. What can be monitored depends on the board but usually >its things like temperatures, voltages and fan speeds. I've seen PCI bus >usage implemented once but unusual stuff like that is rare. >lm_sensors implements drivers for the various SMBus controllers used on >various boards. Ontop of that you have many different programs like gkrellm >to do the actual monitoring for you. > > > >>And yes, the driver issue is an >>interesting one. I had a tough time with the driver for the NIC until I >>installed the latest FC4 kernel. I really need to learn how to compile a >>Fedora kernel, even though in this case there wasn't the need -- it >>would have been more work. >> >> >Yeah - I warned you about that issue :-) I have a bunch of really cheap >network cards so its no problem but it shows the issue I meant. Imagine its >2am, you only got FC4 cd's and no spare network card... enjoy :-) > >As for the kernel rebuild - Bryan posted a link to a kernel rebuild document >recently he has written on one of the fedora mailing lists recently. Look for >it - it has all you need. > > > >>I really I don't have any complaints about the board and I like it; I'm >>glad to have saved some money while I'm still focused on applications >>programming. I'm using it as a desktop board obviously. I am more than >>pleased with the CPU speed. I recently did an 'rpmbuild -ba' of Firefox >>1.5 from source code that completed so quickly I didn't quite realize it. >> >> >You're so lucky. I'm stuck doing most of my work on a P4 2.4Ghz laptop with >the slowest drive I've ever seen... Compile for my current project (small, >just about 15000 C++ lines I've written over the past few weeks) takes almsot >3 minutes - on my desktop FX-55 it takes about 45 seconds... >Ordered a Dell with a Centrino Duo - but it won't be shipped till mid febrary >*grrrr* > >Peter. > > > His blog on kernel rebuilds is here, I think http://thebs413.blogspot.com/2005/10/kernel-26-on-fedora-based-systems.html I believe Bryan is trying to say that as a user you can make your actual kernel changes somewhere between steps 2 and 3. I tried posting a comment to this effect and ended up signing up for my own blog, much to my surprise. As to the slow laptop hard drive, it is possible to change the drive to a faster one for about $100 or so. I did exactly that with my wife's Compaq nx9010 notebook. Now I have a 7200 rpm drive running FC4. However, I made no attempt to migrate data from the old drive to the new one. The old drive is sitting in a nice antistatic bag, too. It was not that hard to install but I still sweated heavily until finally pushing the power button. Some day maybe I'll be able to change a laptop's motherboard. I know -- keep dreaming on that one. Bob From lamont at gurulabs.com Mon Jan 9 18:25:20 2006 From: lamont at gurulabs.com (Lamont R. Peterson) Date: Mon, 9 Jan 2006 11:25:20 -0700 Subject: opteron 146 thoughts on motherboards In-Reply-To: <43C1BB21.3000501@speakeasy.net> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <1136612117.4774.14.camel@bert64.oviedo.smithconcepts.com> <43C1BB21.3000501@speakeasy.net> Message-ID: <200601091125.26106.lamont@gurulabs.com> On Sunday 08 January 2006 06:23pm, Robert L Cochran wrote: [snip] > I stand corrected over the memory, then. I've also changed my 3-pin PLED > connector to a 2 pin configuration as suggested earlier simply by > pulling one pin out and putting it in the middle position of the > connector shell. I hadn't realized one can do that. Now my power LED > works. I'm starting to wonder how to crimp those little connectors for > other electronics projects. Ah, that brings back memories...those days gone by when I used to make thick-net cables. :) No, thank you...I don't want to go back and you can't make me do it. :) Realistically, it's not that hard, but you want to get the right tools if you're going to do it more than once. Years ago, I