AMD Turion 64 2800+ First Look!!

[sxs112]

http://img221.exs.cx/img221/1130/dscn02150jx.jpg
AMD Turion 64 2800+ - 1.6 GHz - 1M L2 - Core 90nm Rev 'E' - TDP 25W Socket754

http://img169.exs.cx/img169/3698/16002oh.jpg
1600Mhz


http://img169.exs.cx/img169/5719/25003cx.jpg
2000Mhz

Tnanks Taiwan 阿格斯 Test

Source (http://www.coolaler.com/ipb/index.php?showtopic=42810)

[=[PULSAR]=]

Yummy :slobber:

I wonder how these do under some serious volts. :banana:

[[XC] moddolicous]

Where did u get that?? I thought they were just focussing on finishing the FX-57. Oh well, nice results. Is that on stock heatsink and everything. Howcome voltage went down while clock went up? Looks good as of now.

[JDizzle]

My DFI wonders the same thing. Maybe it's the dynamic clocking that accounts for the difference in voltages?

[afireinside]

Damnnnnn those are nice. Will there be a 3200+ with a 10x multi? :D I smell 3.5ghz on Dice... Unless these have sub zero issues like the winnies? :(

[frostedflakes]

You sure about the 25w TDP. The D in the beginning of the OPN denotes 35w.

http://www.amdboard.com/low-power_mobile_roadmap.html

Looks like the Turion line will be 25-35w, but the mobile Semprons will be the only ones in 25w territory. Lancaster is rated at 35w.

Looks sweet. Can't wait for Lancaster. :)

[saaya]

sweet! 2ghz with 1.05v :o

i wonder what that would be like? 15W? :o

[frostedflakes]

Is that right? I figured it was a CPU-Z fudge. :eek:

If rev E0 can do 2GHz at 1v... wow...

[saaya]

i thougt so as well, but it reads the volts correctly on the first cpu-z screenie and 2ghz with 1v isnt THAT far off, ive seen 2ghz with 1.2v already...

happy bday sxs btw! :toast: :D

[Sgt_Strider]

Damn, that's amazing guys.

[TEDY]

turion = 754 amd 64 mobile ?

[shuRe]

will the turion be dual or single channel (im guessing single) with a better mem controller than other newcastles?

im hoping ill be able to put it in my DFI 250gb :p:

[EclectiK]

turion = 754 amd 64 mobile ?

Correct, there will not be any mobile for 939.

ECK

[EclectiK]

will the turion be dual or single channel (im guessing single) with a better mem controller than other newcastles?

im hoping ill be able to put it in my DFI 250gb :p:

I don't think so, but maybe I'm wrong, in fact, the E0 have improved memory controller, but that don't mean it will support dual channel.

ECK

[JDizzle]

It's single channel,d ual channel just isn't possbile with Socket 754,

[computersmsa]

A64 4200+
Turion 2800+
What now ? FX57 N :D
Where have you them ? :D

Great ... My A64-M 3400+ 62W CG S754 doesn't make better than 1.1 GHz with 1V !!
A don't have testing my Winnie @ 1V :)

[MrQ3W]

What are you waiting for? overclock it more!

PS. my 3000+ does 2GHz at 1.0V too. :D

[Kanavit]

looks awesome!

[saaya]

A64 4200+
Turion 2800+
What now ? FX57 N :D
Where have you them ? :D

Great ... My A64-M 3400+ 62W CG S754 doesn't make better than 1.1 GHz with 1V !!
A don't have testing my Winnie @ 1V :)

thats cuz its 130nm :)

[[XC] moddolicous]

A64 4200+
Turion 2800+
What now ? FX57 N :D
Where have you them ? :D

Great ... My A64-M 3400+ 62W CG S754 doesn't make better than 1.1 GHz with 1V !!
A don't have testing my Winnie @ 1V :)
yea, he's right. How have you been getting all this new stuff that we dont have access to? I wouldnt mind testing stuff out for AMD.

[saaya]

lol, hes not testing it and theres no way any of you will get stuff from amd to test, sorry to pop your dreams :D

just stop asking where it comes from, if he would answer your questions that somebody we are so thankfull for showing us those informations might lose his job or even get sued by amd, so just sit back and enjoy those wonderfull results :D

[terrace215]

You sure about the 25w TDP. The D in the beginning of the OPN denotes 35w.


I've never seen TDP encoded in an AMD OPN before. Do you have a link?

(Fab51 is wrong to split the first 3 letter code into 2 letters + 1 "power letter", as it clearly DOESN'T WORK with many example A64's and Opterons!)

So the new code of "TMD" must be taken as a whole. We don't know the official TDP until it is announced. (But they've already said their will be 25W and 35W parts, and we already *have* the 35W line (current 90nm LP Mobile A64).)

[frostedflakes]

Neither had I in any AMD spec sheets, but according to this AM is the family (a.k.a. Athlon64 mobile in the current mobiles) and the letter after it is the power design. A = 81.5w for DTRs, N = 62w for transportables, and D = 35w for low-power. Looks like mobiles fall under the Turion family now (TM), but the next number should still note TDP.

http://fab51.com/cpu/guide/opn-64-e.html

[terrace215]

Neither had I in any AMD spec sheets, but according to this AM is the family (a.k.a. Athlon64 mobile in the current mobiles) and the letter after it is the power design. A = 81.5w for DTRs, N = 62w for transportables, and D = 35w for low-power. Looks like mobiles fall under the Turion family now (TM), but the next number should still note TDP.

http://fab51.com/cpu/guide/opn-64-e.html

I just edited my post above.

Fab51 is wrong, IMO. The existing parts out there do not support the interpretation of the 3rd letter as power.

(Winchester has the same ADA code as Newcastle, for instance, with different TDP.)

You take the first 3 letters together as the "family" of part, and there may be different TDPs within it. That's what the data to date indicate.

[frostedflakes]

Good point, I can't say I really have an arguement to that. All the other info seems to fit, though, so I really have no reason to believe they're wrong.

But notice the roadmap posted in my first post also puts the Lancasters at 35w and Semprons at 25w. Plus they appear to operate at the same voltage (K), so I don't see how TDP could be any lower unless there have been major tweaks with rev E0 to reduce heat output (look how much of a difference it made from C0 to CG, so it's possible).

Not like it really matters, though; until the product comes out, this is all speculation. :)

[isp]

wow, with the voltage there you could have a silent but powerful system...

[trans am]

http://www.theinquirer.net/?article=21613

You made it to the inquirer!!!
Look how they mispelled our sites name. "OUR FRIENDS at Extreme Systems managed to get a picture and some basic test of AMD's upcoming Centrino killer."

Must not be very good friends if they can't even manage to spell the sites name correctly lol.


So basically this is an Oakville with 1MB L2 and runs with lower volts.

[frostedflakes]

Default voltage is still the same as Oakville, but looks like because of tweaks in revision E0 it is able to run at very low voltages, and because of this should probably OC well. :)

[Ashe]

Anything known about when they will hit the market?
Will they run on every Motherboard? (with the right bios) Would be great cause i have a dusty unused s754 mobo right here ;)

[frostedflakes]

The only solid date I have heard is 4/18, but who knows if it's true or not.

Not all motherboards support mobiles. I've heard of people having good luck with the DFI, Chaintech VNF3-250, and MSI K8N Neo, but not really much about other S754 boards.

[[XC] moddolicous]

lol, hes not testing it and theres no way any of you will get stuff from amd to test, sorry to pop your dreams :D

just stop asking where it comes from, if he would answer your questions that somebody we are so thankfull for showing us those informations might lose his job or even get sued by amd, so just sit back and enjoy those wonderfull results :D
Sorry saaya, just got a little curios :( . I wouldnt want the guy to lose his job over me wanting to test this kind of stuff.

[IvanAndreevich]

trans am
Don't forget that this is E0 so SSE3, SS etc.

[jjcom]

yeah, AMD must have there 90nm process tweaked pretty good now. Thats an early model and its doing well. but thats only one chip. So we wait and see what these things will do in the future lets hope well

jjcom

[WeStSiDePLaYa]

i think the s754 platform is where the true high oc's are comming from.

[terrace215]

Good point, I can't say I really have an arguement to that. All the other info seems to fit, though, so I really have no reason to believe they're wrong.

But notice the roadmap posted in my first post also puts the Lancasters at 35w and Semprons at 25w. Plus they appear to operate at the same voltage (K), so I don't see how TDP could be any lower unless there have been major tweaks with rev E0 to reduce heat output (look how much of a difference it made from C0 to CG, so it's possible).

