PDA

View Full Version : pics of tejas hsf


saaya
03-22-2004, 03:21 AM
:eek:

http://www.theinquirer.net/?article=14851

................sick...... i remeber reading of a tejas ES 2.8ghz chip that was dissipating 150W+ ...sounds like the final chips will run at least as hot as this ES...

faruquehabib
03-22-2004, 03:25 AM
if thats the real hsf for them....wow, looks like nvidia gpu's all over again :D

barneybear
03-22-2004, 03:28 AM
thats friggin insane ! might make a good cooler for a prescott:p:

Peen
03-22-2004, 03:30 AM
wow!.. it seems like it would transfer heat better with the fins closer to the die , if you know what i mean, but what do i know.
glad I got a chiller now :)

SupaMan
03-22-2004, 04:01 AM
might just be an exotic design so that it looks cool. Werent there rumors that tejas would be 2 cores on one chip?

beatnic
03-22-2004, 04:03 AM
omg what!?

Intronic
03-22-2004, 04:03 AM
That's a huge HS yes, but it's not Intel's. :p:

QuadDamage
03-22-2004, 04:44 AM
Originally posted by faruquehabib
if thats the real hsf for them....wow, looks like nvidia gpu's all over again :D

none of nVidia's gpus run hot, it's the DDR-II memory.

st0nedpenguin
03-22-2004, 05:15 AM
Looks like it's the end of small rackmount cases for servers then. :eek:

`schr0et
03-22-2004, 06:28 AM
Who's up for some barbeque?

I got some steak that'll fit over that giant grill... er heatsink

You could probably make a good shish-ka-bob too ;)

st0nedpenguin
03-22-2004, 06:30 AM
I hear the word steak?

Count me in. :slobber:

Tomsawyer
03-22-2004, 06:43 AM
Got me thinking of camp chaos and thier spoof of Metallica.

FIRE BAD!

Fisk
03-22-2004, 06:50 AM
wow.. huge ^^

Farbror
03-22-2004, 07:23 AM
wtf ?!?! How are u going to get that sh*t in to the case because it´s so fu*king big....

Is´t that to heavy for the moderborde..think it´s gonna break...... :confused:

berkut
03-22-2004, 07:38 AM
I dont get it... why do they stick with that heatpipe- :banana: :banana: :banana: :banana:pipe.... ant they just copy a alpha PEP66 and make it bigger ??

Farbror
03-22-2004, 09:53 AM
Originally posted by berkut
I dont get it... why do they stick with that heatpipe- :banana: :banana: :banana: :banana:pipe.... ant they just copy a alpha PEP66 and make it bigger ??

No way ? get out of here?? :mad:

TheWeaseL
03-22-2004, 10:31 AM
Looks horrid. The only reason I could seem them even attempting a design like that would be to try and use the BTX style motherboards since the cpu is put right infront of the intake fan...Still fugly though...

lalPOOO
03-23-2004, 10:29 AM
I don't know why everyone hates it so much, as long as those heatpipes work it should be a pretty good thermal solution. It'd be a lot sexier if those fins were copper though.

Tomsawyer
03-23-2004, 10:37 AM
Reminds me of way back when the Celeron 300A was king

Computernerd used to sell the whopper celery cooler

It was a huge chunk of aluminum and copper, had 4 fans on it and must have weighed about 4 pounds. I bought one and it roared lol

http://www.sharkyextreme.com/hardware/cooler/cn.jpg


Amazing that old bx board survived, i eventually sold the whole set up to someone then got the watercooling bug and off I went.

Hallowed
03-23-2004, 11:38 AM
Why is this in Xtreme Overclocking and not Intel?



Yeah, seen it a few times now. Think the goal is to get the fins out in front of the BTX intake fans.

Firelord-OCHW
03-24-2004, 08:10 AM
Thats gross!

Stick with me water anyday!

YoupY
03-24-2004, 09:13 AM
OMG, I hope for intel this kind of solution isn't nessesary, but where else are go going to leave 150W with an air cooler.

Maybe watercooling should be the standard for this proc.

Metro
03-24-2004, 10:19 AM
I wonder if its not time to Intel to begin selling Watercooling kits with is CPUs. It begin to be a serious issue air coling is processors. I dont thing the BTX cases are the answer. They are better but not enough.
Intel is in trouble:cool:

Shockerr
03-24-2004, 02:09 PM
Well, air cant last forever, but I think it has a bit of time left. Im guessing that in the future, they will use heatpipes to pipe heat to use the computer case, or heatsinks, in different, high airflow areas.
But the with the current rate (plotted data into Excel from 1994 to 2004, giving power dissipation every 2 years for intel chips), the curve is exponential. In 2024, a CPU would have a heat output of 6000 watts (the sun puts out 6000watts per square cm, which is roughly the size of a CPU......).
In 2008, it would be about 240 watts, 2010, 350

IvanAndreevich
03-24-2004, 05:55 PM
I would love to read a review of one of those prototypes. Seems like they are very effecient and quiet. However, they are HUGE.