got down to 30 minutes per thick-net cable, with the good crimping tool designed for those pins. Any reasonable electric supply store will be able to help you, and show you how to make the crimps. -- Lamont R. Peterson Senior Instructor Guru Labs, L.C. [ http://www.GuruLabs.com/ ] -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: not available URL: From hahn at physics.mcmaster.ca Tue Jan 10 20:19:53 2006 From: hahn at physics.mcmaster.ca (Mark Hahn) Date: Tue, 10 Jan 2006 15:19:53 -0500 (EST) Subject: opteron 146 thoughts on motherboards In-Reply-To: <200601082203.31132.loony@loonybin.org> Message-ID: > > that I haven't thought of. I'm not sure I fully understand what > > lm_sensors will give you -- is this the ability to remotely check things > > like the CPU temperature? Using SNMP? > lm_sensors is the underlaying code for accessing SMBus implementations. SMBus > (System Management Bus) is a very simple serial bus based on I2C that is used > to monitor the system. What can be monitored depends on the board but usually > its things like temperatures, voltages and fan speeds. I've seen PCI bus right. for interest, "real" servers also tend to have IPMI, which is a network-able standard that often includes a small coprocessor, and also gives access via SMBus sensors. among the most valuable features of IPMI-type remote-management interfaces is the ability to cold/warm reset, power on/off, and often have some access the bios/console. I'm not sure why, but SNMP doesn't seem to go along with IPMI-type remote management widgets. regards, mark hahn. From andy at andydeano.co.uk Wed Jan 11 22:21:45 2006 From: andy at andydeano.co.uk (Andy Dean) Date: Wed, 11 Jan 2006 22:21:45 +0000 Subject: opteron 146 thoughts on motherboards In-Reply-To: <43BF2EAB.2000100@speakeasy.net> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> Message-ID: <1C294B07-1944-44D7-9FA1-5F03A79B4EA7@andydeano.co.uk> Hi thanks for all the replys regarding motherboards, there has certainly been some interesting and informitive post with regards to opteron motherboards. The supermicro H8SSL-i looks a very good board high end wise, and anyting with the nvidia NForce 4 ultra for lowend stuff really, very tempted by the asrock board though as well. Does anyone know of a motherboard that supports both Pci-x and Pci-e x16 as i've not found one yet. Finally probably a stupid question, would there be any advantages in running a 3ware 8006-2LP in a standard Pci slot on an Nvidia ultra board, rather than SATA-2 drives on the nvidia ultra SATA-2 onboard controller? i ask as i've been offered one very cheap :) Thanks Andy Dean On 7 Jan 2006, at 02:59, Robert L Cochran wrote: > The archives of this list has a considerable amount of discussion > about motherboards and Opteron processors. Spend a little time > reading them. I wanted to upgrade to a dual core processor myself > but had cost concerns. And it was a bit confusing reviewing the > available motherboards. Someone suggested an Asrock 939Dual-SATA2 > motherboard, and with the v1.5 BIOS update it is quite nice, I like > the board. You might want to look at it. A big advantage is that > you can recycle your existing memory sticks and video card to the > Asrock. > > Bob Cochran > > Andy Dean wrote: > >> Hi >> >> am looking at getting an opteron 146, and would like some input >> with regards to motherboards, whether it's worthwhile going for a >> tyan tomcat k8e or just for a asus nvidia sli board of some >> description, or something else if any one has some recomedations. >> Mainly to be used as a desktop system really. Just wanting to get >> the best that i can with the budget i have and would value your >> input. >> >> TIA >> >> Andy Dean >> > > > From b.j.smith at ieee.org Wed Jan 11 22:28:02 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Wed, 11 Jan 2006 14:28:02 -0800 (PST) Subject: opteron 146 thoughts on motherboards