Not like it really matters, though; until the product comes out, this is all speculation. :)

But the power info doesn't fit, so you do have reason to believe it is wrong. :)

That power roadmap is old.

See here: http://aceshardware.com/forums/read_post.jsp?id=115119165&forumid=1

[Mats]

That's a 35 W model according to this:
http://www.x86-secret.com/index.php?option=newsd&nid=848 >> T64 ML 30 (going FX style, no more confusing P-rating, just strange codes...)

About the three letters, I think it just shows if it's desktop/DTR/low power, no TDP.

[UCmajewski]

hmm wonder the voltage on that thing?

[frostedflakes]

Should be 1.35v for the 35w chips. Don't know about 25w.

Mats, interesting link. Looks like this is 35w, but not for the reason I originally thought.

What I find interesting is they'll have 25w and 35w versions all the way up to 2.2GHz (3400+, I'd assume). The 2.4GHz chip will only come in 35w though. I also didn't know there would be 512kb versions.

[jjcom]

I don't think so. 1.35v was for 35W on the AXP-Ms. So I would guess 1.3 or less for these chips.

jjcom

[saaya]

http://www.theinquirer.net/?article=21613

You made it to the inquirer!!!
Look how they mispelled our sites name. "OUR FRIENDS at Extreme Systems managed to get a picture and some basic test of AMD's upcoming Centrino killer."

Must not be very good friends if they can't even manage to spell the sites name correctly lol.


So basically this is an Oakville with 1MB L2 and runs with lower volts.

fuad doesnt luv me anymore :(


:lol:


im pretty sure its 1.3v, a 3500+ winchester consumes around 36W of power under load, may load probably 40W or 45W ,then 1.3v should be just about right for 35W absolute max heat production/power consumption

[WeStSiDePLaYa]

Ii would say volts less than 1.25vcore. as my 35w newcastle is 1.25vcore

[frostedflakes]

All I know is current Oakvilles have the OPN AMDxx00BKX4LB, and this new chip is TMDML30BKX5LD. According to Fab51 and all the forums I've visited K is supposed to denote 1.35v.

[saaya]

Ii would say volts less than 1.25vcore. as my 35w newcastle is 1.25vcore

? 35W newcastle?

[frostedflakes]

Before 90nm there were 35w low-power Newcastles. They had a default Vcore of 1.2v I believe.

At the same voltage 90nm runs cooler than 130nm though, which is the reason why they don't have to lower it to 1.2v to maintain the same 35w rating.

[Revv23]

so the second two pics arent working... what did they contain?

[saaya]

wlel anything between 1.3 and 1.35v sounds about right for 35W :)

and 2ghz with 1v is truly a centrino killer! :slobber:

[terrace215]

Yes, there are other ways of lowering power besides voltage. Different transistors, better strained silicon (e.g. DSL) for lower leakage, etc.

[terrace215]

wlel anything between 1.3 and 1.35v sounds about right for 35W :)


There should be 1.6GHz (2800+) and 1.8GHz (3000+) 25W parts, along with a 35W 2GHz (3200+) at launch.


Amdboard didn't copy the slide right! See here:

http://www.theregister.co.uk/2004/06/14/amd_low-power_mobiles/

And the slide:

http://66.102.7.104/search?q=cache:bLP3T2DsoV4J:www.hardwareluxx.de/story.php%3Fid%3D491+Lancaster+25W&hl=en&client=firefox-a

You have to look closely, but many of the Lancaster parts are 25W, not 35W.

[saaya]

so the second two pics arent working... what did they contain?

http://img169.exs.cx/img169/3698/16002oh.jpg

http://img169.exs.cx/img169/5719/25003cx.jpg

[reject]

wow, didnt notice it was 1mb
i want one!

[terrace215]

Despite the "3rd letter encodes power" garbage, it appears the pictured part was a 35W part, not a 25W part.

But we are due to get 25W 1.8GHz and 35W 2.0GHz in Q1! Rev E, SSE3, 754!

x86-secret has the scoop:

http://babelfish.altavista.com/babelfish/trurl_pagecontent?url=http%3A%2F%2Fwww.x86-secret.com%2F%3Foption%3Dnewsd%26nid%3D848&lp=fr_en

At a few days of the official advertisement of the AMD Turion 64 (probably in CeBIT), we could get complete information to us on this new processor. Intended to compete with Pentium M in the field of the portable PC, Turion 64 is in fact only one version "Low Voltage", consuming less, of Athlon 64. Thanks to the process 90nm, AMD can thus lower the supply voltage of its Athlons and offer of CPUs Mobile not exceeding 35 Watts, or 25 Watts according to versions'. Equipped with the revision E of the core K8 (having the SSE3), Turion 64 will be equipped with 1 Mo of L2 mask and the support of the DDR400 on only one channel. Let us see the roadmaps and the frequencies of launchings:


http://www.x86-secret.com/pics/news/q105/table.png

First thing, AMD gives up here P-rating of the type "3200+", but adopts a new notation. This time, one speaks about "Turion 64 ML-30" or "Turion 64 MT-28". Ml indicating a CPU consuming 35 Watts maximum and MT, 25 Watts. The following figure being a kind of P-rating, based on the frequency and the size of the mask. Turion 64 MT-30 will function, for example, to 1.6 GHz and will be equipped with 1 MB of L2 mask. This lucky find marketing probably will muddle still a little more the purchasers, who had already extremely to make with the Processor number INTEL and P-rating of AMD. Limited to 2 GHz for the version 35 Watts and 1.8 GHz for the version 25 Watts with launching, Turion 64 should gain 200 MHz per quarter until A the end of the year.

[saaya]

thx for the infos! :)

the naming of the cpus is kinda lame imo, trying to make them look really powerfull, mt-37 is obviously supposed to look close to a 3700+ wich is a joke with 2ghz 1mb and only single channel memory :/

depsite the naming of those cpus they look very very promising though, this is a great upgrade option to 754 users extending the life of that plattform for several years! :banana:

they wont become the overclockers choice as the mobile barton and mobile clawhammer chips though i guess... current chips can already to 2ghz with 1v, i wonder why amd i going for 1.2-1.35v :confused:

[terrace215]

I'm not sure they are supposed to look like anything, really.

They need some scheme that is reasonably short, but allows for two different power ratings, yet indicates the direction of increasing performance.

And most people will be comparing them against Dothans. In that sense, it is actually interesting that the model schemes no longer bear any resemblence to one another.

[Athens[2004]]

What are you waiting for? overclock it more!

PS. my 3000+ does 2GHz at 1.0V too. :D


Mine 2 , and windows prisntscrntable with < 1v :banana:

[krampak]

How could I calculate the wattage of my CPU? Mine is prime stable at 2'1GHz 1'1V with and XP90 and I haven't tried harder (will do soon).

http://img209.exs.cx/img209/9007/21ghz112v0oi.th.jpg (http://img209.exs.cx/my.php?loc=img209&image=21ghz112v0oi.jpg)

The cpuz verified screen (http://valid.x86-secret.com/show_oc?id=5694) doesn't say the voltage :(

[panumas05]

I hope Turion get 3.3GHz on air overclock.

:banana: :banana: :banana:

[Napoleonic]

hmmm, tell me how can that processor be overclocked with lesser voltage? what causing that? why will AMD sell them with higher voltage?

do you think this can beat Dothan?

[VVJ]

I think there is a little cheat! First of all, please, look at the OPN number. It is TMDML30BKX5LD, but sys112 said that was AMD Turion 64 2800+ processor operated at 1.6GHz. I think we have the real AMD Turion 64 ML 3000+ 1.80GHz processor. According to OPN number (that is real) the processor has a default core voltage of 1.350V (letter K). Now take a look at cpuz screens. What do we see? We see that the processor operates at 1.6GHz with 1.400V core voltage. I think this is too high for the mobile 0.09 u 1.60GHz processor.

[OC Detective]

Its 1.6Ghz with 1Mb of L2 cache VVJ not 1.8Ghz

[saaya]

hmmm, tell me how can that processor be overclocked with lesser voltage? what causing that? why will AMD sell them with higher voltage?

do you think this can beat Dothan?

maybe because it was in a desktop board with better cooling. lower temps -> higher oc at the same vcore or lower vcore for the same oc
and this might be a good chip, maybe the average turion cpus only run that clockspeed with a higher vcore with laptop heatsink... and dont forget that they always leave some headroom in clockspeeds as well to make its always stable.

and last but not least i really think that amd managed to get more out of their chips then they expected! they seem to run much better with lower volts than amd probably thought they would when they came up with the specs of the turion plattform. :)

[THunDA]

:slobber:

Me needs to subscribe to this thread :)

[Sentential]

Perhaps this will prlong S754 even further. I cant imagine what it would be like with a 35W 2.4ghz mobile with 1 meg of cache. On phase it should clock to increible heights

[UCmajewski]

hopefully the board support is there for Turion 64 ... might have to rethink 939 purchase if they still will have decent 754 chips out. Also can we confirm AMD wont bring 939 to the notebook? Isnt sempron going 939 sometime in the future?

