Dual Core AMD Opteron. First screenshot!

[VVJ]

Hi guys! It’s interesting to see the screenshots of the AMD Athlon 64 X2 processor here. And I decided to make a small, but highly exclusive screenshot of the Dual Core AMD Opteron processor 275. AMD Corp. has permitted me to post it here, but only one screen, because my CBId tool shows too much confidential info, he-he :) OK, no problem. So, here it is:

http://cbid.amdclub.ru/work/cbid_opt275.gif

It’s a low-power 68W AMD Opteron processor based on the newest E6 core!

[agenda2005]

1.2V Vdimm??? MY gawd. What the hell is AMD doing. 2X1024MB L2 cache, priceless. 68W TDP. I'm short of words. Thanks VVJ. We want more than screenshot please.

[perkam]

OMG thats crazy. The number of features on that retarded...though are they gonna release single processor versions of that and how much is it going to cost :eek: :eek: :eek: !!!

At any rate, Thank you for making XS the place to show it exclusively, VVJ :up:

Though interestingly enough, how'd u become XIP with only 38 posts omg wow quite an accomplishment there.

However, if posts like that is how u did it, then I can understand ;) a post thats worth a thousand words....NICE :D

Perkam

[Poki]

Why is it showing HT? I assume it only shows it because it's a dualcore.... Are there any secrets that you can share?

[perkam]

Why is it showing HT? I assume it only shows it because it's a dualcore.... Are there any secrets that you can share?

Prolly why the Elite Bastards results showed a 4800+ X2 Dual Core kicking EE butt in Cinebench.

Makes sense, maybe AMD's way of splitting threads is getting more and more similar to Intel's for dual cores :confused:

Perkam

[Ailleur]

X86secret explored that and said that the HT bit is actually enabled on dual core amds to make use of software optimisations thaat were made for intel's hyperthreading

[ozzimark]

that's a pretty slick idea. hopefully it won't cause any compatibility issues.. though i don't see why it would.

[perkam]

that's a pretty slick idea. hopefully it won't cause any compatibility issues.. though i don't see why it would.

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam

[Gogeta]

Wow this is amazing. Slap two of those in a Tyan nF4 Opty board and hold on to your shorts!

[ozzimark]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation
that's true, but i don't see how intel could sue amd on the grounds of a bit being enabled. besides, they have some crazy cross licensing agreement, so i think amd might actually be able to put intel's HT ideas into their cpu, just have to call it something different (like x86-64 vs EMT64; same thing, different name)

:)

[Sentential]

Price? :woot:

[NiCKE^]

Wow looks amazing! Same as above, Price?

[agenda2005]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam

LAW SUITE? It's just an instruction set, no BIG DEAL about it. AMD have the right under cross-licenscing agreement to use Intel processor instructions. Why is AMD not sued for using SSE3 in Venice and SD which have HT instructions.

[[XC] moddolicous]

why does it say the processor core is italy?? Is that really what core it has? Thanks for sharing that with us.

[Ailleur]

why does it say the processor core is italy?? Is that really what core it has? Thanks for sharing that with us.

The program is only as reliable as the kid who wrote it so, whos to say.
Paris is semprom iirc, so, its probably wrong.

[notoriousformula]

Wow looks amazing! Same as above, Price?


Found this Click (http://www.aslab.com/products/workstations/customize/marquisk825.cgi) , select the CPU drop down list, you'll see you have to pay an additional $1106 for the "265", $1366 for "270" and $1689 for "275" 2MB L2 cache dual core processor.

so far, Dual Processor prices = :woot:

[[XC] moddolicous]

yea, but the performance = :woot:

[Thorry]

OMG this is a very sick CPU.... 1.2 vcore.... :o

Wish I had a few of these in my server :D

[terrace215]

Hi guys! It’s interesting to see the screenshots of the AMD Athlon 64 X2 processor here. And I decided to make a small, but highly exclusive screenshot of the Dual Core AMD Opteron processor 275. AMD Corp. has permitted me to post it here, but only one screen, because my CBId tool shows too much confidential info, he-he :) OK, no problem. So, here it is:

It’s a low-power 68W AMD Opteron processor based on the newest E6 core!

AMD allowed you to post this? I find that extremely hard to believe.

Also, the DC Opteron 275 has a 95W TDP, not 68W.

1.2v seems low, based on the X2 running at 1.3v... you sure this isn't a fake? :)

Another problem: the OPN is bogus, or 1.2v is bogus.

"FAA" encodes current Rev E4 Opteron package, voltage, and case temp.

So that portion of the OPN says the voltage is: 1.35/1.4v, not 1.2.

Busted.

[Lostfaith]

x.i.p = very elite important member, like sxs112.

they are definately not bullsh*tting us, so don't think lightly of them terrance215

[agenda2005]

AMD allowed you to post this? I find that extremely hard to believe.

Also, the DC Opteron 275 has a 95W TDP, not 68W.