In-Reply-To: <1C294B07-1944-44D7-9FA1-5F03A79B4EA7@andydeano.co.uk> Message-ID: <20060111222802.12786.qmail@web34110.mail.mud.yahoo.com> Andy Dean wrote: > The supermicro H8SSL-i looks a very good board high end > wise, and anyting with the nvidia NForce 4 ultra for lowend > stuff really, very tempted by the asrock board though as well. Don't forget the nForce Pro 2200 (and 2200+2050). They _are_ available in single socket mainboards. > Does anyone know of a motherboard that supports both Pci-x > and Pci-e x16 as i've not found one yet. For single socket AMD, no, not yet. Only PCIe x8 and PCI-X. There are several dual-socket mainboards that add an AMD8131 dual-PCI-X 1.0 (or AMD8132 dual-PCI-X 2.0) to CPU #0, while using a nForce Pro 2200 on CPU #0 and, optionally, a nForce Pro 2050 on CPU #1. > Finally probably a stupid question, would there be any > advantages in running a 3ware 8006-2LP in a standard Pci slot > on an Nvidia ultra board, rather than SATA-2 drives on the > nvidia ultra SATA-2 onboard controller? i ask as i've been > offered one very cheap :) The problem is that the 3Ware will quickly _saturate_ your bandwidth on the legacy 32-bit @ 33MHz (133MBps) with modern drives. The 3Ware Escalade 8006 clearly wants >>100MBps for itself for reads (in RAID-1, reads/writes in RAID-0), and the higher port count 8506 products can easily exceed 250MBps at reads (and over 200MBps at RAID-10 writes in the 8 port version). On a workstation, that will kill anything else on the PCI bus (e.g., audio). That's why PCI-X is highly recommended for even those older cards. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does *** From maurice at harddata.com Thu Jan 12 00:19:35 2006 From: maurice at harddata.com (Maurice Hilarius) Date: Wed, 11 Jan 2006 17:19:35 -0700 Subject: opteron 146 thoughts on motherboards In-Reply-To: <1C294B07-1944-44D7-9FA1-5F03A79B4EA7@andydeano.co.uk> References: <2843189D-0D92-443A-AA8B-D0A656CC60EB@andydeano.co.uk> <43BF2EAB.2000100@speakeasy.net> <1C294B07-1944-44D7-9FA1-5F03A79B4EA7@andydeano.co.uk> Message-ID: <43C5A097.3070700@harddata.com> Andy Dean wrote: > > .. > > Does anyone know of a motherboard that supports both Pci-x and Pci-e > x16 as i've not found one yet. > Not yet, except in dual socket. Coming probably May, in ASUS. Socket939 Opteron, ATX form factor, both PCI-X and PCI-E, dual GbE, etc. A little bit more than the SuperMicro, lets say around $315 > Finally probably a stupid question, would there be any advantages in > running a 3ware 8006-2LP in a standard Pci slot on an Nvidia ultra > board, rather than SATA-2 drives on the nvidia ultra SATA-2 onboard > controller? i ask as i've been offered one very cheap :) > The 3WARE is a RAID controller, the on chip stuff is "FRAID" (Fake RAID). Of course a "standard PCI slot" if only 32 bit, 33MHz, is pretty slow.. -- With our best regards, Maurice W. Hilarius Telephone: 01-780-456-9771 Hard Data Ltd. FAX: 01-780-456-9772 11060 - 166 Avenue email:maurice at harddata.com Edmonton, AB, Canada http://www.harddata.com/ T5X 1Y3 From dfxinfx123 at hotmail.com Thu Jan 12 21:53:50 2006 From: dfxinfx123 at hotmail.com (chris) Date: Thu, 12 Jan 2006 15:53:50 -0600 Subject: compiling latest kernel on amd64 In-Reply-To: <20060112170013.C042D73C73@hormel.redhat.com> Message-ID: I get successful compiles but when I make install it complains that it is missing the sata_nv driver. My hd's are sata so obviously kernel panic on boot. I have sata enabled in my .config, (under scsi settings not the deprecated version). Anyone else experiencing this problem? Chris Burger Head of IT ABTG From jmacfarland at nexatech.com Thu Jan 12 22:13:09 2006 From: jmacfarland at nexatech.com (Jeff Macfarland) Date: Thu, 12 Jan 2006 16:13:09 -0600 Subject: compiling latest kernel on amd64 In-Reply-To: References: Message-ID: <43C6D475.2060602@nexatech.com> If you compiled it as a module, you