[shuRe]

im thinking that if i buy a cheapo a64 laptop now then i can get a turion and replace the newcastle :) thats if a bios is released to support the turion on mobile motherboards.

[terrace215]

I think there is a little cheat! First of all, please, look at the OPN number. It is TMDML30BKX5LD, but sys112 said that was AMD Turion 64 2800+ processor operated at 1.6GHz. I think we have the real AMD Turion 64 ML 3000+ 1.80GHz processor. According to OPN number (that is real) the processor has a default core voltage of 1.350V (letter K). Now take a look at cpuz screens. What do we see? We see that the processor operates at 1.6GHz with 1.400V core voltage. I think this is too high for the mobile 0.09 u 1.60GHz processor.

No, the ML-30 is the 1.6GHz 1MB L2 35W part.

The MT-30 is the 1.6GHz 1MB L2 25W part.

He just didn't know the new official name, and went with the old PR rating for a 1MB 1.6GHz part, which was 2800+.

[MaSell]

oh yeah... any sponsors :slobber: :slobber:

[VVJ]

terrace215,
I think you are right! By the way, that BIOS doesn’t support AMD Turion 64 processor, because BIOS must name it by AMD Turion(tm) 64 Mobile Technology ML-XX where XX is a model number.

[terrace215]

Yes it looks like (at least) CPU-Z couldn't decode it. Maybe he had to guess and set the voltage manually?

[UCmajewski]

3.3ghz would be nice!

Are all revision E's going to be that seemingly overclockable or is this just a Turion thing?

[mcnbns]

All E stepping CPUs should overclock like mofos.

[don_vercetti]

42s at 2ghz for an a64 (which is what the turion is loosely based around, i assume) isn't all that good. Is it?

Still 2 ghz on 1v is bloody impressive. That must run very cold indeed. How high has it been clocked with higher voltage?

[jjcom]

42s at 2ghz...doesn't sounds all that bad to me, single channel

jjcom

http://www.xtremesystems.org/forums/showpost.php?p=747046&postcount=65
take a look there, I know not the same platform, but good comparison IMO

[WeStSiDePLaYa]

guys vcore WILL BE LESS THAN 1.25V! my 35w newcastle mobile has 1.25vcore at 1.8ghz. these are 90nm so there will be a vcore drop, also there are 25w parts. im thinking 1.1vcore.

[|3ourne]

maybe because it was in a desktop board with better cooling. lower temps -> higher oc at the same vcore or lower vcore for the same oc
and this might be a good chip, maybe the average turion cpus only run that clockspeed with a higher vcore with laptop heatsink... and dont forget that they always leave some headroom in clockspeeds as well to make its always stable.

and last but not least i really think that amd managed to get more out of their chips then they expected! they seem to run much better with lower volts than amd probably thought they would when they came up with the specs of the turion plattform. :)

Saaya , ull be surprised how many winchesters are stable while undervolting. I myself have owned two winchesters , a 3000+ and a 3500+ , both were stable at 1.158 V and 1.2 V respectively . By stable I mean any given benchmark , 3d or prime95 or S&M or stresscpu. They were completely stable at stock speeds.
I undervolt my processor as a part of my burn-in process , thanx to Sentinel who has posted in this very thread and whose burn in guide is posted over at OC Forums. I have had many pms requesting me to make a burn in guide and I am currently working on one. I hope Sentinel doesnt mind , its basically his concept which has worked great for me till now . :toast:

[frostedflakes]

But the thing is that at the same voltage 90nm runs cooler. So the voltage won't need to be lower than the Newcastles to have the same 35w rating. According to the OPN, it's 1.35v. I'm not sure about the voltage on the 25w chips yet.

Also, I contacted an online store about these Turions, and one of the sales staff mentioned they have one MT-30 in stock. My reaction was: :eek:

I'm not sure if it's a sample or what. If he is allowed to resell it, though, I'll be sure to buy it. At the least I'm going to have him send me the OPN so we can find out what voltage these 25w chips are rated for.

[majkel84]

2 Gigs on 1V is impressive and looks like centrino killer indeed, but i t doesn't means that Lancaster would clock high, even with higher voltages.
It was preety often to see Winnies that made 2,7G on 1,45 but woudnt ever go 2,9 even on 1,7V.
But SS may change that... i think
Turion is tremendous to show the 754 power !!! :banana:
Get use to 3G on air :banana: :D

[terrace215]

But the thing is that at the same voltage 90nm runs cooler. So the voltage won't need to be lower than the Newcastles to have the same 35w rating. According to the OPN, it's 1.35v. I'm not sure about the voltage on the 25w chips yet.

Also, I contacted an online store about these Turions, and one of the sales staff mentioned they have one MT-30 in stock. My reaction was: :eek:

I'm not sure if it's a sample or what. If he is allowed to resell it, though, I'll be sure to buy it. At the least I'm going to have him send me the OPN so we can find out what voltage these 25w chips are rated for.

Do let us know what the OPN is!

[frostedflakes]

Will do. I couldn't have even imagined they'd have one already.

He also said that they'd have more by April. So I'm guessing the 4/18 release date I read about somewhere is fairly accurate.

[Bloody_Sorcerer]

yahoo! hope for us 754ers after all! :banana:

[frostedflakes]

It was just a sample. I doubt they're too anxious to break NDA and sell some kid a Turion mobile. :D

Looks like the CPU is currently installed in a system, so he couldn't get me an OPN.

[terrace215]

It's often more difficult to purchase mobile CPUs anyhow.

But if it is in a system, he could run CPUZ...

[frostedflakes]

I don't know if the guy knows about CPU-Z. I can go ahead and shoot him an email about it, though.

The guy works for memoryexpousa.com. They have Turion listed in their AMD mobile section. If the prices are anywhere as reasonable as their current mobiles, I think this would be a great place to get Turion.

[perkam]

But the question remains, will we be able to buy it off retailers in quantities AND are we perfectly sure that they will be fully compatible with desktop boards.

Cause if they'll work with 754 desktop boards, then I finally understand what the fuss behind Turion is all about ;)

Perkam

[WeStSiDePLaYa]

But the question remains, will we be able to buy it off retailers in quantities AND are we perfectly sure that they will be fully compatible with desktop boards.

Cause if they'll work with 754 desktop boards, then I finally understand what the fuss behind Turion is all about ;)

Perkam


oh im getting one, im sure the dfi guys will get it to work ;)

[perkam]

oh im getting one, im sure the dfi guys will get it to work

I know, it just might delay my Intel DDR2 upgrade. As long as AMD keeps SSE3 in them :up:

Perkam

[frostedflakes]

The DFI supports all current mobiles. If the existing BIOS versions don't support Turion, I'm sure they'll release one that does.

Although none of the motherboard manufacturers want to admit it publicly, I think many quite a few of the most popular have included microcode support for Athlon64 mobiles. I think the Chaintech VNF3-250 and MSI K8N Neo are also compatible with low-power mobiles.

[saaya]

Saaya , ull be surprised how many winchesters are stable while undervolting. I myself have owned two winchesters , a 3000+ and a 3500+ , both were stable at 1.158 V and 1.2 V respectively . By stable I mean any given benchmark , 3d or prime95 or S&M or stresscpu. They were completely stable at stock speeds.
I undervolt my processor as a part of my burn-in process , thanx to Sentinel who has posted in this very thread and whose burn in guide is posted over at OC Forums. I have had many pms requesting me to make a burn in guide and I am currently working on one. I hope Sentinel doesnt mind , its basically his concept which has worked great for me till now . :toast:

yeah, ive seen several chips at 2ghz+ with 1v stable, i didnt know that current chips already clock that sweet with low vcores! but those were all really good chips, if this is average for a turion it will revive the 754 market in no time! :D

and about availability... i think it will be the same as with current mobile chips, they are not everywhere, but you can buy them for a decent price in big shops :) might just take some time before they are in stock :)

and i hope amd can cope with the huge demand so prices dont shoot up :D

and about bios support, dfi will definately support them, and if they do all the bios modders will see how they changed the bios to work with a turion and can hopefully apply the same changes and transfair the code to the bios files of other 754 boards :)
prepare for a lot of work tictac! :D

[perkam]

Cos they're single channel, does that mean they'll lose points in 3dmark 01/03 or 05 ??? or does single channel affect encoding audio video etc ???


