Slashdot Mirror
First Look At Intel Tejas & Socket 775
Anonymous Indian writes "The snoops at Anandtech have unearthed some details and photos of Intel's rumored Tejas 90nm CPU which draws 150 watts of power, a 50% jump compared to Prescott. It's also got an interesting locking mechanism instead of the traditional metal clip from hell for most processors." There's not much info beyond the photos, but it's still interesting to see what lies ahead for Intel.
So much power use from the first 90nm process cpu? There can be only one reason for this... the rumored twin core intel CPU. Let's see how hyperthreading run's on THIS baby!
Is that dissipated heat?
If so, ouch! that system is gonna need a huge heat sink or water cooling will be needed.
I would loves to see standardized water cooling parts for computers.
That way all you would need to do is buy the hosess and your favorite pump.
Regarding those metal clips from hell, I've always wondered why Intel and AMD never followed the examples of Alpha, Sun, SGI, etc. machines. They usually have bolt on heatsinks that either bolt the CPU and heatsink together or sandwich the CPU between the mainboard and the heatsink.
It takes care of the flat head screwdriver ruined mainboard, and there are no clips to break off the socket itself. I'm glad they are finally changing the way x86 sockets work.
Cthulhu Saves.
It's interesting to see how what was once consideres "high-end" eventually makes it's way into the consumer desktop stuff.
For example, this chip appears to use a "pinless" package design. Instead of little pins that fit into the socket, it has little ever-so-slightly raised 'nubs'. These 'nubs' simply sit on top of contact points in the socket.
This pinless design was being used by Compaq for the Alpha CPU as early as 2000, so this isn't a new packaging technology. The only problem Compaq had with it, was keeping all the little 'nubs' firmly in contact with the corresponding points on the socket. They used some sort of plastic clip design at first, which ended up with a high failure rate - not for the CPU, but for the plastic retaining mechanism itself. That's probably why intel is using that big beefy metal retaining clip.
I guess intel learned something from all those Compaq Alpha engineers they bought a few years ago.
I'd rather be a conservative nutjob than a liberal with no nuts and no job.
the "lack of metal clip", however, is so exciting that I am sure that we are to see perpetual world peace from its announcement any minute now.
n33t0x0rs. And it also may approach the heating capacity of an AMD chip, so now I'll be able to install an Intel to heat my room instead of my present AMD incumbent.
I find "While it is unlikely that Tejas includes two discreet Prescott cores on die, there is a chance that the two cores (if they exist) could be sharing data caches and maybe other units. A multicore Tejas would explain the jump in power consumption, and it is in line with Intel's strategy although it does seem sooner than expecte" interesting.
The first Pentium was a stripped-together combination of 2 486-like CPUs, with shared parts. To describe this as "two discreet Prescott cores on die... sharing data caches and maybe other units..." brings together this similarity. Perhaps Intel is trying the same strategy in its journey into 64-bitness that it has tried and succeeded at in the past (hedging its bets if its present design doesn't work out too well)? Just a thought, but IMHO not an unreasonable one.
--
FreeNET user? Comfortable with the adverse selection?
That's quite different from the standard today where the arm locks the pins of the cpu into the socket of the motherboard. I could be wrong, but that's what I think I see in the pictures.
> Now, it seems like they've lost that advantage.
I think that depends what the tradeoff is. Why would the cpu have such a high power output from a 90nm process? Either it's a huge cache inside, or there is more than one core. That sounds exciting. It may be very hot running but it has the advantage of space.
Just which applications this finds use in, I cannot tell.
a macgirl web cam. live 24/7
that's my first thought when I saw the picture. Why the CPU is looking so... um.. dirty (looks like corrosion)? ;) I'm just curious.
;)
I'm not an overclocker specialist, or whatever, and really I don't want to be a troll here
perhaps answer to this question will be modded informative (as this question itself is not
#
#\ @ ? Colonize Mars
#
RTFA! Oh wait, this is SlahsDot.
Realize that this processor is a Prototype, fabbed on a process that doesn't reflect Intels true capabilities. So criticism as to it's heat dissipation is at best pre-mature and at worst, downright off topic.
Sometimes I wish I was a plumber, then I'd know how to deal with other people's shit.
Is it going to be 64-bit? If not, a new chip is really a waste of money.
By 2006, most everything will be 64-bit, and Intel needs to realize this and stop making 64-bit chips just for the server market.
ehm, it would actually drain quite quickly. A regular car battery supplies about 14.4 volts when it's full (even though it's called a 12 volt battery), and can supply about 40 Ah, when drained slowly (less when drained quickly). 150 / 14.4 ~= 10 Amps, so it would last only 4 hours, and that's only for the CPU.
Mini-ITX P4
I can see them trying, but I don't really see it. Hyperthreading as it is generates a lot of extra heat that isn't exactly proportional to the performance gain (I've noticed this on my own 2.8C). I know there are some dual Xeon systems that have hyperthreading enabled, so they must have partially addressed the problem the inquirer put as...
"An instruction thread might well favour one kind of execution unit, but who's to say the other threads, the ones that are running in parallel with the 'troublesome' one, won't be favouring the same execution units? Assume you've got two cores per die, and that's two threads running simultaneously. Chuck in Intel's much-touted HyperThreading technology and you'll have four threads being processed in parallel, two per core, all being 'hopped'. Chances are that at least two of them will be using the same execution units, so swapping them round isn't going to change anything."
But in a dual core system, heat is a much bigger problem, so it would make more sense to up the clock speed over enabling hyperthreading whereas the heat on a dual Xeon would be easier to eliminate because there are 2 processors with their own cooling solutions. There's also the issue that stardard operating systems don't support 4-way systems for residential computing. There's also the fact that Intel COO Paul Otellini said "We'll go from putting HyperThreading in our products to putting dual-core capability in our mainstream client processors over time." That implies that hyperthreading is used as a stepping stone for the consumer, as some programmers optimize their code to be run on hyperthreaded systems, it will also boost performance on future dual core systems. Anyone else have any ideas?