will need to add appropriate entries to modules.conf and rebuild the initrd image. chris wrote: > I get successful compiles but when I make install it complains that it is > missing the sata_nv driver. My hd's are sata so obviously kernel panic on > boot. I have sata enabled in my .config, (under scsi settings not the > deprecated version). Anyone else experiencing this problem? > > Chris Burger > Head of IT > ABTG > -- Jeff Macfarland (jmacfarland at nexatech.com) Nexa Technologies - 972.747.8879 Unix Administrator GPG Key ID: 0x5F1CA61B GPG Key Server: hkp://wwwkeys.pgp.net From b.j.smith at ieee.org Thu Jan 12 22:04:01 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Thu, 12 Jan 2006 14:04:01 -0800 (PST) Subject: compiling latest kernel on amd64 -- initrd? In-Reply-To: Message-ID: <20060112220401.16143.qmail@web34111.mail.mud.yahoo.com> chris wrote: > I get successful compiles but when I make install it > complains that it is missing the sata_nv driver. My hd's > are sata so obviously kernel panic on boot. I have sata > enabled in my .config, (under scsi settings not the > deprecated version). Anyone else experiencing this > problem? You need an initial root disk (initrd) with sata_nv (which has dependencies on scsi_mod and sd). Anything that uses SCSI is a module so it must be put in an initrd if it's the boot device. If you have a scsi_hostadapter alias in /etc/modprobe.conf (e.g., alias scsi_hostadapter sata_nv), when you run "mkinitrd", it should automatically generate it (and all dependencies). So literally all you need to do is run: mkinitrd /boot/initrd-(your_ver/label) (your_ver/label) And add that initrd to GRUB. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does *** From dmccormick at wvmcc.com Sun Jan 22 16:24:30 2006 From: dmccormick at wvmcc.com (David McCormick) Date: Sun, 22 Jan 2006 11:24:30 -0500 Subject: [Ndiswrapper-general] Lockup with modprobe Progress I think In-Reply-To: <43D18353.6050409@wvmcc.com> References: <43BBC396.6060401@wvmcc.com> <20060105173627.9326DEF27C@wolfe.lmc.cs.sunysb.edu> <43D0E803.50408@wvmcc.com> <43D0F779.1050002@yahoo.de> <20060120151910.D427916F4F0@smtp.lmc.cs.sunysb.edu> <43D177C0.9030907@yahoo.de> <43D18353.6050409@wvmcc.com> Message-ID: <43D3B1BE.8040801@wvmcc.com> By adding the two lines: alias wlan0 ndiswrapper options ndiswrapper ifname wlan0 to /etc/modprobe.conf, the computer no longer crashes when I run modprobe ndiswrapper. If I remove the line options ndiswrapper ifname wlan0 it again crashes on modprobe. The following are the the error messages that I get when running modprobe ndiswrapper. /var/log/messages; Jan 21 08:07:30 localhost kernel: ndiswrapper: Unknown parameter `ifname' Jan 21 08:07:30 localhost modprobe: FATAL: Error inserting ndiswrapper (/lib/modules/2.6.14-1.1656_FC4/misc/ndiswrapper.ko): Unknown symbol in module, or unknown parameter (see dmesg) Jan 21 08:07:45 localhost kernel: ndiswrapper: Unknown parameter `ifname' [root at localhost ~]# modprobe ndiswrapper FATAL: Error inserting ndiswrapper (/lib/modules/2.6.14-1.1656_FC4/misc/ndiswrapper.ko): Unknown symbol in module, or unknown parameter (see dmesg) [root at localhost ~]# dmesg; ndiswrapper: Unknown parameter `ifname' Any ideas on how to correct this? David From P.C.M.Chiu at rl.ac.uk Tue Jan 24 18:13:05 2006 From: P.C.M.Chiu at rl.ac.uk (Chiu, PCM (Peter)) Date: Tue, 24 Jan 2006 18:13:05 -0000 Subject: motherboard comparison Message-ID: Does anyone has any experience with Tyan S2882-D and S2891 motherboards in their support with Suse 10/9.x with Dual Opteron 254 cpus? In particular, we would like to know if there is any significant performance difference between these two with 4x2GB DDR400MHz memory. Many thanks. Peter From b.j.smith at ieee.org Tue Jan 24 19:30:21 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue, 24 Jan 2006 11:30:21 -0800 (PST) Subject: motherboard comparison