Perkam

[Mats]

Although none of the motherboard manufacturers want to admit it publicly.....

I've seen one exception, MSI K8NM. Although I wouldn't call it "official support".
http://www.msi.com.tw/program/support/bios/bos/spt_bos_detail.php?UID=622&kind=1

[DrJay]

But the question remains, will we be able to buy it off retailers in quantities

How many you lookin' for?

[perkam]

How many you lookin' for?

LOL only one, but I might not buy one until the Turion 3400+ comes along. That would some nice 90nm+SSE3 sweetness :D

Perkam

[elec999]

Will this cpu cost alot when its released.
Thanks

[LowRun]

Will this cpu cost alot when its released.
Thanks

Yes

[TEDY]

damn it why did i move to 939 system then ??

why not 939 mobile winchester ?

[Bloody_Sorcerer]

Cos they're single channel, does that mean they'll lose points in 3dmark 01/03 or 05 ??? or does single channel affect encoding audio video etc ???


Perkam
with athlon 64s, the performance difference between single-channel and dual-channel is somewhere around 5%-10% in most things, though with specific benchmarks i can't say for certain.

[perkam]

Yes

It would be a direct competitor to the dothans so dont expect it to be cheap ;) Expected prices should be something like:

~$200 for Turion64 2800+
~$250 for Turion64 3000+
~$300 for Turion64 3200+

Remember, the only other procs in the next few months that will have SSE3 and 1mb cache will be the 4200+ :eek: so cheap is out of the question :up:

EDIT: Prices revised cos 3700+ costs $340 (which is amazing) :) Interestingly enough INtel's dothans cant do double channel either, afaik, so that will not be a reason for a lower price, neither will be the fact that most other newer procs on 754 will be value semprons.

Perkam

[Franky 4 Finger]

I thought dothans were only limited by their current chipset. I believe that they will have dual channel on the Alvisonm chipset.

[ss284]

Franky is right, the alviso chipset supports dual channel with dothans.

[perkam]

Franky is right, the alviso chipset supports dual channel with dothans.

Probably, and thats also why they are and will continue to be unreasonably priced whereas the Turion64s (as they wont be comign 939) will be much more competitively priced.

Perkam

[Franky 4 Finger]

They arent unreasonably expensive until you get over 2ghz.

[perkam]

They arent unreasonably expensive until you get over 2ghz.

LOL The only ones that arent unreasonable are the 725 (1.6) and 735(1.7) and they too are $50 to $100 more than Winnies at the same speed rating. And the 745 costs more than the 3800+ in addition to the fact that mobos for these procs are same cost as the DFI LP SLI-DR ($230) and dont even have dual-channel. And the alviso has yet to come and will be even more expensive due to exclusivity :rolleyes: so pls dont tell me it isnt unreasonable.

Perkam

[ozzimark]

remember that amd has to take market share away from intel, so for about equivilent performance, the chip should be much cheaper (similar to how it is on the desktop right now, minus the extreme high end, fx's and EE's)

[perkam]

1x512 crucial ballistix pc3200 [230mhz :: 2.5-2-2-10 - 2.6v]

Whoa !!! Ozzimark, thats ballistix u have there, give it some volts man, 230 at cas 2.5 is...well un-xtreme :D

And yes you're right Turions will make great competitors specially at competitive prices, though I'm still waiting for more news on its feature-set.

Perkam

[ozzimark]

Whoa !!! Ozzimark, thats ballistix u have there, give it some volts man, 230 at cas 2.5 is...well un-xtreme :D
a limit of my memory controller. i need 2t to go much past 230mhz now
i used to be able to do 245 at 2.65v.. it's been slowly declining
and 2-2-2 only gets me to 209mhz. my old bh-5 that can do 220 at 2-2-2 with 2.8v on an nf2 acts the same way. :stick:

(read: i'm dying for a new cpu, but i'm holding out for a mobile 90nm :toast: )

[_Eduard_]

Anyone have any idea how long it will take untill these Turions hit the market?

[perkam]

Anyone have any idea how long it will take untill these Turions hit the market?

EDIT: See thread here: http://www.xtremesystems.org/forums/showthread.php?t=52361

April 18th seems to be it ;)

Perkam

[Rabbi_NZ]

Man I want one!


How many you lookin' for?
can you get them DrJay?

[frostedflakes]

I'm interested too. Why not do a bulk order and sell them to XS members? If you'd be willing to sell them for a good price (at or slightly over cost) I'm sure there would be people here willing to buy them. I know I would be one.

[perkam]

Why not do a bulk order and sell them to XS members?

That you only need to do with really expensive cpus and those ones you dont care about the steppings for like xeons, ppl here like to search for their own procs just so that they can get better steppings, get better prices or get a combo or something. It also leaves the commitment part out of it so you can make your decision on your own time.

I dont think that'll be an issue really. But I just hope ALL turion's clock well and not just the ones with good steppings. cos if i see a SINGLE CBBID TURION LOL, I"M GONNA :flame: it !!

Perkam

[frostedflakes]

I'm just afraid that only a handful of vendors will carry these, and because of this the few that do will sell them for rediculously high prices. :(

[JDizzle]

Looks like AMD has launched the Turion. Who's going to be the first person to buy one?



http://img.photobucket.com/albums/v239/Johnieboi/Turion.jpg

AMD Official Link (http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~95256,00.html)

Anantech Link (http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=2374)

[Tenknics]

More Turion news:

http://anandtech.com/cpuchipsets/showdoc.aspx?i=2374

[perkam]

Interesting, they're prices are quite competitive relative to Dothans, but can we say the same about performance ;) ??

Prices:

35W Models
AMD Turion 64 ML-37 2.0Ghz, 1MB $354
AMD Turion 64 ML-34 1.8Ghz, 1MB $263
AMD Turion 64 ML-32 1.8Ghz, 512K $220
AMD Turion 64 ML-30 1.6Ghz, 1MB $184

25W Models
AMD Turion 64 MT-34 1.8Ghz, 1MB $268
AMD Turion 64 MT-32 1.8Ghz, 512K $225
AMD Turion 64 MT-30 1.6Ghz, 1MB $189

25-Watt models more expensive by only $5.00 :up:

Perkam

[ozzimark]

hmmm.. i'm very tempted to go with the mt-34.
any word on the voltage for the 25 watt models? i'm surprised that they're only $5 more.. i was expecting like 25-50 :stick:

[Franky 4 Finger]

I wonder if they can compete when Dothan goes dual channel

[=[PULSAR]=]

Does anyone have a clue of the max operating temps of these chips?

[ozzimark]

need only post that once :D

look at the opn. the X before the 5LD indicates the same 95º max die temp as the old mobile chips.

[isp]

So maybe my old s754 board is good for something after all... ??? :hehe:
Let's speculate on the default voltage... 1.25? 1.3?

[=[PULSAR]=]

I believe its 1.4 :)

[ozzimark]

k = 1.35v :toast:

i did some calculations that imply that the 25w line needs 1.15v to meet the tdp.

[_Eduard_]

w00t. Definately getting the MT-30, only like 200 euros :slobber: and it'll do at least 2.5ghz im SURE

[perkam]

Hmmm....from what I've been reading the 512kb ones are Oakvilles and the 1MB ones are Lancasters which isnt good since Lancasters are the super winnies ;) Can someone correct me if I'm wrong ??

Perkam

[frostedflakes]

Technically I guess they're Oakvilles, but they won't be like the current D0 Oakvilles. They'll use the E3 core. Basically a Lancaster with 1/2 the cache.

[perkam]

Basically a Lancaster with 1/2 the cache.

Are you perfectly sure it isnt a different core ?

Perkam

[frostedflakes]

I'm 99% sure. The whole Turion line is marketed as having SSE3 instructions, and current Winchesters don't have these features. So I couldn't see what else it would be other than a Lancaster core with 512kb L2.

[ozzimark]

Technically I guess they're Oakvilles, but they won't be like the current D0 Oakvilles. They'll use the E3 core. Basically a Lancaster with 1/2 the cache.
oakville is still based on the D0 core that the winchesters are though...

i still think lancaster is the 512kb part while newark is the 1mb part.

[krampak]

And which software uses SSE3?