1.2v seems low, based on the X2 running at 1.3v... you sure this isn't a fake? :)

Another problem: the OPN is bogus, or 1.2v is bogus.

"FAA" encodes current Rev E4 Opteron package, voltage, and case temp.

So that portion of the OPN says the voltage is: 1.35/1.4v, not 1.2.

Busted.

Cool down man. This is legit. The Vcore might be incorrect, but the screen shot is 100% legit. It's simple. The guy is X.I.P.

[VVJ]

Not long ago I have signed NDA and I very proud of this. Therefore I’m afraid of to lost it, because it cost a lot.


Also, the DC Opteron 275 has a 95W TDP, not 68W.
Soon you’ll understand you are mistaken :)


1.2v seems low, based on the X2 running at 1.3v... you sure this isn't a fake?
No it is not! VCore is 1.200V. Today, I’ll try to post a little more info regarding this CPU.


Another problem: the OPN is bogus, or 1.2v is bogus.
"FAA" encodes current Rev E4 Opteron package, voltage, and case temp.
Do you know something about OSP suffix? I’m not sure you do :) Besides, FAA encodes not core revision, the last two letters encode the revision. The A-voltage defines a variable voltage, not definite as you suppose. For example, it may be 1.200V or 1.250V. The letter F says that CPU uses a new organic package.


Busted.
:)

[aoc007]

x.i.p = very elite important member, like sxs112.

they are definately not bullsh*tting us, so don't think lightly of them terrance215

I think its xtremely important person ;)

Does anybody know if this will be making an appearance in socket939? Id love to see one in a SLI-DR.

[STEvil]

pair of these dropped into a "cheap" dual 940 board would be very attractive....

[metro.cl]

I think its xtremely important person ;)

Does anybody know if this will be making an appearance in socket939? Id love to see one in a SLI-DR.


its socket 940, one more pin.


great info, hope you can show some benchs soon

[terrace215]

Not long ago I have signed NDA and I very proud of this. Therefore I’m afraid of to lost it, because it cost a lot.


Soon you’ll understand you are mistaken :)


No it is not! VCore is 1.200V. Today, I’ll try to post a little more info regarding this CPU.


Do you know something about OSP suffix? I’m not you do :) Besides, FAA encodes not core revision, the last two letters encode the revision. The A-voltage defines a variable voltage, not definite as you suppose. For example, it may be 1.200V or 1.250V. The letter F says that CPU uses a new organic package.


:)


I would be very happy to be wrong, trust me.

But if 68W is possible with the 275, why aren't they releasing the 280 right now, too?

Also, I didn't claim that FAA encoded the core revision.

OSP is the prefix for DC Opteron. I can accept that, as OSA is the prefix for SC Opteron.

Next we have FAA.

F = organic package, like the Rev E4 Opterons.
A = voltage
A = die case max temp of 71 C

Then 6CB.

6 = 2MB total L2
"CB" is the core revision. I could believe that stands for E6 dual core.

My trouble is with the "A" voltage line. For single core E4 parts, that signifies either 1.35v or 1.4v.

So you are telling me that the A really only means "variable"... Hmmm.

I suppose that is possible.

It seems odd, though, that they would not want to be more specific, doesn't it?

....

Also, were you joking when you said AMD gave you permission to post information about the 275 before the NDA expires on Thursday? Because why would AMD do that?

.....

Regarding 68W... I suppose if AMD can produce 35W 2.2GHz Turions, then a 68W 2.2GHz DC Opteron 275 is not out of the question.

[NotoriousMike]

Regarding 68W... I suppose if AMD can produce 35W 2.2GHz Turions, then a 68W 2.2GHz DC Opteron 275 is not out of the question.

Yep, which acutally is sort of wierd, i think AMD should have went with the slapped together core design like intel. Other then the HT links being connected, there is not much diffrence. Plus you can bin the chips MUCH more effiecently.

[cupra]

that is look so nice :slobber:

[VVJ]

terrace215:

But if 68W is possible with the 275, why aren't they releasing the 280 right now, too?
Don’t forget about low-power AMD Opteron HE and EE processors having an extremely low TDP of 55W and 30W respectively.


OSP is the prefix for DC Opteron. I can accept that, as OSA is the prefix for SC Opteron.
As far as I know, an OPN number of the newest Opteron processors doesn’t contain a letter that points to the dual core presence. Perhaps, new model numbers x65, x70, x75 point to this.


My trouble is with the "A" voltage line. For single core E4 parts, that signifies either 1.35v or 1.4v. So you are telling me that the A really only means "variable"... Hmmm.
;)


Also, were you joking when you said AMD gave you permission to post information about the 275 before the NDA expires on Thursday? Because why would AMD do that?
A man who works in the AMD Connectivity Solutions Group ask me to make a very sexual surprise for Intel :D By the way, he is a registered member on this forum :)

[b0bd0le]

ooooooh GIMME!!! I WANT!

any word on prices?

are they really going to be $1600?