In-Reply-To: Message-ID: <20060124193021.83184.qmail@web34110.mail.mud.yahoo.com> "Chiu, PCM (Peter)" wrote: > Does anyone has any experience with Tyan S2882-D and S2891 > motherboards in their support with Suse 10/9.x with Dual > Opteron 254 cpus? > In particular, we would like to know if there is any > significant performance difference between these two with > 4x2GB DDR400MHz memory. Depending on the model of the older S288x, the BIOS/APIC registers might only support 166MHz DDR (DDR333/PC2700) synchronous clock for the memory, even though the CPUs support 200MHz DDR (DDR400/PC3200). Haven't checked BIOS updates on that because I haven't had really any of those boards, although you'll probably want it for the dual-core CPU support (if there is one available). The newer S289x were clearly designed with 200MHz DDR (DDR400/PC3200) in mind, as well as dual-core processors. BTW, the memory capabilities have _nothing_ to do with the HyperTransport chips used for I/O. I.e., the 288x is AMD8151, AMD8131 and/or AMD8111 combinations. The 289x is nForce Pro 2200, nForce Pro 2050 and/or AMD8131 combinations. Those have to do with I/O and peripherals. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does *** From cochranb at speakeasy.net Wed Jan 25 22:15:12 2006 From: cochranb at speakeasy.net (Robert L Cochran) Date: Wed, 25 Jan 2006 17:15:12 -0500 Subject: Server For Tutoring Organization Message-ID: <43D7F870.6070402@speakeasy.net> I've recently begun consulting for an organization that provides tutoring assistance in a specific area. I'm specifically being asked to fix their network and application software issues. I've just started there, and I'm accordingly just learning their network and workstation mix. The organization's leader is definite that a new server machine is wanted. Currently, a Dell Optiplex GX260 is running in a database server configuration. The organization has 20 student classes, approximately 10 full time staff and 100 tutors. The 10 "lab" computers are in constant use and the 4D server software is getting a lot of use. There are another 12 or so computers for the regular staff. Everyone is complaining of slow response times except for internet access. Everyone wants to replace the Dell with something more powerful. Is there one server box that can keep up with the demands of such an organization and accomodate some growth? Surprisingly enough, it looks like a rack enclosure was given to the organization, and the Dell box is enthroned on the bottom of this. Some of the rack mounts are used to support the monitor which is a few feet above it. I'm not sure of this is a 1U rack. While I am not new to networking, I am new to patch panels and rack mounted servers. I have a lot to learn here. Advice is appreciated. (Is there also a mailing list where I can ask questions about hardware such as patch panels and how to trace which computer a particular termination is serving?) Thanks Bob Cochran From dmccormick at wvmcc.com Thu Jan 26 16:24:52 2006 From: dmccormick at wvmcc.com (David McCormick) Date: Thu, 26 Jan 2006 11:24:52 -0500 Subject: Ndiswrapper on 64 bit Message-ID: <43D8F7D4.4090309@wvmcc.com> I would like to know if anyone has gotten ndiswrapper to run on a 64 bit machine. I can get it to run on a 32 bit AMD laptop but can't get it to load the module on a 64 bit AMD FX-51 box. David From berryja at gmail.com Thu Jan 26 16:43:03 2006 From: berryja at gmail.com (Jonathan Berry) Date: Thu, 26 Jan 2006 10:43:03 -0600 Subject: Ndiswrapper on 64 bit In-Reply-To: <43D8F7D4.4090309@wvmcc.com> References: <43D8F7D4.4090309@wvmcc.com> Message-ID: <8767947e0601260843w2e47d543qe4751b4da9cffd22@mail.gmail.com> On 1/26/06, David McCormick wrote: > I would like to know if anyone has gotten ndiswrapper to run on a 64 bit > machine. I can get it to run on a 32 bit AMD laptop but can't get > it to load the module on a 64 bit AMD FX-51 