· Content
· News
· Articles
· Mailinglists
· Knowledgebase
· Trouble Tickets
· Files
· Glossary
· Links
· Compatibility Lists
· Forums
Welcome to our website
To take full advantage of all features you need to login or register. Registration is completely free and takes only a few seconds.
Tyan prepares to unveil Thunder K8QS!
Posted by: Jim_ on: 03/10/2004 07:37 AM [ Print | 17 comment(s) ]
I noticed over at Digit-Life that Tyan is preparing to release yet another Opteron board, the Thunder K8QS.
Tyan is going to unveil its new server motherboard, Thunder K8QS (S4880), supporting up to 4 x AMD Opteron processors and featuring 4 x PCI-X, 1 x PCI slots, 4 x Serial ATA (RAID 0, 1, 10), dual-channel Ultra320 SCSI, dual-port Gigabit Ethernet, and supporting up to 20GB DDR400 (10 x DIMM sockets). Despite such features, the company positions the novelty as an entry-level product.Entry-level? Woah. They also have a picture of the sexy Tyan Transport GX28 over here. Tyan has said that they'd like us to look at some of their barebones server products, so you never know what products might make their way into the 2CPU.com secret, underground test bunker.
« 2CPU.com Raffle: 512MB Kingston Data Traveler 2.0! · Tyan prepares to unveil Thunder K8QS!
· Army to Gates: Halt the free software »
Comment
|
opus13 misanthrope. Posts: 1628 Joined: 2002-04-05 |
i do hope that a K8QS raffle is in the future |
Comment
|
Forge misanthrope. Posts: 704 Joined: 2001-05-12 |
I would not mind fiddling a quad Opteron at all, but only if it comes with some 8** Opterons. :\ Otherwise, getting that 'entry-level' board loaded would cost several grand. Registered Linux user 82133 (li.org has a short memory) |
Comment
|
spikegifted I do what? Posts: 6144 Joined: 2000-07-24 |
How come Digit-Life, a Russian web site, get to play with all the good stuff, all the time?? spikegifted.net "At the very beginning of the 19th century chemists generally thought that compounds from living organisms were too complicated in structure to be capable of artificial synthesis from non-living things, and that a 'vital force' or vitalism conferred the characteristics of living beings on this form of matter." Well, how wrong were they? So much for dogma... |
Comment
|
mickdunde Registered User Posts: 716 Joined: 2001-08-17 |
it's russia, maybe you better do not ask :rolleyes: PoD | mickdunde |
Comment
|
ingen_alls Unregistered |
This info is just the blurb in tyans monthly newsletter: http://www.tyan.com/support/html/newsletter16.html |
Comment
|
BlueCoder Registered User Posts: 118 Joined: 2000-04-22 |
I was a bit conserned with that 20GB number that it each processor wouldn't have it's own memory but quite to the contrary it appears to be a full NUMA system according to the pic of the motherboard. That should yeild some pretty impressive memory throughput numbers. Looks like a xeon killer. |
Comment
|
hellodeadcat SMP Guru Posts: 627 Joined: 2003-02-06 |
Bunch of nice pic here White trash Jesus freaks for Bush '04. Join us, YOU will be saved. |
Comment
|
Forge SMP Guru Posts: 704 Joined: 2001-05-12 |
I'm off to register 4CPU.com.... No really... One of those plus an AGP slot = wet dreams for months. Registered Linux user 82133 (li.org has a short memory) |
Comment
|
spikegifted I do what? Posts: 6144 Joined: 2000-07-24 |
spikegifted.net "At the very beginning of the 19th century chemists generally thought that compounds from living organisms were too complicated in structure to be capable of artificial synthesis from non-living things, and that a 'vital force' or vitalism conferred the characteristics of living beings on this form of matter." Well, how wrong were they? So much for dogma... |
Comment
|
Krogoth255 Registered User Posts: 30 Joined: 2004-02-09 |
Damm, how to fit and cool all the those Opertrons? That isn't a Personal Computer anymore what you got youself is a personal big iron! Barton 2800@2083 Mhz, ASUS A7N8X 1.4, RADEON 9700 PRO, 3x256MB Corsair PC3200, 5 ATA HDDs=540GBs, LTR-107T LTR-52327S, PX-708A ,TBSC, Logitech Z5500 and Fortron 530W |