STEvil
03-24-2004, 06:38 PM
I'll bet the heatsinks suck.

saaya
03-24-2004, 06:54 PM
yepp i think they suck, they can remove a lot of heat, but they are not designed to cool the cpu to low temps, just keep it under 70/60°C under load thats it...

you wont get very good temps with such a monster on a current cpu i guess.

dont like the idear of default watercooling, then you would HAVE to get a phase change unit if you wanna oc seriously ...

solo
03-24-2004, 07:19 PM
looks good, but doesn't look too efficient to me

maybe the heatpipes will work well, who knows

IvanAndreevich
03-24-2004, 08:24 PM
>>they can remove a lot of heat, but they are not designed to cool the cpu to low temps
Self-contradicting statement? :D If it can cool a Prescott to 65C then it will cool a cooler CPU proportionally better as well.

solo
Those are some HUGE heatpipes. Remember, this is supposed to deal with 150W peak heat dissipation.

STEvil
03-24-2004, 08:37 PM
IvanAndreevich - If the phase material used inside the heatpipe only changes at a given temperature it will never go below that temperature, and if you think about it, it will not scale linearly to CPU power as temperature increases as it will not have enough time to become cool, but will instead stay hot and be "stuck" at the top where it is doing no good.... just continually heating more and more.

kommando
03-24-2004, 10:32 PM
Thats why i don't like air cooling :)

shrae
03-24-2004, 10:57 PM
The transfer of energy based on heatpipes is an interesting design, and it CAN work. The idea of the heat pipes is to constantly recirculate the energy. It doesn't necessarily mean that there is a great deal of mass flow going on.

Just think about what happens inside a pot containing water at boiling temperature (with just enough energy to keep it boiling). The temperature gradient is constant over time as you move vertically within the steam, but there is still a lot of energy being lost.

Of course, I don't see how the design couldn't be more efficient with fins closer to the base.

Hallowed
03-24-2004, 11:22 PM
I believe the purpose of that HSF is to put the cooling element directly behind the twin intake fans called for by the BTX design standard.

If you recall the BTX specifications place the processor immediately behind the intake fans but is still recessed far enough that it misses the flow. Hence heat pipes to alleviate this.

IvanAndreevich
03-25-2004, 12:18 AM
STEvil
Well, if they were smart, they would've used a material that would change at a low temp. That would make it pretty much universal :) I guess if you have a link about the types of materials used in these commercially available heat pipe coolers, I'd be happy to read it.

Formann
03-25-2004, 12:47 AM
I believe the purpose of that HSF is to put the cooling element directly behind the twin intake fans called for by the BTX design standard.

I agree. If the Tejas generate ~150w it will be a very wice thing to do, or else the case-temp will become very high.


The box Tejas come in is probably huge and has a Asetek/nVentiv logo on it :p:

Farbror
03-25-2004, 01:43 AM
Originally posted by Formann
I agree. If the Tejas generate ~150w it will be a very wice thing to do, or else the case-temp will become very high.


The box Tejas come in is probably huge and has a Asetek/nVentiv logo on it :p:

h0h0...Then they will not sell anything to dell , hp and that sh*t because they don´t have the cooling for it....

But it would bee a nice thing to do :

The New dell dimeson XXXXXXXX with nVentiv mach II GT.... :stick:

apathy^2
03-25-2004, 01:48 AM
i can see a new additional $100 premium for retail tejas cpu

saaya
03-25-2004, 05:08 AM
Originally posted by IvanAndreevich
STEvil
Well, if they were smart, they would've used a material that would change at a low temp. That would make it pretty much universal :) I guess if you have a link about the types of materials used in these commercially available heat pipe coolers, I'd be happy to read it.

its not that simple .if you use a coolant that evaporates at a lower temp it will not be able to remove as much heat as a coolant that evaporates at a higher temp because it will also need a lower temp on the fins to change its phase to liquid again and is less efficient.

same with a prommie, if you use a coolant with a lower evap temp you will need more energy to compress it again to make it change its phase to liquid again. in a compressor based phase change unit this means a higher work load on the compressor which is why compressors run hotter and draw more power when you swith to a coolant that evaporates at a lower temp and usually also means that the max WATTs the unit can remove becomes less.

a heat pipe is a self regulating system however and this means its a lot harder to balance it.

i dont think this cooler will cool any better than the current termaltake heatpipe copper coolers wich means they are not bad if you give tham a lot of cfm, but first of all they need the just said high cfm and second this cooler is a lot bigger wich means its more epxensive. now keep in mind that the termaltake heatpipe cooler wich is a lot smaller already costs ~50$...