[Hallowed]

Awesome chip, but that price is a bit extreme.

Thanks for the sneak peek VVJ. Much appreciated. :D

[lv_88]

the prices looks way to high right now, hopefully they will drop a bit :slobber:

[dpa]

terrace215:
A man who works in the AMD Connectivity Solutions Group ask me to make a very sexual surprise for Intel :D By the way, he is a registered member on this forum :)

that has to be Pedro.

[aoc007]

its socket 940, one more pin.


great info, hope you can show some benchs soon

I know, I was asking if anybody knows if there might be a socket939 variant....

[Clew]

It's impressive but it isn't the first screenshot. HKEPC has a little preview here:
http://www.hkepc.com/hwdb/dualcore-opteron-3.htm

They also show a pair of these in a dual config, how about this?
http://www.hkepc.com/hwdb/dualcore-opteron/4coretest.jpg

These are Opteron 866 on Egypt core.

[cpulloverclock]

Egypt, Italy, lol AMD is funny

[aoc007]

And Turion64 over IDF or whatever it was haha, they have a good sense of humor.

[sojouner]

wwwwwaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaaaaaaaaa cries the dual core p4 @ 2 x2.8 ghz and running @ 244 watts . :slapass: :slapass: :shocked: :moon:

[Korpse]

11x200
htt link x5

HYPER THREADING - i thought only intel done this?

[VVJ]

Guys, I’ll try to post a new screenshot of the P-States tab this evening if I’ll be provided, of course. It’s seems the processor fully supports Cool’n‘Quiet technology and even has some performance states.

[perkam]

But what about Opteron 175...just as there was the 150/152 opterons ???

Perkam

[[XC] leviathan18]

i saw in a post while ago that amd enable some instructions like the ones HT has just to take advantages in programs that uses HT right now so this chips will gain perfomance is that programs.

the cores have city names and country names...

[Mr. Tinker]

terrace215:
A man who works in the AMD Connectivity Solutions Group ask me to make a very sexual surprise for Intel :D By the way, he is a registered member on this forum :)
THAT is hilarious. Take that Intel. Sexual indeed - most of us here would agree. And that HT bit include is brilliant. *goes to change avatar*

[Ailleur]

i saw in a post while ago that amd enable some instructions like the ones HT has just to take advantages in programs that uses HT right now so this chips will gain perfomance is that programs.

the cores have city names and country names...


Righty-o

http://www.x86-secret.com/?option=newsd&nid=870

[terrace215]

It's impressive but it isn't the first screenshot. HKEPC has a little preview here:
http://www.hkepc.com/hwdb/dualcore-opteron-3.htm

They also show a pair of these in a dual config, how about this?

These are Opteron 866 on Egypt core.


Yeah, and for those, DC Opteron 875 (866 was the old name), voltage is 1.35v. OTOH, that chip was made late last year, so it is possible that DSL was not incorporated yet, hence a newer 275 could have the lower voltage of 1.2v shown above.

[angrysquirrel]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam

Intel and AMD have a technology sharing agreement, like how the Pentium 4 6xx series uses AMD's x64 instruction set and how the a64 uses SSE and SSE2, and how the new a64 x2 uses SSE3.

http://www.dvhardware.net/article2053.html

Edit: Looks like someone beat me to it, I should have read the whole tread...commence with the :slapass:

[VVJ]

Hi guys! As I promised before, here is another portion of sweet screenshots!

As you can see, the processor is in the minimum power state:

http://cbid.amdclub.ru/work/CBId01.gif

and more interesting screenshot reports its three performance states…

http://cbid.amdclub.ru/work/CBId02.gif

The maximum TCase spec is 71 Celsius degree.

[goodcooper]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam


maybe this isn't "HT" per se, but pacifica? i thought pacifica wasn't for a while?
maybe its a similar technology that just has the bit enabled, so that brain reads it as HT... there are many ways for AMD to get around the litigation

(i love the 'sexual suprise' for intel... )

i'm taking that as meaning this suprise for intel is the 1.2v and the HT: Yes portions

[terrace215]

Hi guys! As I promised before, here is another portion of sweet screenshots!

As you can see, the processor is in the minimum power state:

and more interesting screenshot reports its three performance states…

The maximum TCase spec is 71 Celsius degree.

If you are allowed, can you tell us if this DC Opteron 275 is a special "low power" 275 that AMD will be selling (like Opteron HE or Opteron EE, although I wasn't expecting that at 2.2GHz), or will ALL DC Opteron 275's run at 1.2v and have a 68W TDP?

thanks.

[Quanticles]

Intel HT = HyperThreading, A method of splittings pipelines into seperate proccesses

AMD HT = HyperTransport, A physical communication link like PCI, SATA, etc, that connects processors together and processors to the chipset. Your 939 has an HT link between the processor and the chipset.

There will be a dual-core 939, that information is posted on AMD's website.