box. > > David Yes, it works just fine as long as you have a 64-bit Windows driver to use with ndiswrapper. There are a few out there. I have personally used it with a Broadcom chipset 802.11g wireless card. Jonathan From msaqer at iastate.edu Tue Jan 31 20:00:10 2006 From: msaqer at iastate.edu (Mohamad Al-Saqer) Date: Tue, 31 Jan 2006 14:00:10 -0600 Subject: DDR or DDR2 Message-ID: <200601311400.10123.msaqer@iastate.edu> Hello all, I am considering to upgrade my computer (which is a 754 athlon 64 3700+ and a k8v motherboard) to an amd 64 939 platforms, and I would like to know which direction shall I take. which of nforce[3,4] and via chipsets for fedora 64 bit is best supported? My main use of this computer will be computations that sometimes require large memory, simulations and also graphics. Regards; Mohamad Al-Saqer From msaqer at iastate.edu Tue Jan 31 20:02:26 2006 From: msaqer at iastate.edu (Mohamad Al-Saqer) Date: Tue, 31 Jan 2006 14:02:26 -0600 Subject: DDR or DDR2 In-Reply-To: <200601311400.10123.msaqer@iastate.edu> References: <200601311400.10123.msaqer@iastate.edu> Message-ID: <200601311402.26122.msaqer@iastate.edu> Sorry as the subject should say: nforce or via. On Tuesday 31 January 2006 2:00, Mohamad Al-Saqer wrote: > Hello all, > I am considering to upgrade my computer (which is a 754 athlon 64 3700+ and > a k8v motherboard) to an amd 64 939 platforms, and I would like to know > which direction shall I take. > > which of nforce[3,4] and via chipsets for fedora 64 bit is best supported? > > My main use of this computer will be computations that sometimes require > large memory, simulations and also graphics. > > Regards; > Mohamad Al-Saqer From b.j.smith at ieee.org Tue Jan 31 20:22:28 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue, 31 Jan 2006 12:22:28 -0800 (PST) Subject: DDR or DDR2 In-Reply-To: <200601311400.10123.msaqer@iastate.edu> Message-ID: <20060131202228.60986.qmail@web34111.mail.mud.yahoo.com> Mohamad Al-Saqer wrote: > Hello all I am considering to upgrade my computer (which is a > 754 athlon 64 3700+ and a k8v motherboard) to an amd 64 939 > platforms, and I would like to know which direction shall I > take. > which of nforce[3,4] and via chipsets for fedora 64 bit is > best supported? ViA has had a nasty and continual habit of changing their peripheral logic on a regular basis. This especially includes the ATA controllers. Although early nVidia chipsets had the associated peripheral support issues, any kernel 2.4.23+ (or backports in newer RHEL3 update releases of kernel 2.4.21) or 2.6.5+ should do just fine. The only issue I've recently run into was limited to the nVidia C51/NV44 (GeForce 61x0 + nForce 4x0) as I noted in the following blog posting, but seems to have been cleared up with a kernel update: http://thebs413.blogspot.com/2005/12/linux-on-nvidia-c51nv44-nforce.html Otherwise, the nForce platform drivers add the nvnet and nvsound modules. http://www.nvidia.com/object/linux_nforce_amd64_1.0-0310.html The nvnet driver is provided for older kernels. The GPL equivalent, which nVidia actively works on, is forcedeth. It should support even GbE models well in 2.4.23+/2.6.5+, although I had to upgrade to 2.6.12+ for C51/nF4x0. The nvsound OSS module is for pre-ALSA setups. Either use ALSA's AC'97 or upgrade or use a kernel 2.6 distro with ALSA if you don't want to use it. > My main use of this computer will be computations that > sometimes require large memory, simulations Consider Opteron then. The Socket-939 Opteron 165 or 175 are great products and tested to 24x7 operation, such as with multi-day (or week) simulations. > and also graphics. nVidia GeForce 7800GTX products are stable and probably the highest performing _general_ card. If you have more complex objects than typical multimedia, entertainment and entry-level CAD, there are better options for 4-5x the cost. If you want to save a few