Comment
|
rchatterjee SMP User Posts: 59 Joined: 2002-02-11 |
Once the opteron boards start switching over to PCI express slots and PCI express video cards start coming out it won't be just a dream for long :-). 2 x AMD Athlon MP 2800+, Tyan Thunder K7 S2462UNG, 1024MB Corsair DDR2100 Registered ECC RAM, 36.7GB Seagate X15-36LP, 18.4GB Seagate X15, 2x 36.7 GB IBM 36Z15, ATI AIW Radeon 9600 Pro, Creative Audigy Platinum Ex, Plextor 40/12/40 CD-RW, Lite-On LTD-163 16x DVD. |
Comment
|
dualboot_2xcpu hooked on sleeping pills Posts: 2397 Joined: 2001-07-16 |
Pretty soon 2cpu.com will have to be xCPU.com x being the number of processors heh. All joking aside, this will hopefully make the opterons cheaper and the dual opteron boards cheaper, so that more people will be able to buy these badboys! alex I finally have SMP. Looking for the next Mrs. Right now (c) 2004. HEMI. |
Comment
|
rmn oh my, it's huge! Posts: 6013 Joined: 2002-01-26 |
Actually, I'm pretty sure there used to be a site called 4cpu.com or .net or something like that (either in Russian or Chinese, I don't remember which, but I don't think it had anything to do with SMP). RMN ~~~ |
Comment
|
jrv-austin Registered User Posts: 257 Joined: 2004-01-19 |
"Iron" is right, as in clothing iron: I bet that thing could produce well over 500 W of heat no matter how careful you are. I'd really consider using ducted exhaust from each CPU cooler rather than the traditional fin & fan arrangement atop each CPU. The traditional approach would dump 400+ W of CPU heat into the chassis, and some of that will inevitably loop back into a CPU cooler intake again. Yuck. An Opteron 8xx has three HT busses. In a quad arrangement that means that each CPU can talk to the other three. But! One CPU needs to use an HT link to talk to the system controller ("south bridge"). And, since this board has PCI-X slots, another HT bus is needed to talk to the PCI-X controller. Based on the pictures my guess is that two of the processors have direct links to the other three, but that the other two are missing the link between them (but both have links to the first two). The latter two use their extra link to talk to the south bridge and PCI-X, i.e.
With this arrangement, if CPU 1 needs to access memory in a DIMM attached to CPU 2, the request has to be routed through CPU 3 or CPU 4. A little more expensive than usual for AMD. Tyan K8W S2885, 2x Opteron 248, 8GB ECC DDR400, 3ware 8506-8 mirrored 2x Raptor 74, 2x 7k250 HP zx6000 1x 1.5 GHz 6MB Itanic2 |
Comment
|
Forge Registered User Posts: 704 Joined: 2001-05-12 |
jrv-austin - That's a good theory and all, but the AMD-8131 PCI-X tunnel is a *tunnel*, not an endpoint. Only the AMD-8111 is an endpoint controller. http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_739_9004,00.html To add to that, the board is fully functional with only one CPU installed. I would theorize that it's not quite the same way you imagine it, but close, with CPU #4 only having one CPU link functional. So CPU #1 would be able to talk to the chipsets or to CPU #2, CPU #2 can talk to #1 and #3, CPU #3 can talk to CPU #2 and #4, and CPU #4 can talk to just CPU #3. I'd imagine it's even more elegant than that, though, with any CPU being able to drop a CPU link and link to the chipsets as needed, but it's hard to say. AMD doesn't provide a lot of docs on 4P implementations ATM. Registered Linux user 82133 (li.org has a short memory) |
Comment
|
jrv-austin Registered User Posts: 257 Joined: 2004-01-19 |
OK, I see Tyan K8W S2885, 2x Opteron 248, 8GB ECC DDR400, 3ware 8506-8 mirrored 2x Raptor 74, 2x 7k250 HP zx6000 1x 1.5 GHz 6MB Itanic2 |
Comment
|
opus13 misanthrope. Posts: 1628 Joined: 2002-04-05 |
what is that bottom slot? it looks like a regular 32/33 slots with an extra 'dickey'. maybe kinda like those old supermicro boards with teh raidport slot for zero channel adaptation? |
