[ozzimark]

i don't really know of much yet.. i think some video encoding stuff does

edit: just saw the new amd roadmap (http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html).. the newark core hasn't been assigned to anything yet, so.... i dunno what's up with it. probably the cheaper and higher wattage 'mobiles'

[computersmsa]

w00t. Definately getting the MT-30, only like 200 euros :slobber: and it'll do at least 2.5ghz im SURE
200 € is VERY expensive ;)
A good Winchester 3000+ @ 130 € can 2.7 GHz en air with Dual Channel and no prob with 2 stick of 512 Mo ;)

[_Eduard_]

200 € is VERY expensive ;)
A good Winchester 3000+ @ 130 € can 2.7 GHz en air with Dual Channel and no prob with 2 stick of 512 Mo ;)

ok, first of all that is a "good" winchester. I'm sure a "good" Turion could do near 3GHz. furthermore, the Turion has 1mb cache, support for more RAM, SSE3, etc

[DrJay]

can you get them DrJay?

No, I was just being a wise @ss.

[americium]

So, Turion been anounced 2 days ago. Wonder if sxs212 is going to let usk now which mobo was he using with that Turion sample he got ;)

[c7775]

well be getting in a couple to sample within two weeks so if anyone
wants a Turion64 2800+ you can PM me and ill see what i can do :toast:

[tritium]

Does any one know if these new processors use Strained Silicon.

[c7775]

enough interest generated :)
whoever PMed me until know, ill contact you
if you did not , then im sorry but we wont be getting
enough chips :)

[ozzimark]

Does any one know if these new processors use Strained Silicon.
it's not confirmed, but i am under the assumption that amd got DSL (http://arstechnica.com/news.ars/post/20041213-4459.html) working with these chips. i don't know if they'll actually tell us if they have or not though. :confused:

[Revv23]

so will upwards multi's be locked?

thats all i want to know.

[frostedflakes]

current mobiles are only half-locked like the desktop chips, so I'd assume it's downward multis only.

[dcg]

w00t. Definately getting the MT-30, only like 200 euros :slobber: and it'll do at least 2.5ghz im SURE

Well, looking a this mobile clawhammer 3400+ (0.13) I really hope Turions will do better than 2.5GHz :)

http://crazypc.ro/forum/index.php?showtopic=710&st=0&#entry9759

3GHz of good ol' clawhammer power, maybe Turions don't look so tempting after all :slobber:

yes, it is romanian, but the screenshots do speak loud and clear :toast:

[Snowman89]

i thougt so as well, but it reads the volts correctly on the first cpu-z screenie and 2ghz with 1v isnt THAT far off, ive seen 2ghz with 1.2v already...

happy bday sxs btw! :toast: :D
well ive seen 2ghz with 1.15v :p:

[shuRe]

the new A64 dtr's seem dam good when it comes to o/cing. Seen loads of 3400+ at 2.7g lately.

Do you think its possible that these new dtr's are AMD experimenting with early versions of turions? or are turions a completely different core?

[jjcom]

Turions are suposed to be a different core than the regular A64, I believe. Correct me if I'm worng. The DTR chips that are being talked about are 130nm and the Turons are 90nm so, their different cores.

jjcom

[ozzimark]

i'm pretty sure jjcom is right.. the old mobiles are just the best of the yeilds in order to make the thermal bin, as a side effect, they overclock really well too.
i'm hoping the turion is similar and doesn't turn out to be a chip that has a very defined top end and isn't insensitive to voltage. time will tell though :toast:

[dcg]

the new A64 dtr's seem dam good when it comes to o/cing. Seen loads of 3400+ at 2.7g lately.


The link i posted is not about a DTR chip..is a true mobile. Sorry if i went off topic :)

[Rabbi_NZ]

Would all those considering buying a Turion for a DFI nF3 LanPartyUT please let me know via PM.


Cheers :toast:

[ozzimark]

are you making a petition?
i already fired off an email to msi tech support and i'm trying to gather interest in the msi forums.
i guess i can try to do the same for the k8n neo platinum, though i don't think as many people here use that board :toast:

[perkam]

Turions are suposed to be a different core than the regular A64, I believe. Correct me if I'm worng. The DTR chips that are being talked about are 130nm and the Turons are 90nm so, their different cores.

Yeah as was stated before by someone Turions are 90nm SOI Lancasters with 1M cache. The 512K versions of the Lancasters are called Oakvilles ;) which is the successor of the Odessa core: 130nm mobile claws everyone above is referring to.

And yes they're confirmed SSE3, Strained Silicon, and 35-25W Voltage Processors.

Perkam

[Grov]

Heheh cool. Might get one if they o/c nicely. though probly not worth it. :p:

[perkam]

though probly not worth it.

Unless u dont mind a score of 40+ even with 3ghz Turion, yes dual channel cripples Turion for desktop implementation but low power requirement means a lot more OCing coming in the way of notebooks ocers.

Perkam

[ozzimark]

And yes they're confirmed SSE3, Strained Silicon, and 35-25W Voltage Processors.
where did you see that? i've been looking everywhere to find an official source that says that amd is actually putting strained silicon into these, but i haven't found any.

[Rabbi_NZ]

are you making a petition?
i already fired off an email to msi tech support and i'm trying to gather interest in the msi forums.
i guess i can try to do the same for the k8n neo platinum, though i don't think as many people here use that board :toast:
not so much a petition as support/help in getting the new BIOS... but if you want to know more please PM, as I wont discuss in open forum

[Penti]

And yes they're confirmed SSE3, Strained Silicon, and 35-25W Voltage Processors.


Well not really, they don't use strained silicon wafers, they use the technology they developed with IBM. Strained Silicon-On-Insulator (sSOI wafers that is). But otherwise is probably totaly correct that this is the first product with it from AMD. There plenty of pages regarding sSOI, includings both IBMs and AMDs.


/me still waiting for dual core laptops :D

[ozzimark]

Dual Stress Liners. i believe it's a form of stress that uses both expansion and comression for the various types of transistors (high voltage and low voltage threshold iirc). google it and you'll come up with something.
i just wasn't sure if rev E had it cause i haven't seen any official word yet.

[_Eduard_]

Whatever it is, it can't be worse :toast:

[frostedflakes]

I remember reading in an article that Troy (rev E4 Opteron) had it, so I guess one could assume all the new revisions do as well. :)

[wickedld9]

Unless u dont mind a score of 40+ even with 3ghz Turion, yes dual channel cripples Turion for desktop implementation but low power requirement means a lot more OCing coming in the way of notebooks ocers.

Perkam


score of 40+ what?

[c7775]

friday ill have some updates once we set up the turions up.

confirmed as arriving : 1X 1.8Ghz 25watt , 1X 1.6GHz 35 watt

[frostedflakes]

Nice. :)

What boards you going to try them in? If they work, I'd very very interested in knowing how the 1.8GHz 25w overclocks (this is the 1mb L2 version, and not 512kb?), as this is what I would probably buy.

[ozzimark]

also, what is the rated voltage for the 25w part?

[Mats]

also, what is the rated voltage for the 25w part?
Until we know I'm guessing 1.2 V (= "Q" in OPN).
The 35 W ML uses 1.35 V and goes up to 2.4 GHz.
The 25 W MT uses 1.?? V and goes up to 2.2 GHz.
I checked CPUPower (http://www.silentpcreview.com/Web_Links+index-req-visit-lid-137.html) (must use comma for decimals): (35 : 1,35 : 2400 @ 1,2 : 2200) = 25.3 W

[ozzimark]

i was thinking a 1.15v part, but that seems way too low for a stock voltage at 2.2ghz

[c7775]

its a 1.1V

and only overclocking ill do is to change the multiplier, our pcbs dont
allow for fsb increases ... will wait until Said goes around to modding
the bios of our reference msi boards to work with those puppies :toast:

[ozzimark]

considering how my conversation with msi tech support is going, it looks like someone is going to need to mod the k8n neo bios to get the turion working. it might already, i don't know, but they're not being very receptive, and claiming that the system won't boot due to the low stock voltages :stick:


but as a side note, 1.8ghz on 1.1v!!! omg, nice.

[frostedflakes]

1.1v is very yummy... I think my mind is made up.