AMD uses a different type of manufacturing process that creates very low power, very low voltage components. I believe they develope it with IBM (partners).

AMD is reporting very high yields on their processors.

The pricing strategy is such that a dual core will be slightly cheaper than two single cores. This is to make as much money as possible (you want AMD to do that so they can afford more R&D). This strategy is being applied to the server/enterprise market (big companies). No home user is going to want to buy opterons anyway. There is no telling what the pricing strategy will be for 939.

939 should be out in a couple months, June I believe.

[cpuz]

Actually dual core Athlon/Opteron really report the Hyper-Threading flag as being set, and also 2 logical CPUs.
IMO this is just for software compatibility, even if the CPU does not support HT stricto senso.

[Ailleur]

Intel HT = HyperThreading, A method of splittings pipelines into seperate proccesses

AMD HT = HyperTransport, A physical communication link like PCI, SATA, etc, that connects processors together and processors to the chipset. Your 939 has an HT link between the processor and the chipset.

There will be a dual-core 939, that information is posted on AMD's website.

AMD uses a different type of manufacturing process that creates very low power, very low voltage components. I believe they develope it with IBM (partners).

AMD is reporting very high yields on their processors.

The pricing strategy is such that a dual core will be slightly cheaper than two single cores. This is to make as much money as possible (you want AMD to do that so they can afford more R&D). This strategy is being applied to the server/enterprise market (big companies). No home user is going to want to buy opterons anyway. There is no telling what the pricing strategy will be for 939.

939 should be out in a couple months, June I believe.

The discussion wasnt about the meaning of HT across both company but was about amd enabling the hyperthreading bit on their cpu

Read http://www.x86-secret.com/?option=newsd&nid=871

[drcrawn]

Intel is in so much trouble it's not even funny anymore :) I'm sold on Opteron more than ever now. And do I even need to mention that XS is fast becoming the place for first looks on just about everything. Love this place. Socket 940 power baby!
As for the voltage, I believe it. It's already been shown the Opty E4s run much cooler and use considerable less power. We are witnessing AMD come into a period of pure greatness and things just keep getting better.

[[XC] leviathan18]

Originally Posted by Quanticles
Intel HT = HyperThreading, A method of splittings pipelines into seperate proccesses

AMD HT = HyperTransport, A physical communication link like PCI, SATA, etc, that connects processors together and processors to the chipset. Your 939 has an HT link between the processor and the chipset.

There will be a dual-core 939, that information is posted on AMD's website.

AMD uses a different type of manufacturing process that creates very low power, very low voltage components. I believe they develope it with IBM (partners).

AMD is reporting very high yields on their processors.

The pricing strategy is such that a dual core will be slightly cheaper than two single cores. This is to make as much money as possible (you want AMD to do that so they can afford more R&D). This strategy is being applied to the server/enterprise market (big companies). No home user is going to want to buy opterons anyway. There is no telling what the pricing strategy will be for 939.

939 should be out in a couple months, June I believe.

you are right but in this chips amd did this:

Athlon 64 Simple Core - CPUID (EAX = 1)
Hex (EDX) : 0x078bfbff
Bin (EDX) : 0b00000111100010111111101111111111
Hex (EBX) : 0x00000800

Athlon 64 Dual Core - CPUID (EAX = 1)
Hex (EDX) : 0x178bfbff
Bin (EDX) : 0b00010111100010111111101111111111
Hex (EBX) : 0x00020800

The bit 28 in register EDX (which is emphasized in red here) corresponds to the support of HyperThreading. However, we can see clearly that this bit is now set to 1 whereas it was to 0 for a standard, Single Core, Athlon 64. In the same way, the bits [23:16] of register EBX, which indicate the number of logicals CPU supported on Intel CPUs with Hyperthreading enabled, are also set to "2" in the case of a DC Athlon 64.

So, It seems that AMD chose to activate the "HyperThreading" bit on those Athlon 64 Desktop CPUs in order to profit from optimizations already done by many programmers for HyperThreading technology. This will make possible for those upcoming Athlon 64 Dual Core to also benefit from work already carried out. For now, we do not know yet if the Dual Core Opterons will also have this bit active.

as i said before just to take advantage in the programs that have some optimizations with hyperthreading.....

[Quanticles]

I thought there was a question about Intel HT and AMD HT, I must have been mistaken.

There's no reason to not take advantage of HyperThreading optimizations using a dual-core scheme, that's what HyperThreading was supposed to represent. I'd imagine Intel is doing the same thing.

[terrace215]

Intel HT = HyperThreading, A method of splittings pipelines into seperate proccesses

AMD HT = HyperTransport, A physical communication link like PCI, SATA, etc, that connects processors together and processors to the chipset. Your 939 has an HT link between the processor and the chipset.



Yes, but in this case, the reference to HT is to hyperthreading support, which AMD's dual core parts will have. Just so that software that checks for HT will run properly, and take advantage of both cores.

