Sneak Preview of AMD 760MP System

mjg writes: "Overclockers Australia have a sneak preview of a pre-release Tyan Thunder SMP motherboard using the AMD 760MP northbridge and a couple of 1.2GHz "workstation" Palomino (to be renamed AthlonMP) CPUs. The benchmarks are interesting, but should be taken with a grain of salt since this is a fairly early pre-release configuration. There are lots of nice images of the motherboard and CPUs included with the preview."

Comments

Google

[Anonymous Coward]

Google caches tons 'o stuff. I'm not saying that it's right, just that they do it. They seem to be doing just fine.

Banner ads still show up because the html for showing them still exists (as do images, etc).

Oh no!

[Anonymous Coward]

You Slashdoted Australia! You bastards!

not a problem . . .

[hawk]

I have one of these coming with 3x15k scsi and 4 ide drives for backup (faster & cheaper than tape).

It all works quite well. THis is a very small campus (ok, not that small--at some points, you can't see all 4 buildings :). I've subcontracted with the student union to fry eggs on the heatsink. They save electricity, the students get breakfast, and I dispose of heat . . .

Re:slashdot should mirror before posting

[Now15]

> If slashdot had a clue, they would mirror XYZ
> site before posting, using xxx.slashdot.org, ie
> overclockers.com.au.slashdot.org or something.
> Keep it mirrored for 1week then flush it.

I doubt any site that was funded by banner advertising would appreciate that.

Plus, I doubt that the person who pays Slashdot's bandwidth bill would appreciate either.

Not to mention the murky copyright issues involved...

Simon

--

Re:metric???

[Hammer]

No, MFLOPS means Million FLoating point Operations Per Second
That has very little to do with Hertz.

Re:slashdot should mirror before posting

[cyba]

1. They don't have right to do this and asking for permission would take a lot of time (aspecially for .au site, considering different time zone).

2. They have no interest in doing this. Mirrored pages would use their bandwidth but wouldn't contain any of their banners.

More info...

[dinky]

http://www.enpc.com/enpc/tyans2thun7d.html
http://www.enpc.com/enpc/tyans2thun7d.html

Release timeframe from Tyan

[Sokie]

I was researching putting together a dual Athlon machine for a client a couple weeks ago and I wanted to know when I could reasonably expect the MB to be available, so I asked Tyan:

"Thank you for your inquiry. Although Tyan has no comment regarding AMD's
plans, nor any chipset issues, I can tell you that our dual Socket A
platform is tentatively scheduled for a mid to late Q2 release.

Regards,
Matt Vernon
Manager of Marketing Communications
Tyan - Innovative Internet Platforms"

Just thought I'd share... :)

-Sokie

Finally!! but ..

[dgb2n]

This will obviously be a major step for AMD into the server market (at least once they go into full production) but I certainly have my doubts about how well any given setup can dissipate the heat from two Athlon processors.

Even though the Palomino is supposed to generate much less heat, I would imagine that two Athlons will pump out a frightening amount.

Also, it would make sense that a 400W or greater power supply would be necessary. Priced one of those lately?

I could use my Thunderbird 900 as a space heater for my mother-in-law if all the CPU and case fans didn't keep her up at night. She knows better than to turn off my server though ;-)

Re:Interresting benchmarks

[Gerad]

From, well, the body of the SUMMARY:

"The benchmarks are interesting, but should be taken with a grain of salt since this is a fairly early pre-release configuration."

What part of that do you not understand?

Putting the MP in AthlonMP

[clevershark]

Actually, if I may play the devil's advocate here -- Apple was marketing its dual processor G4 systems "MP" at least a year ago, way before "XP" was even the proverbial "glint in the milkman's eye" over at Redmond! This being said, 'MP' has been the short way to say "multi-processor" for as far back as I can remember, so I'm not claiming they came up with it.

And Apple was the first to decide on X as the monicker for its BSD-based OS, though at least it had the taste to point out that X was merely the roman numeral for 10.

Of course they royally screwed that explanation by (currently) referring to the system as OS X 10.0 (!). Must be something in the air in Cupertino :-)

Microsoft's naming efforts are rather pathetic in this area. The XP "brand" was a clear steal from Apple's "OS X", unless Steve Ballmer is going to claim that the new versions of Office and Windows (are you eXPerienced?) will make you high like you're on acid, or lose your virginity, or (more likely) lose all the hair on top of your head, presumably by making you pull it all out in frustration.

It's not to say that "OS X" was terribly original in light of Douglas Coupland's "Generation X" (1989?) and "The X Files", but at least OS X/10 did follow OS 9.

And wouldn't you know, there is an OS X native solution for those who use X-10 home automation devices... it's all a conspiracy, I tell ya!

Now if only Apple could find a way to make sure people don't think Microsoft came up with it first...

Re:400W PS

[connorbd]

See, what you do is you take the guts out of a CD mechanism and put the metal case into the drive bay nearest the CPUs. Then you get a shingle or a small pizza peel and you have an easy way to make lunch in your server. Make sure you leave the back on the drive case, though, and you might want to put a thin layer of firebrick (can probably get that at a model railroad shop) on the bottom of the case so it holds the heat; otherwise all you've got is an easy-bake oven and that's no fun except for cookies.

Actually, I kind of like that mobo design. The CPU sockets are staggered so you can actually put ovens in two different drive bays. Twice the pizza, twice the fun -- perfect for a LAN party...

An AMD sympathizer (stuck with Intel) just resurrecting an old Mac addict joke from the early Pentium days...

/Brian

Interresting benchmarks

[boaworm]

When i look at the benchmarks in SiSoft Sandra, i find it rather amusing that the Dual P3 1Ghz is faster then the Dual Athlon 1.2Ghz...

A few days ago there was a thread telling us a Single Athlon 1.2 Ghz was faster then intel's new XEON 1.7 Ghz...

Whats wrong with this picure... the P3 1Ghz is faster then the Xeon 1.7 Ghz ? :)

Guess there are some finetuning to make on this dual configuration. Awesome Mobo though, just wonder about the price tag...

Re:Interresting benchmarks

[C0vardeAn0nim0]

Probably because this is an early prototype that needs some optimizations.

Sometimes I read Tom's Hardware for news on processors and stuff, and according to their benchmarks Athlon is faster for somethings (like integers and floating point when using it's internal x87 unit) while P4 is faster for other things (multimedia and floating point when using it's brand new SSE unit).

And yes, clock by clock P3 can be faster than P4 in some benchmarks.

Re:Slashdotted

[Agg]

It's back up. FYI we had Apache set to load a few too many clients so ran outta memory. Reboot, trim it back and it seems to be working. Slow, and I imagine some people are getting "sod off, we're busy" messages, but at least people are seeing pages (eventually) now.

Re:Slashdotted

[Agg]

Yes, that's right. You're very clever and we're all very proud of you. Run along now, the adults are talking.

let me guess

[Proud Geek]

It's /.'d, so I can't read it, but let me guess. This:

The benchmarks are interesting, but should be taken with a grain of salt since this is a fairly early pre-release configuration.

means that an Intel based system whooped it ass.