Kanavit
03-25-2004, 05:45 AM
Originally posted by STEvil
I'll bet the heatsinks suck. I agree, i never liked copper based heatsinks, and especially heat pipes. They suck.

nailbomb
03-25-2004, 06:00 AM
I agree, i never liked copper based heatsinks, and especially heat pipes. They suck.

Jah, like the SP-94 and such :rolleyes:

Kanavit
03-25-2004, 06:14 AM
Originally posted by nailbomb
Jah, like the SP-94 and such :rolleyes:

i like this better. SLK948U
http://images10.newegg.com/productimage/35-109-114-01.JPG

than this..
SP-94
http://images10.newegg.com/productimage/35-109-012-01.JPG

Copper is a better conductor than aluminum, aluminum has better heat dissapation.

Soulburner
03-25-2004, 10:17 AM
Kanavit...those are BOTH 100% copper...the SP is better...

STEvil
03-25-2004, 11:18 AM
Copper good.

Heatpipe on stilts bad.

IvanAndreevich
03-25-2004, 06:04 PM
>>its not that simple .if you use a coolant that evaporates at a lower temp it will not be able to remove as much heat as a coolant that evaporates at a higher temp because it will also need a lower temp on the fins to change its phase to liquid again and is less efficient.

True, but look at those fins @ the top. They look pretty efficient, with a high surface area.

>>but first of all they need the just said high cfm and second this cooler is a lot bigger wich means its more epxensive

Neither the CFM nor the cost are the problem for the server market.

You have a point though.

sandman
03-25-2004, 06:07 PM
I don't get why you are all worrying os much about it.

a barton at 1.9V produces close tot hat, and you can aircool that just fine.

IvanAndreevich
03-25-2004, 08:30 PM
Uh... 150W for a Barton is 3.0 Ghz @ 2V. Are you gonna tell me you can aircool that just fine?

saaya
03-25-2004, 10:00 PM
Originally posted by IvanAndreevich

Neither the CFM nor the cost are the problem for the server market.

what? the most important factor in servers is money! theinquirer has an article about it as well, intel basically shot its own foot by telling everyone to use btx. its almost impossible to get a dual xeon potomac server with btx design cooled, and even if they find a way it will cost quite some extra money wich means the price diference between an amd and intel solution widens even more.

no, it wont be able to cool a barton@3ghz, the barton only runs at 3ghz with low temps you cant reach with air cooling.

IvanAndreevich
03-25-2004, 10:55 PM
saaya
>>the most important factor in servers is money!
Factors of importance in servers:
1) reliability (always #1 in any kind of server)
2) performance
3) cost
The last 2 get switched back and forth depending what kind of srever it is.

>>no, it wont be able to cool a barton@3ghz, the barton only runs at 3ghz with low temps you cant reach with air cooling
I know just getting back at the people who say that 150W are OK to remove with aircoolers :)

Penti
03-26-2004, 02:45 AM
Originally posted by YoupY
OMG, I hope for intel this kind of solution isn't nessesary, but where else are go going to leave 150W with an air cooler.

Maybe watercooling should be the standard for this proc.
150W is fine with a good heatsink, high rpm fan and ambient below 30°C.

Anyway i would rater have a good heatpipe solution then a german watercooling kit with a bad radiator. The bad water kitt will not be able to remove ~150W of heat from the cpu. A good designed heatpipe would. I would prefer heatpipe in servers and OEM computers, big heatexchanger away from the cpu to get good airflow trought it would handle alot, ofcourse it's not for lowtemp setups, but the tejas have to and will run fine at 70-90°C load.

Penti

saaya
03-26-2004, 04:36 AM
h3ll yeah! those innovatek kits are crap! its a myst to me how they ever managed to get such a large market share with those overpriced crappy sets...

well a btx case is THAT small that it will soon get imporrible to have a dual systerm aircooled inside of it because its just too little room. as every server room has air condition to remove the heat it would make more sence to get a pair of waterblocks and a pump wich will cost less than such a huge heatsink, and use the air condition to directly cool water ie the cpus...

YoupY
03-26-2004, 04:06 PM
Originally posted by IvanAndreevich
Uh... 150W for a Barton is 3.0 Ghz @ 2V. Are you gonna tell me you can aircool that just fine?

Also the contactarea of the die is a lot smaller then with the trejas which will probaly have a hs. It's easier to remove 150W from a large surface then from a small one.

IvanAndreevich
03-26-2004, 08:06 PM
YoupY
Of course! However, even though Tejas has a heat spreader, the die is actually small. The contact between the core and heat spreader is NOT perfect!