This has been known for some time, announced last year by AMD, despite Samuel of x86-secret thinking he discovered something.

[terrace215]

Intel is in so much trouble it's not even funny anymore :) I'm sold on Opteron more than ever now. And do I even need to mention that XS is fast becoming the place for first looks on just about everything. Love this place. Socket 940 power baby!
As for the voltage, I believe it. It's already been shown the Opty E4s run much cooler and use considerable less power. We are witnessing AMD come into a period of pure greatness and things just keep getting better.

It may be possible with DSL / a different transistor mix.

E4 Optis are nice, but they run at 1.35v/1.4v at top clock, like Venice and San Diego.

1.2v is just amazing.

MT-xx Turions might run at 1.2v as well, I suppose.

[terrace215]

The Inquirer called this back in January, upon review.

In fact, they were even more optimistic, saying there would be a 55W HE part at 2.2GHz and 1.15v.

http://www.theinquirer.net/?article=20420

Still, 68W and 1.2v is very, very good for dual core 2.2GHz.

So I guess the answer is, VVJ is showing us the 2.2GHz HE Dual core Opteron.

[aoc007]

The Inquirer called this back in January, upon review.

In fact, they were even more optimistic, saying there would be a 55W HE part at 2.2GHz and 1.15v.

http://www.theinquirer.net/?article=20420

Still, 68W and 1.2v is very, very good for dual core 2.2GHz.

So I guess the answer is, VVJ is showing us the 2.2GHz HE Dual core Opteron.

Please AMD release it to s939 as an FX cpu (maybe with higher voltages/clocks?) crosses fingers. Although I have heard the reports that AMD has no plans for a dual core FX cpu, so you guys dont need to reiterate that ;) But really, shouldn't they have a competitor to the EE?

[karelke]

The Inquirer called this back in January, upon review.

In fact, they were even more optimistic, saying there would be a 55W HE part at 2.2GHz and 1.15v.

http://www.theinquirer.net/?article=20420

Still, 68W and 1.2v is very, very good for dual core 2.2GHz.

So I guess the answer is, VVJ is showing us the 2.2GHz HE Dual core Opteron.

That's an old roadmap.
Here's (http://www.tweakers.net/ext/f/55322/full.jpg) one from February.

[terrace215]

That's an old roadmap.
Here's (http://www.tweakers.net/ext/f/55322/full.jpg) one from February.

True. Interesting... OTOH, they are currently supplying 68W single-core Opterons at 2.4GHz to HP for blade servers, so I guess they want a drop-in replacement for that.

Perhaps there's a new category in between 55W and 95W.

[terrace215]

Please AMD release it to s939 as an FX cpu (maybe with higher voltages/clocks?) crosses fingers. Although I have heard the reports that AMD has no plans for a dual core FX cpu, so you guys dont need to reiterate that ;) But really, shouldn't they have a competitor to the EE?

Well, whatever you call it, the A64 X2 4800+ *is* a dual San Diego, i.e. a dual FX, with 2 x 1MB L2 at 2.4GHz.

That is, 2 x FX-53, i.e. 2 x 4000+.

This will blow the 840 EE out of the water.

[dippyskoodlez]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam

No, AMD could actually implement it if they wanted. they have a technology sharing agreement. thats how AMD uses SSE, SSE2, etc.

[perkam]

AMD HT = HyperTransport, A physical communication link like PCI, SATA, etc, that connects processors together and processors to the chipset. Your 939 has an HT link between the processor and the chipset.


Right...uh huh...so if HT was HyperTransporT bus, then can u pls educate us on what HTT would be :rolleyes:

Perkam

[dippyskoodlez]

Right...uh huh...so if HT was HyperTransporT bus, then can u pls educate us on what HTT would be :rolleyes:

Perkam

lol. :p:

Gotta love comments like that ;)

[some_young_guy]

THe idea is slick, but compatibility would be the last of AMD's worries. It's patented by Intel....so I smell a lawsuit...and a big one. Only that AMD knows that if Intel does pursues that route, it has its own case against Intel's monopolizing influence in the market which could get Intel into a VERY sticky situation.

Perkam

I don't know why it would be a problem since AMD owns the patent to Hyperthreading.

patent 5,944,816 (http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=12&f=G&l=50&co1=AND&d=ptxt&s1=5944816&OS=5944816&RS=5944816)
Link 2 (http://www.theinquirer.net/?article=6773)

[dippyskoodlez]

I don't know why it would be a problem since AMD owns the patent to Hyperthreading.

patent 5,944,816 (http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=/netahtml/search-bool.html&r=12&f=G&l=50&co1=AND&d=ptxt&s1=5944816&OS=5944816&RS=5944816)
Link 2 (http://www.theinquirer.net/?article=6773)

That, too.. lol :D

AMD and intel have a cross-technology license, so they can use what they want when they want prettymuch.

[Quanticles]

Alot of people call it HT, not HTT. Talk trash if you want.