As long as it's under $300, I'll be picking up an MT-34. :)

[c7775]

considering how my conversation with msi tech support is going, it looks like someone is going to need to mod the k8n neo bios to get the turion working. it might already, i don't know, but they're not being very receptive, and claiming that the system won't boot due to the low stock voltages :stick:


but as a side note, 1.8ghz on 1.1v!!! omg, nice.

ill let you know how the neo works with teh chip by early next week at the latest m8 :) take care

[ozzimark]

thanks, i appreciate the effort. seriously :toast:

*figures out a way to work for amd*
:banana: :banana:

[Mats]

1.1 V???
Heh, awesome... well they have the possibility to ramp up to 1.2 V when the faster MT comes out. The question is if the ML will stay at 1.35 V. If so, then this 1.8 GHz @ 1.1 V CPU have a TDP = 17.4 W... :toast:

[ozzimark]

where do you get this 17.4w from?
undervolt a 35w chip from 1.35v to 1.1v..
wattage = 35 * (1.1 / 1.35)² = 23w :stick:

[Mats]

where do you get this 17.4w from?
undervolt a 35w chip from 1.35v to 1.1v..
wattage = 35 * (1.1 / 1.35)² = 23w :stick:
You forgot the speed drop from 2.4 to 1.8 GHz.
:stick:

[ozzimark]

oh, my bad, i though he was trying to say something else.

though it definitly looks like the MT will be more speed binned and ML will be more thermally binned.

[trans am]

Has anyone seen a turion for sale yet?

[isp]

I haven't, anyone know ETA?

[ozzimark]

april 15th iirc.

[Mats]

ill let you know how the neo works with teh chip by early next week at the latest m8 :) take care

Any news?

[ozzimark]

agreed. i'm dying to know how these turions are working out :banana:

[krille]

Ummm... how come this thread has reached the news just now? I mean, it's rather old news, this thread has been around for quite some while?

[trans am]

I called Monarch Computer and asked about the Turions and he wasn't sure what I was talking about. they have been getting tons of calls about the Venice. I told them thry should order some. Since Monarch is a system builder, they should have no problems getting them he said. So hopefully by mid april they will have some Turries! Now that we are on the subject of 90 nm 754, has anyone here played with a Palermo Sempron? I was thinking of getting a cheap one for $80 just to mess around with.
http://www.xbitlabs.com/articles/cpu/display/sempron-2600.html
http://www.newegg.com/app/ViewProductDesc.asp?description=19-104-226&depa=1

[matt9669]

Ummm... how come this thread has reached the news just now? I mean, it's rather old news, this thread has been around for quite some while?A glitch, goof, error, mixup, mistake. Speak of it again and I shall be forced to :slapass: :lol:

[krille]

A glitch, goof, error, mixup, mistake. Speak of it again and I shall be forced to :slapass: :lol:

Sorry, Sir! Won't do no more, Sir! Thank you, Sir! :eek:

[[XC] leviathan18]

i want to see acer ferrari with turion 64 :O

[_Eduard_]

so basicly these turions are san diego s754 chips with single channel?

[ozzimark]

yes, but some of them have 512kb of cache instead of 1mb

[saaya]

so basicly these turions are san diego s754 chips with single channel?

amd told me theres a difference, there are different transistors you can chose when building a cpu, some consume more power and switch faster, some consume less power and switch slower, just an easy example.

amd told me that the turion uses a different set of transistors than the regular a64s in 90nm and uses around 30% less power in some situations than a desktop a64 with the same vcore and clockspeed.

i dont know if this i true though, or if its just marketing :D
after all having a different mask for the turion cpus ould mean a lot of extra work...

[saaya]

since turion is now released and there are/will be reviews of turion cpus already/shortly i unstuck this thread to make some more room for new stickies in this section :D

[cadaveca]

have you seen any for sale? i'll try one out! (gotta stop buying cpu's, and get some other parts..3 cpu's extra is too many already)

[tristancarton]

saaya, any chance you could throw us a link to where they are available/released? i can't find them anywhere (presale or not), I thought that they weren't going to be released until mid april? (perhaps a difference on semantics?)

...since turion is now released and there are/will be reviews of turion cpus already/shortly...

[Revv23]

april 4th i belive is the correct date.

[Raul]

is there little, little chance to AMD release Turion on s939 ? NO ? :(

[HKPolice]

Does anyone have OC results yet (More than 2Ghz)? The latest article from anandtech:

Much as we suspected, all of the power optimizations that went "into" Turion 64 are all transistor level optimizations. Basically selecting transistors that provide better thermal and power characteristics at the expense of lower switching frequencies. Given that the Turion 64 runs at multiple speed grades lower than the fastest desktop Athlon 64s, this tradeoff makes sense, but it also means that Turion 64 is no Pentium M killer.

http://anandtech.com/cpuchipsets/showdoc.aspx?i=2387&p=4

It might hit a 'wall' when going above ~2.5ghz or something...

[ozzimark]

yeah, it appears that way :stick:
oh well. we can always hope. if not.. there's still newark :banana:

[robberbaron]

is there little, little chance to AMD release Turion on s939 ? NO ? :(

No, the successor to the Turion will go on a socket similar or the same as M2.

[Raul]

pitty, but Stay Tuned :D Venice is arrive soon :D

[_Eduard_]

too bad. I was hoping for the new "athlon XP mobile" under the A64's

[Rabbi_NZ]

a bit premature aren't we?

I'm going to reserve judgement until there are some actual user results coming through...

[majkel84]

a bit premature aren't we?

I'm going to reserve judgement until there are some actual user results coming through...

has anybody already got Turion?
They aren't avalible anywhere, are they ??

[perkam]

Turions sure have funny steppings :D lol "Cabse, wait for me !!!" lol

Though this is a *P*W so it should be one of the ones that oc well.

http://www.watch.impress.co.jp/akiba/hotline/20050409/image/nt644.jpg

In any case they're coming and should be bringing updated chipsets to the 754.

Perkam

[Mats]

Though this is a *P*W so it should be one of the ones that oc well.
Dunno, but I thought *P*W was the most common (http://dissolved.50megs.com/939Winchesters.htm) on A64. AFAIK, that code only tells you when it's manufactured,
and "CABSE" being the part of the OPN that's really interesting, just like CBBHD, AQYHA and JIUHB was.

[ozzimark]

iirc, all amd proc's are xPxW :toast:

[Mats]

iirc, all amd proc's are xPxW :toast:
Yeah, me too. Maybe there are some typos in the OC database....

Not every CPU got a four letter code (http://www.xtremesystems.org/forums/showthread.php?t=56248):

In the MPM the first M means that the assembly lots have been combined and not made on one specific day thats all. The second M means that the fabrication cores came from different batches too. The lack of a W at the end means that wafer lots used in the process could be mixed. Nothing to do with "speed re-binning". The P is for the assembly plant of Penang.
The CBBHD or CBBID fabrication stepping is way more important.

[ingentingmendeg]

http://img221.exs.cx/img221/1130/dscn02150jx.jpg
AMD Turion 64 2800+ - 1.6 GHz - 1M L2 - Core 90nm Rev 'E' - TDP 25W Socket754

http://img169.exs.cx/img169/3698/16002oh.jpg
1600Mhz


http://img169.exs.cx/img169/5719/25003cx.jpg
2000Mhz

Tnanks Taiwan 阿格斯 Test

Source (http://www.coolaler.com/ipb/index.php?showtopic=42810)

hmm, looks similar to the Dothan

[frostedflakes]

Some new Turion-related info:

DFI recently released the 4/8 BIOS for NF3 250Gb motherboards. One of the most interesting features is revision E0 support. Hopefully Turion will be able to install in these boards without much or any trouble.

As far as availability goes, memoryexpousa.com claims they are working on getting them. The impression myself and others got from talking with them is that it's not a matter of if, but when. I'd hope another couple weeks to a month and they should be available. Not sure what models they plan to carry; hopefully all of them.

As the pic above points out, all the 25w models will begin with TMS. The 35w models are TMD. Voltage appears to be 1.2v ("Q") for the 25w models.

[Rezag1000]

any word on pricing and what K8 HSF will be compatable with turion?

[frostedflakes]

http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_609,00.html

From what I gather, though, these are just recommended prices that are set by AMD. In the end, I think it all comes down to what the store selling the product decides to charge. But expect them to be around those figures.

The only two heatsinks I can think of that work with mobiles out-of-the-box are the Thermalright SLK-948U and Alpha PAL8150. You can make almost any other heatsink work with minor modifications.

[Fzero]

will these only come in S754 or will there be a S939 version?

[frostedflakes]

AFAIK all mobiles will stay on 754 until M2 comes in 2006. Then AMD will end this multiple socket madness and all desktop and mobile processors will consolidate onto one socket, M2. Opterons will have their own socket.

[ozzimark]

i believe the opteron socket weill be S1 (presumable s for server, m for mainstream)



Yeah, me too. Maybe there are some typos in the OC database....

Not every CPU got a four letter code (http://www.xtremesystems.org/forums/showthread.php?t=56248):
ah, i forgot about those mixed cpu's. though i do recall seeing some that are MPMW, which really confused me.