ThorinMan
03-26-2004, 10:44 PM
I was reading about future cooling solution in some pc magazine today, that theres lots or research in to really tiny fans. They said there are allready a few they have made that are the size of a pin head. Then the sick a whole crap load of them all over the transitors. I thought it would be pretty tight. But hey thats just be i would like to see 10 million little tiny fans on a cpu. Just dont touch it too hard or you might screw your cpu up.

cos
03-26-2004, 11:03 PM
Originally posted by ThorinMan
I was reading about future cooling solution in some pc magazine today, that theres lots or research in to really tiny fans. They said there are allready a few they have made that are the size of a pin head. Then the sick a whole crap load of them all over the transitors. I thought it would be pretty tight. But hey thats just be i would like to see 10 million little tiny fans on a cpu. Just dont touch it too hard or you might F*** your cpu up.

in which magazine?

saaya
03-26-2004, 11:42 PM
??? that doesnt make any sence to me, you will never get the same airflow out of 2 40mm fans a 80mm fan can offer you, plus the 80mm fan is cheaper to manufacture, lasts longer and produces less noise.

aoc007
03-26-2004, 11:50 PM
Originally posted by ThorinMan
I was reading about future cooling solution in some pc magazine today, that theres lots or research in to really tiny fans. They said there are allready a few they have made that are the size of a pin head. Then the sick a whole crap load of them all over the transitors. I thought it would be pretty tight. But hey thats just be i would like to see 10 million little tiny fans on a cpu. Just dont touch it too hard or you might F*** your cpu up.

Or get many microscopic cuts on your finger lol.

ThorinMan
03-27-2004, 10:57 AM
PC modder. I picked it up at the store the other day. It looked interesting. What i am tlaking about is on page 10. Here i will take a pic so you guys can read it.

ThorinMan
03-27-2004, 10:58 AM
and this is the magazine

IvanAndreevich
03-27-2004, 12:10 PM
Yeah to get the airflow of a 80 mm fan you need 4 40 mm fans or about that @ the same RPM with the same fin configuration.

SupaMan
03-27-2004, 01:49 PM
is that magazine any good? I got CPU for a while, mostly to read the columns that kyle and anand wrote.. but the reviews and stuff in there were pretty weak. Is this spinoff worth a subscription?

ThorinMan
03-27-2004, 03:03 PM
i have only read like the first 20 pages. This one is a great OC guide for newbs i thoguht it was going to be more indepth. It tells the bassics of OCing cpu, ram, gpu. Also talks about different cooling solutions from at to phase change. Nothing to deep ofcorse just basics from what i can tell. Its like 200 + pages and it looks to be all articles no ads. Its good entertainment reading for me a school. There is also a big section or reviews about the newest mobos. But i havent read that far. So far i think it. But i think its directed for beginer over clockers.

Skeet
03-27-2004, 03:07 PM
Hah you guys actually have mags on modding over there?:slobber: wow we are so after over here man..

celemine1Gig
03-27-2004, 05:08 PM
I think what they were referring to in their article could've been nanotechnology, used to cool semiconductors. BUT, AFAIK, this has nearly ZERO to do with fans. It's a completely different story. :)

IvanAndreevich
03-27-2004, 11:31 PM
Did this turn into the future cooling solutions thread?

saaya
03-28-2004, 04:51 AM
Originally posted by celemine1Gig
I think what they were referring to in their article could've been nanotechnology, used to cool semiconductors. BUT, AFAIK, this has nearly ZERO to do with fans. It's a completely different story. :)

read the article, 0.5mm fans are not nano technology at all, not even close.

and as IvanAndreevich pointed out those fans will NEVER be able to cool a cpu unless they spin at an extremely high speed (article is talking of 200rpms :rolleyes:)

koensa
03-28-2004, 06:22 AM
whats the tejas cpu?
i know its intel's

but what is it?
pentium5?or something more advanced(64bit ) like :itanium or gallatin?
or XEON like?

saaya
03-28-2004, 07:06 AM
northwood -> prescott -> tejas

causticVapor
03-30-2004, 01:58 PM
hmm

Methinks at least some good will come out of this...i.e. cases/cooling solutions will be designed with thermal headroom for the superhot intels, that way we can enjoy super-quiet, fast AMDs.

That and the shift to smarter, not only higher clock speed CPUs

ns_ripper
03-30-2004, 02:12 PM
what in hells name ... :shakes:

that's way to big for servers.

that peace of :banana: :banana: :banana: :banana: doesn't even fit in an 4u rackmountable case if you ask me. and that costs lots of money. rackspace isn't for free !!

if that realy is the new design of intel's cooler wel ...
They got :owned: for sure by the high end AMD hammer cpu's.


If you want to build superclusters with those Intel proc's you are damn sure heating the earth's atmosphere :eek:

pïezo electrical pumps are being tested for laptop purpose, I think that servers wil soon follow :D