[[XC] leviathan18]

chill out ppl...

[dippyskoodlez]

Alot of people call it HT, not HTT. Talk trash if you want.

Geez.. no need to get all angry :P

The CBID screenie says "hyperthreading" not HT. ;)

[Quanticles]

If I was angry I would have put :mad:

:p:

[Quanticles]

I didnt mean to kill the thread x.x

[[XC] leviathan18]

so athlon 64x2

only have 512kb of cache per core, 1mb total?

and some opteron dc have 1mb per core 2mb total?

i tought every dc would have 1mb per core 2mb total cache

we need to see if a heavy multithread program gains a lot of perfomance with 2 cores and how perfom the EE vs any athlon 64x2

this is just a nice year, the best thing we amd buyers can keep our mobos :D

[terrace215]

so athlon 64x2

only have 512kb of cache per core, 1mb total?

and some opteron dc have 1mb per core 2mb total?

i tought every dc would have 1mb per core 2mb total cache

we need to see if a heavy multithread program gains a lot of perfomance with 2 cores and how perfom the EE vs any athlon 64x2

this is just a nice year, the best thing we amd buyers can keep our mobos :D

No, some A64 X2 are 1MB x 2, some are 512K x 2.

[aoc007]

Well, whatever you call it, the A64 X2 4800+ *is* a dual San Diego, i.e. a dual FX, with 2 x 1MB L2 at 2.4GHz.

That is, 2 x FX-53, i.e. 2 x 4000+.

This will blow the 840 EE out of the water.

Ah, I didn't know that, havn't been keeping up on the tech news lately like I used to. San Diego is FX-57 right? Or is it the 90nm FX-55?

[Hector Ruiz]

hehe...Amd 65u quad :slapass:

[terrace215]

Ah, I didn't know that, havn't been keeping up on the tech news lately like I used to. San Diego is FX-57 right? Or is it the 90nm FX-55?

San Diego is the name of the desktop Rev E. Socket 939 1MB L2 parts.

So, the new 3700+, 4000+, FX-55, (and when it launches in Q3, the FX-57) are all based on "San Diego".

(Its sibling is "Venice", Rev E. Socket 939 512K L2 parts, like 3000+, 3200+, 3500+, 3800+.)

[[XC] leviathan18]

mmm i see cheap dc athlon with 512kb cache per core and i think more speed and lower ones with 1mb per core just like newcastle and clawhammer in socket 754?

[terrace215]

No, more like the reverse. We seem to have data points indicating:

A64 X2 4800+ = dual core 2.4GHz, 1MB L2 x 2
A64 X2 4400+ = dual core 2.2GHz, 512K L2 x 2

And my guess for the 3rd part is:

A64 X2 4600+ = dual core 2.2GHz, 1MB L2 x 2

-------------

Basically, I think they are only producing a 2 x 1MB L2 die right now, on socket 940 (for DC Opteron), and will put it on 939 for (DC desktop A64 = "A64 X2").

I think that the 4400+ part, 2.2GHz 512K L2 x 2 is actually the same large die, just with 1/2 the cache diasbled for each core.

[[XC] leviathan18]

fault cache chips?

interest in fact higher clocks with more cache...

now the question is a real perfomance gain having the extra 512kb of cache in a athlon 64? at the same clock speed?

the extra cache doesnt bring larger pipelines to the structure??? or more complex desing and more latency?

more cache doesnt affect perfomance???

dreaming mode on just imagine a dc fx with unlocked multi and 2.8ghz for each core dreaming mode off

[terrace215]

fault cache chips?

interest in fact higher clocks with more cache...

now the question is a real perfomance gain having the extra 512kb of cache in a athlon 64? at the same clock speed?

the extra cache doesnt bring larger pipelines to the structure??? or more complex desing and more latency?

more cache doesnt affect perfomance???

dreaming mode on just imagine a dc fx with unlocked multi and 2.8ghz for each core dreaming mode off


For single core:

2.2GHz 512K L2 = "3500+" ; 2.2GHz 1MB L2 = "3700+"
2.4GHz 512K L2 = "3800+" ; 2.4GHz 1MB L2 = "4000+"

So, going from 512K L2 to 1MB L2 results "on average" in a GAIN of 200 PR points.

[VVJ]

If you are allowed, can you tell us if this DC Opteron 275 is a special "low power" 275 that AMD will be selling (like Opteron HE or Opteron EE, although I wasn't expecting that at 2.2GHz), or will ALL DC Opteron 275's run at 1.2v and have a 68W TDP?
thanks.

Yes, you’re right, An OSP275 is a special 68W processor. Moreover, as far as I know, currently AMD doesn’t have an AMD Opteron 68W processor with models of 1xx and 8xx. Also, AMD will announce much hotter dual core OSA-Opteron with a TDP at least 90W.