hundred, the 7800GT is the best bang for the buck right now. BTW, the latest nVidia ForceWare 80 are available for Linux and have SLI support. http://www.nvidia.com/object/linux_display_amd64_1.0-8178.html But I don't know what professional applications can take advantage of it. http://download.nvidia.com/XFree86/Linux-x86_64/1.0-8178/README/appendix-w.html -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does *** From b.j.smith at ieee.org Tue Jan 31 20:27:45 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue, 31 Jan 2006 12:27:45 -0800 (PST) Subject: DDR or DDR2 Message-ID: <20060131202746.62609.qmail@web34111.mail.mud.yahoo.com> "Bryan J. Smith" wrote: > Consider Opteron then. The Socket-939 Opteron 165 or 175 are > great products and tested to 24x7 operation, such as with > multi-day (or week) simulations. Consider the nVidia nForce Pro 2200 instead of the nForce4 as well. If you want SLI and/or an extra x4-8 PCIe channels for a storage controller, go for the Pro 2200 + 2050 combination. It's available in a select, few single socket mainboards. If you need PCI-X, get one with the AMD8131 (or AMD8132) added, although those are almost always dual-socket. I'd mention the ServerWorks HT1000, but it doesn't have PCIe video, just x8 for storage. I guess the HT2000+HT1000 is an option, but then I've only seen those in dual-socket (and I like the nFPro2200+2050+AMD8131 better there). -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does *** From msaqer at iastate.edu Tue Jan 31 22:42:40 2006 From: msaqer at iastate.edu (Mohamad Al-Saqer) Date: Tue, 31 Jan 2006 16:42:40 -0600 Subject: DDR or DDR2 In-Reply-To: <20060131202228.60986.qmail@web34111.mail.mud.yahoo.com> References: <20060131202228.60986.qmail@web34111.mail.mud.yahoo.com> Message-ID: <200601311642.41022.msaqer@iastate.edu> On Tuesday 31 January 2006 2:22, Bryan J. Smith wrote: > Consider Opteron then. ?The Socket-939 Opteron 165 or 175 are great > products and tested to 24x7 operation, such as with multi-day (or > week) simulations. what is an advantage would the opteron offer me over the regular 939 athlon if both have same FSB and cache, considering the price difference. For instance: Lowest price for opteron 175 is: $496.25 while for X2 4400+ is $459.50 where both have 2.2GHz with 2MB cache and 1000MHz FSB. Actually I am considering a dual socket with two single core processors. I guess this will reduce my choices to only 2xx opteron series, right? From b.j.smith at ieee.org Tue Jan 31 23:50:30 2006 From: b.j.smith at ieee.org (Bryan J. Smith) Date: Tue, 31 Jan 2006 15:50:30 -0800 (PST) Subject: DDR or DDR2 In-Reply-To: <200601311642.41022.msaqer@iastate.edu> Message-ID: <20060131235031.97471.qmail@web34103.mail.mud.yahoo.com> Mohamad Al-Saqer wrote: > what is an advantage would the opteron offer me over the > regular 939 athlon if both have same FSB and cache, > considering the price difference. What price difference? Have you actually priced the new Socket-939 Opterons?!?!?! They're maybe $20 more. They are: A. Tested to higher tolerances/specifications B. Always 1MiB L2 cache (2MiB L2 in dual-core versions) C. Always lead Athlon 64 in various improvements (like FX) > For instance: > Lowest price for opteron 175 is: $496.25 > while for X2 4400+ is $459.50 > where both have 2.2GHz with 2MB cache and 1000MHz FSB. God forbid, $35 more. > Actually I am considering a dual socket with two single core > processors. I guess this will reduce my choices to only 2xx > opteron series, right? Correct. Some of the Socket-940 Opteron 2xx are quite cheap these days. Of course, you'll then pay more for Registered DDR SDRAM -- something the Socket-939 Opteron doesn't have to do. -- Bryan J. Smith Professional, Technical Annoyance b.j.smith at ieee.org http://thebs413.blogspot.com ---------------------------------------------------- *** Speed doesn't kill, difference in speed does ***