[Rezag1000]

what kind of minor modifications are we talking about? Anything too complicated for a first time builder?

[ozzimark]

just tweaking the heatsink so that it applies the same pressure about 2mm lower (the thickness of the missing heatspreader)

it might be a bit intimidating at first, but if you know what you're trying to do, it's relatively simple.

[tristancarton]

just to let you guys know i sent the email below (or a variation) to these sites on saturday. newegg, monarchcomputers, zipzoomfly, mwave, chiefvalue, and frontierpc. some of the emails i sent to both the purchasing and service areas. so far i only have emails back from both monarch sales and service. i will update this post with the responses as i get them.

--
email template that i sent out.(small variations depending on store)
Newegg,
Long time customer waiting for AMD’s new mobile socket 754 chip to come in stock at newegg. They were announced by AMD as available and with prices on March 10th. ( http://www.amd.com/pricing ) Any chance you could inform me when (and/or if) you will have these in stock?
my name
my email address
--
Monarch sales response 4/11
my name,
Presently we do not have an official ETA on that chip.
Thanks,
Eric
--
Monarch service response 4/11
Thank you for your inquiry.

We have not received definitive information on the Turion processors. As soon as we do we will have it on the website.

Let me know if there is anything else I can help you with. Sincerely, Rick Monarch Computer Sales
770-939-8886
800-611-0875
--
Newegg service response 4/11
Dear Valued Customer,

Thank you for contacting Newegg. All information we are provided with including pricing and product availability is listed on our website in a real time format. Our inventory runs so fast that we are unable to tell exactly what products is in stock and their prices.
We generally recommend customers utilize the SHOP BY CATEGORY and SHOP BY BRAND links at the top of our site because they usually work best when trying to locate a specific type of product or particular make/model. After narrowing your search with one of the aforementioned links, you may also be presented with ADVANCED SEARCH options which are very useful when shopping for certain product specifications.

Here is a link just for your reference. http://www.newegg.com/app/viewProductDesc.asp?description=19-103-477&depa=0 Please note, we post every bit of information the manufacturer provides us with for your convenience. We can only guarantee what is on our website.

If you can not find the item you want on our website, you may submit a product suggestion to our marketing and purchasing departments directly, please visit www.newegg.com and click on the YOUR IDEAS https://secure.newegg.com/app/ShallWeCarry.asp box on the left side of our home page. Our standards of excellence allow Newegg the ability to provide cutting-edge products at bargain prices. We are experts on delivering the best the world has to offer so we always welcome your ideas.

If you have any further questions or concerns, please visit our FAQs page.
--
Newegg purchasing response
waiting....
--
Chiefvalue service response 4/11
Dear Valued Customer,
Thank you for your most recent e-mail.We currently do not have any information on the availability of this item. We will post this item on our website if or when it becomes available.

Thank you,

Edgar Zambrano
Chief Value.Com Customer Representative
888-611-4165
--
zipzoomfly service response
waiting....
--
mwave service/product response
waiting....
--
frontierpc tech response
waiting....
--
these will be updated as i get responses back

hope this helps guys. i really want one of these cpus. also as an fyi the new dfi nf3 beta bios 2005/04/08 found here http://www.dfi.com.tw/Support/Download/bios_download_us.jsp?PRODUCT_ID=2840&STATUS_FLAG=B&SITE=NA supposedly supports new EO a64s. (although not sure if this in relation to new 90nm eo semprons, turions, or both) props to felipev on dfi-street for seeing this.

Major Reasons of Change:
1. Update Memtest v1.55.
2. Fix certain USB device cause system hang up.
3. Fix C&Q failure with certain memory modules.
4. Support New E0 Stepping Athlon64. (oskar wu rules :) )

[Revv23]

^very nice works sir, this will be very helpful indeed

[Mats]

I wonder how much those slower transistors will affect performance, maybe that's why Turions won't have the traditional P-rating (because Turion 2600+ wouldn't look good). I really hope it's because that P-rating is just outdated and confusing, otherwise we can all go and buy a cheap Oakville, or a Sempron...

[frostedflakes]

To be honest, I don't know if I even believe that whole transistor thing. Sounds like marketing BS to me. Guess we'll see, though. I plan to purchase an MT-34 as soon as they're available, and will of course be sure to post results.

[Revv23]

To be honest, I don't know if I even believe that whole transistor thing. Sounds like marketing BS to me. Guess we'll see, though. I plan to purchase an MT-34 as soon as they're available, and will of course be sure to post results.


i agree

especially when i see a chip doing 2.0ghz at 1.0v

[ozzimark]

--
Newegg service response
waiting....
--
Newegg purchasing response
waiting....
--
a while ago (3/14 actually) i got this from them regarding carrying turions

"We are currently in the process of carrying this item. We will post any information that becomes available on the website. If you have any further questions or concerns please do not hesitate to contact our customer service at 800-390-1119 or email us at service@newegg.com."


edit:

To be honest, I don't know if I even believe that whole transistor thing. Sounds like marketing BS to me. Guess we'll see, though. I plan to purchase an MT-34 as soon as they're available, and will of course be sure to post results.
just out of curiosity, why do you think the cto of amd would say that unless it's the truth?

[tristancarton]

ozzi, marketing. pure marketing. the main marketing push against amd mobile chips is their power consumption. pentium-m chips consume amazingly low amounts of power.(and are still relatively fast) if amd comes out with new tech saying "we have special amazing low power silicon" it will look good.

the above and one more thing. from what i have heard many 90nm a64 s939 cpus can 1.6-2 ghz speeds at low voltages already. why mess with a process if the current process can hit these consumption limits.

in any case we should know soon. (at the latest hopefully a few weeks) btw the response above is pure speculation on my half.

just out of curiosity, why do you think the cto of amd would say that unless it's the truth?

[ozzimark]

i see. i was looking at it from an enthusiast standpoint, so that makes thing a bit different..

however if it's a normal revE chip, it's thermally binned.. quite heavily. seeing as the TDP for the normal revE chips appears to be 67w at 1.4v.. dropping that to 1.35 (stock voltage of 35w model) only brings down the wattage to 62.3w
the rest of it has to come from either the fact that it's slower, or the cores themselves are that good that it puts out that little heat.. both of which may be able to amount to almost 30w, but it's kinda iffy.


that's just my view on it though, which is why i do believe that turion is made a bit different than the rest of the revE chips.
also note that there are 1mb and 512kb chips. i would suppose that they're all make with 1mb, but the ones with a defect in the cache get turned into the 512kb parts, effectively improving yeilds and making a seperate line more justifiable

[frostedflakes]

Not to mention the statement might discourage some enthusiasts from purchasing Turion for desktop boards. And I think we are all aware of AMD's stance on mobiles. They plain do not want them to be used in anything other than laptops.

Also, the 67w TDP number is the thermal envelope, and includes all current and future processors. This means 3800+ and beyond. The 3800+ is 2.4GHz, and future chips may go as high as 2.8GHz or 3.0GHz for all we know. This is why TDP is so high. It's a loose number to cover an entire line.

On the other hand, Turion has a defined thermal limit. We already know the max clock rate, 2.4GHz.

[Mats]

I wouldn't call it pure marketing, since it runs slower AFAIK (someone tell me if I'm wrong...). I thought they meant lower power and performance at the same speed, but that sounds like real BS, and really stupid to use in a mobile CPU. I mean come on, it's like starting a new MHz war just to get the attention in the market where that's the least important. You're right guys, let's wait and see, as usual...
Now we got FX, ML and MT. I guess the regular desktop chips will follow when they switch to socket M2.

[Mats]

Ingentingmendeg: Do you know Swedish :toast: , or where did you get your nick from?

[frostedflakes]

Mats, that's the impression I got also. That they were using transistors which switched slower (performed worse than "normal" transistors), but had better thermal characteristiscs (ran cooler). Like I said, I don't know if I believe it. It just doesn't make any sense that AMD would start manufacturing a completely different CPU when current Venice/San Diego can easily be undervolted and rebadged as mobiles.

[Mats]

What happened with Voodoo|Minions T64?????
Have I missed something?