[i found nemo]

notice the core name is italy and another core name is venice, aren't those both in the italian area??? i think amd is doin this by geographical location lol, good cpu though keep me updated

[aoc007]

San Diego is the name of the desktop Rev E. Socket 939 1MB L2 parts.

So, the new 3700+, 4000+, FX-55, (and when it launches in Q3, the FX-57) are all based on "San Diego".

(Its sibling is "Venice", Rev E. Socket 939 512K L2 parts, like 3000+, 3200+, 3500+, 3800+.)

So Venice is San Diego with half the cache and is the basis for the 512kb cache Athlon64 X2s?

[VVJ]

notice the core name is italy and another core name is venice, aren't those both in the italian area??? i think amd is doin this by geographical location lol, good cpu though keep me updated

I think it is too ridiculous to name the processor core by name of countries and towns. Spitfire, Thunderbird, Palomino, etc sound much proper.

[perry_78]

Nice information, though I believe people have spent too much time discussing the authenticy of this thread. Please, get over it, CPU-Z and CBI have never been exactly accurate regarding processors which are yet to be released.

[Dagalidis]

:slobber: :slobber: :slobber:

Let's save some money for my next upgrade.... :D

[goodcooper]

dreaming mode on just imagine a dc fx with unlocked multi and 2.8ghz for each core dreaming mode off

or just a decent overclocking workstation board w/ nforce pro and 2x opteron 252s...

troy core = venice/sandeigo like, sse3, fixed mem controller, should overclock very well, if only they had a beefy enough board to support overvolting 2x cpus...

(come on asus k8n-dl) maybe if we cried enough about it asus will enable numa AND OVERVOLTING in their next bios revision :D

then all we'll need is some registered pc5000

[MaSell]

:slobber: Great stuff :slobber:

[ZENNZZO]

Not long ago I have signed NDA and I very proud of this. Therefore I’m afraid of to lost it, because it cost a lot.


Soon you’ll understand you are mistaken :)


No it is not! VCore is 1.200V. Today, I’ll try to post a little more info regarding this CPU.


Do you know something about OSP suffix? I’m not sure you do :) Besides, FAA encodes not core revision, the last two letters encode the revision. The A-voltage defines a variable voltage, not definite as you suppose. For example, it may be 1.200V or 1.250V. The letter F says that CPU uses a new organic package.


:)

terrace215...pwnd, my man. :slapass:

[notoriousformula]

Dual Core Opteron Pricing (http://www.xtremesystems.org/forums/showthread.php?p=823608#post823608)

[IvanAndreevich]

Excellent stuff VVJ. Интересно :)

[cpulloverclock]

launched today in desktop version

4200+ = 2.2GHz 512kb
4400+ = 2.2GHz 1MB
4600+ = 2.4GHz 512kb
4800+ = 2.4GHz 1MB

Venice, San diego and Athlon X2...

available in june

[aoc007]

No word on prices?

[Rys]

http://img.hexus.net/v2/rys/dcopteron/osa875fkm6bs_s.jpg

:woot:

Big one here (http://img.hexus.net/v2/rys/dcopteron/osa875fkm6bs.jpg). The box I have for testing is 2 x Opteron 875, 4GiB, nForce4 Pro. More info when we (http://www.hexus.net) publish my article. Feels so good to be able to talk about them, now the NDA has expired :banana:

[STEvil]

:drool:

Man, you're lucky ;)

[Stumpjumper5200]

2.2GHz dual core, 2x 1MB L2 caches.......1.2V?!?!

AMD = :banana: :banana: :banana: :banana:

[IvanAndreevich]

Everyone is cheering for these chips.. They own, no question about it. The prices make them crap, though.

[nordik]

:drool:

Man, you're lucky ;)

I second that :toast:

[Woody-san]

AMD Opteron Dual Core manufacturer published pricing in Tokyo:

Socket 940::
1) Opteron 875 US$2,800 (Y292K).
2) Opteron 870 US$2,300 (Y237K).
3) Opteron 865 US$1,600 (Y167K).

4) Opteron 275 US$1,400 (Y143K).
5) Opteron 270 US$1,100 (Y116K).
6) Opteron 265 US$940 (Y94K).

Availability now. the 2xx DC will ship retail im may 2005. Samples are available to selected customers ...

:slobber:

[VVJ]

Congratulations! The Single core AMD Opteron processor with Socket 939 really exists!!! :toast:

[STEvil]

????

[aoc007]

Yeah man are you sure you didnt miss-type?

[OvaKilla]

i'll take two.

[VVJ]

aoc007:
man, I'm sure, I can type the letters with the keyboard without mistakes! :D Moreover, CBId is ready to fully identify them! It even knows about OPNs of these processors. I've just announced the second Release of CBId 7.50. See the news. BTW, there are no more confid. news here from me, I have some troubles.

[smegz]

terrace215:


A man who works in the AMD Connectivity Solutions Group ask me to make a very sexual surprise for Intel :D By the way, he is a registered member on this forum :)

This is a sham! PCSG (Personal Connectivity Solutions Group) doesn't even work with processors. CPG (Computational Products Group) does. PCSG makes things like the PIC...dummy.

[VVJ]

This is a sham! PCSG (Personal Connectivity Solutions Group) doesn't even work with processors. CPG (Computational Products Group) does. PCSG makes things like the PIC...dummy.

Oh, man, please, do be ridiculous. And you can’t convince me, because I work with PCSG for a year. PCSG provides answers to many customer questions! :) Its technical documentation includes a lot of information: technical manuals, data sheets, design guides, application notes and so on. I’ll be provided with everything I want… dummy!

[smegz]

What you say is ridiculous. I truly believe that PCSG can get you a great deal of info...on embedded stuff...not Opterons. Don't believe me...check it out for yourself. Anyone with the most minimal web skills can look up this info.

PCSG
http://www.amd.com/us-en/ConnectivitySolutions/ProductInformation/0,,50_2330,00.html

CPG
http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118,00.html

Flash (Spansion)
http://www.amd.com/us-en/FlashMemory/ProductInformation/0,,37_1447,00.html

Don't be mad because your information is wrong...obviously your informant doesn't know what he's talking about. If he works for PCSG, he would not have acccess to this kind of stuff. He might be able to show off the embedded processor that is in the PIC (their flagship) but not a pre-release Opteron...give me a break.

[cha0t1c1]

these shots are siiick, really I feel like throwing up...man I can render without sleeping between each frame....I love AMD, definetly going to exchange those 244's of mine for a couple of 270's, I'll be happy tofork out $3000+, for such insanity, just for bragging rights :D

[Pivo504]

sexxyy cpu

[Saruji]

why are people arguing like little school girls when there is manly overclocking to do?

[cha0t1c1]

any successful overclocking using a dual motherboard? :slobber: :shock2:

[smegz]

why are people arguing like little school girls when there is manly overclocking to do?

Simple...he's obviously lying about his source, he is the programmer, he CAN fake results. I'm not convinced that this is genuine. It shows up on no roadmaps...it's a fake to promote his program. There are too many unanswered questions...notice how he clammed up when pushed.

[iboomalot]

Everyone is cheering for these chips.. They own, no question about it. The prices make them crap, though.

some of you people ^^^^^^^^^^^ are cheap skates go buy a skt754 2800+ or a Barton and stop b!tching please.

$550 for a Dual CPU 3500+ speeds with 1meg of L2 that should be clockable to 2.4 - 2.6 gig which would be like having TWO FX-53s or FX-55s that costs $500 / $810 for ONE CPU much less two CPUs.

I can't wait to get mine, all the gaming speed of my current 3500+ with a 50% faster encoding and my tradestation software is designed for dual CPUs so I should see a 80%+ increase in speed on my # crunching.

9-10 months ago the new 3500+ NC cost $500 now you get two plus twice the cache for the same price.

think its pricey, thats your personal opinion ,but you have cheaper options that you can use, instead of dumbing down the AMD cpu to an Intel level. :stick:

[smegz]

I'm not convinced that this is genuine. It shows up on no roadmaps...it's a fake to promote his program. There are too many unanswered questions...notice how he clammed up when pushed.

I think I need to qualify this statement....help if I read my post before submitting. The Paris core and Opteron 200 series are on the roadmap. What is not is a 68W part. See the images at AMDBoard.com for details.

AMDBoard Images (http://www.amdboard.com/amdroadmap.html)

Opterons are shown in 2 flavors...95W and 55W (last image on page), not 68W. This combined with the fact that VVJ is supposedly getting his data from someone in PCSG, a group not known to deal with processors, leads me to believe that something is not Kosher.

[smegz]

The Paris core and Opteron 200 series are on the roadmap.

See what I mean...Italy...not Paris :clap:

[VVJ]

Man, don’t dig further! You deal with the Russian mafia! :worm: You need some additional info on dual-core Opteron? I can provide it here.

[smegz]

Man, don’t dig further! You deal with the Russian mafia! :worm: You need some additional info on dual-core Opteron? I can provide it here.
Russian Mafia? You? :lol2: Don't make me laugh man. So go ahead...enlighten me. What pearls of wisdom do you have regarding DC Opterons? What rabbit can you pull out of your questionable hat? Nothing...that's what. Now had thse screens come from an independent source and not from the programmer, they may carry more weight. It would have been more believable if you had said that the part was 55W but to pull 68 out of your hat...

[TSFroggy]

Russian Mafia? You? :lol2: Don't make me laugh man. So go ahead...enlighten me. What pearls of wisdom do you have regarding DC Opterons? What rabbit can you pull out of your questionable hat? Nothing...that's what. Now had thse screens come from an independent source and not from the programmer, they may carry more weight. It would have been more believable if you had said that the part was 55W but to pull 68 out of your hat...

:rolleyes:

[knob]

:P

10 char