[ozzimark]

i was wondering the same thing mats. i was one of the few who was lucky enough to contact him before it was too late :toast:

fr0stedflakes: what you say does make a lot of sense. though various chips put out verying amounts of heat at the same speed. i'm not sure how much, but i think enough that some of the faster running chips are thermally binned to slower speeds to make the TDP.

also, not trying to disagree with you, just consider this:
fastest 130nm chip at 89w is 2.4ghz. at 2.6, it got bumped to 104w. from that, i deduce that the 89w figure is meant for the 2.4ghz parts
then consider the 3400+ mobile: 1.4v, 2.2ghz and 62w.
bumping a 2.4ghz a64 to 2.2 and 1.4v would drop the power to 71w, not the 62w that the mobile rated for... somewhere amd is either getting cores that really do put out a bunch less heat, or there is some room left in the TDP figure, and it was needlessly raised for the fx-55.

but then again, it is known that the fx-55 has strained silicon, while the normal a64's don't. i don't know what happens to the cores that don't do 2.6ghz, but i won't go into that part of my theory for now.
but, if what the cto said about lower power transistors is true, then it would make sense that the same is true for the fx-55, thus it has higher power, faster transitors in it to allow for better yeilds at 2.6ghz. this would also explain the larger increase than expected in TDP when going from 2.4ghz to 2.6ghz ;)

again, both are just my wildly speculative theories, take with a grain of salt :toast:

[jjcom]

does make sense tho, if AMD has slightly different process changes made for different A64s.

jjcom

[Gogeta]

A 3200+ 35W Lancaster with 1x512 BH5 in a HTPC sounds REALLY nice to me. :)

[frostedflakes]

Could be. But I believe the reason for higher TDP with the FX-55 was the use of strained silicon, not faster transistors. Note that Venice, which employ DSL (basically strained silicon, but under a different name), run slightly hotter than Winchester. Same # of transistors, though, and same type I'd assume. For whatever reason, strained silicon/DSL increases thermal output.

[ozzimark]

that's quite possible too, though i was under the impression that SS was supposed to reduce leakage some, thus lowering power consumption and heat. i might be getting my facts mixed up though

however, you better not be basing that hotter statement off of the one review that compared power consumption and temps :toast:

edit: this is getting waaayy o/t sorry :slap:

[frostedflakes]

What else are we going to talk about until Turion is available. :p

I guess I should probably wait until more reviews are out. I've only found one that compares Winchester to Venice at the same clock speed and voltage with the same cooling. Something still could've been messing up the results. The difference was pretty significant, though. About 2*C I think. It was the 3200+ Venice review at Xbit-labs.

[ozzimark]

2ºC (alt-167 for º ;)) isn't really much of a difference.. i've seen two different mounts with the same heatsink on the same cpu change more than 2º

though, assuming that it was the same both times, it appears that the venice that they had for testing is warmer than the winchester. i'd like to see comparisions between other venice and winchesters though, because if my thought on how no 2 cpu is the same in terms of heat output, it would kinda make any comparision between two chips void. i want an average of like.. 50 different chips :banana:

(ps, sorry for being so difficult. it's in my nature to disagree :toast: )

[frostedflakes]

Hey no problem man I may be jumping to conclusions.

[ozzimark]

same with me. or i could be way off base with my assumptions on how amd is making their chips.. though i've thought about it for a long time, and what i stated before is about the conclusion i've come to, and i'm not about to change my mind until i see some solid evidence ;)

though, i guess time will tell on how turion really does oc

[reject]

i want turion but im worried that my evap will crack the core :(

[ozzimark]

cores are more durable than most people make them out to be. just don't do anything stupid with it. i've been running my cpu with the core looking like this since late december. it's been through quite a few remounts and i move the case around a bit and nothing bad has happened yet ;)

http://www.freewebs.com/ozzimark/core.JPG
(note that it's not really clean there, thus not as shiny as it would be)

[reject]

:ROTF: hehe that makes me feel a bit more comfortable!

[frostedflakes]

Holy crap. Did your Zalman cause that?

I accidentally took a tiny chunk off the core of my old desktop 2800+. The chip still worked, but it did hurt overclockability.

[ozzimark]

nope. sandpaper did. if you ever lap your core, don't do it too much :slap:

[saaya]

woooaaaa that cpu still works? :0

soi = less leakage
ss = faster transistors = higher clockspeed :)

[reject]

im wanting a turion, if the die doesnt crack it will be easier to aim small evaps onto cause u dont need to get it perfect alignment
when they come out if ive gpot cash i can do the first -40 tests

[ozzimark]

woooaaaa that cpu still works? :0

soi = less leakage
ss = faster transistors = higher clockspeed :)
yup, it's my 24/7 rig

and thanks for the correction. i did a bit of research, and it seems that DSL increased and decreases leakage, depending on whether the transisor is being stretched or compressed.
http://arstechnica.com/news.ars/post/20041213-4459.html

[perkam]

Good Work keeping this thread active guys, especially ozzimark, cos they actually had to get it off sticky as it getting so unpopular.

Perkam

[ozzimark]

thanks perkam

one thing i would like to bring up though.. note in the roadmap (http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html), under the turion, replacing the mobile clawhammers is Newark. for a long time, i was thinking that Newark would be included with Turion, the Lancaster being the extra low voltage, 512kb parts, while Newark would be the slightly higher wattage, 1mb parts.
now that everything has fallen into place, and Newark still isn't claimed for anything, i'm trying to figure out what will happen to it.
perhaps we will see a 50-55w 90nm DTR part that will have better compatibility that is based on thermally binned san diegos, just like the DTR's and mobiles of before that were thermally binned clawhammers?

if anyone has any ideas, feel free to chip in :toast:

[Gogeta]

thanks perkam

one thing i would like to bring up though.. note in the roadmap (http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_608,00.html), under the turion, replacing the mobile clawhammers is Newark. for a long time, i was thinking that Newark would be included with Turion, the Lancaster being the extra low voltage, 512kb parts, while Newark would be the slightly higher wattage, 1mb parts.
now that everything has fallen into place, and Newark still isn't claimed for anything, i'm trying to figure out what will happen to it.
perhaps we will see a 50-55w 90nm DTR part that will have better compatibility that is based on thermally binned san diegos, just like the DTR's and mobiles of before that were thermally binned clawhammers?

if anyone has any ideas, feel free to chip in :toast:

That would be awesome. Some of the current 3400+ DTR Clawhammers clock like a dream!

[ozzimark]

that's my thought.

and my problem is... my birthday is in the end of may. i want to upgrade my clawhammer that has a terrible memory controller and can only clock to 2.3ghz before needing a terrible amount of voltage.

so, i either get a 3400+ mobile, or see what newark will really be and try to be one of the first and evangelize it's awesomeness :D

[tristancarton]

ozzi? do you have magical powers? amd announced the 3700 mobile today. http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_543~96667,00.html

wouldn't seem like too big of a deal but the feature model number comparison now shows the 3000,3200,3400, and 3700 on 90nm process and "3DNow!™ Professional technology, SSE2, SSE3".

http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_10220_10221^10222,00.html

so it seems as if an E stepping regular s754 mobile 90nm (12x multiplier :) 1mb cache) has just happened. *

* this was just announced so avaiability won't be for a long time. (turion was announced on march 10th and we still can't find them)


perhaps we will see a 50-55w 90nm DTR part that will have better compatibility that is based on thermally binned san diegos, just like the DTR's and mobiles of before that were thermally binned clawhammers?

if anyone has any ideas, feel free to chip in :toast:

[reject]

my cpu died today so im now defianalty in the running for some kind of s754 1mb chip!
the cpu is :toast: i hope they are out soon as i have learned its best to wait, but i cant wait long!

[frostedflakes]

90nm DTRs? Sweet. You called it ozzimark. :)

[ozzimark]

i'm just damn lucky when it comes to guessing :toast:

now it looks like my option from upgrading to a 3400+ mobile claw has changed to a 3700+ mobile newark! yay, i'm so happy.
now to find a way to pay the ~350 for the cpu

[ozzimark]

double post, i know.. BUT

the 3700+ that came out isn't DTR. that's been around for a while.
it's a real mobile A64
check the pricing
http://www.amd.com/us-en/Corporate/VirtualPressRoom/0,,51_104_609,00.html
before, the mobile line only went to 3400+ :toast:

[frostedflakes]

So it's a 62w mobile?

I wonder if it operates at a lower voltage than desktop San Diegos. AMD could probably pull off the 62w power rating at 1.4v (or 1.35v, whatever the rev. E desktop chips are).

[ozzimark]

i'm not sure. seeing as the mobile clawhammers were 27w lower (about 30%) than the normal destops.. that's where i get my 50-55w guess from before
(70% of 67 is actually 47, but i doubt it would go that low.. mostly because i think the voltage will only drop .05v instead of the .1v before)

[CrimeDog]

http://www.coolaler.com/ipb/index.php?showtopic=48113