AMD Phenom II X6 | Turbo CORE 101

[FlanK3r]

only some days to launch jupiii

[Oliverda]

These are being sold in china still no review??

I mean no forum info, etc?

http://forum.coolaler.com/showthread...=235891&page=3

[Manicdan]

http://forum.coolaler.com/showthread...=235891&page=3

cool thanks for the link

its weird how cpuz calls the bottom cpu, the one they OCed a 1095T

yet i love how its 1.40v for 4.2ghz, crazy compared to my PII 940 which is 1.35v stock for 3ghz.

[tbone8ty]

X4 1090T @ 1.4v 4200mhz

[ajaidev]

http://forum.coolaler.com/showthread...=235891&page=3

Thanks just what the doc ordered

here is a nice pic of a girl i found in the link you gave me....



Enjoy

X4 1090T @ 1.4v 4200mhz

He disabled 2 cores because the score in CPU Queen did not increase with the increase in freq.

http://item.slide.com/r/1/274/i/MGYX...Ko9hEfpU58bHp/

http://item.slide.com/r/1/263/i/MKIv...MpSS9yUKo_w25/

[PatRaceTin]

Damn low voltage i want it

[Nintendork]

Gotta get a sexy.

/me smash piggy bank

[64NOMIS]

you forgot R which stands for Rapid

Turbo Core Opportunistic Rapid Execution...lol

Can.Overclock.Real.Easy

[JumpingJack]

Looking over the coolaler info, this is gonna be the first CPU from AMD in some years that may cause Intel to rethink their pricing on the high end....

[OhNoes!]

looking over the coolaler info, this is gonna be the first cpu from amd in some years that may cause intel to rethink their pricing on the high end....

qft.

[qurious63ss]

Looking over the coolaler info, this is gonna be the first CPU from AMD in some years that may cause Intel to rethink their pricing on the high end....

Pricing no. You may see another higher bin released sooner then expected though. Having the IPC and process lead is a great advantage which Intel will use when it comes to maintaining market leadership.

[JumpingJack]

Pricing no. You may see another higher bin released sooner then expected though. Having the IPC and process lead is a great advantage which Intel will use when it comes to maintaining market leadership.

I am not so sure -- it is unlikely to touch the i7-980X, that is certain; however, that is also Intel's EE or $1K CPU -- pretty exorbitant.

But it should be a good match against Lynnfield, and give the i7-930 a run for it's money.

I am of the impression this will be the next Q6600/i7-920 ... it is surely going to be a hit.

Jack

[Voodoo²]

Something is no clear for me. Many people says that turbo core kicks in when 3 cores are idle. I think this has to be clarified. It is not the same 3 idle cores than 3 active cores or threads. Because if that is the case and the 960T being just a thuban with 2 disabled cores, What will happen to turbo core in this model? Just one thread/core will be turbo core enabled?

[JumpingJack]

Something is no clear for me. Many people says that turbo core kicks in when 3 cores are idle. I think this has to be clarified. It is not the same 3 idle cores than 3 active cores or threads. Because if that is the case and the 960T being just a thuban with 2 disabled cores, What will happen to turbo core in this model? Just one thread/core will be turbo core enabled?

Hard to say exactly the implementation. But as I read it, and many here will likely argue, when the OS sees a load to put the chip into a lower power state, if there are 3 or less threads running then Turbo will certainly kick in, with any combination of the 3 cores.

At a glance, AMD appears to have implemented a turbo feature similar to what Intel did with Penryn in the mobile space by redefining the P-states, and hard coding in the microcode rules for accessing the higher P state on active core transparent to the OS.

I am anxious to see this in action myself and work with the CPU to figure out how they did it.

[villa1n]

I am not so sure -- it is unlikely to touch the i7-980X, that is certain; however, that is also Intel's EE or $1K CPU -- pretty exorbitant.

But it should be a good match against Lynnfield, and give the i7-930 a run for it's money.

I am of the impression this will be the next Q6600/i7-920 ... it is surely going to be a hit.

Jack

Indeed.

Now we just have to see what the bios support on all the boards is going to be like, if it will be all, or just select ones from the big mobo companies.

[JumpingJack]

Indeed.

Now we just have to see what the bios support on all the boards is going to be like, if it will be all, or just select ones from the big mobo companies.

I am 100% certain they will all support it, at least all the AM3 boards. (EDIT: http://www.hitechlegion.com/music/amd_simon.mp3 , AM2+ is certainly supported)

I am not certain though in one AMD slide that states the NB interlocks the power draw. That has me confused.

[Oliverda]

1035T (95W) HDT35TWFK6DGR
1055T (95W) HDT55TWFK6DGR
1055T (125W) HDT55TFBK6DGR
1075T (125W) HDT75TFBK6DGR
1090T (125W) HDT90ZFBK6DGR


1. H = Phenom
2. D = Desktop CPU
3. Txx = Thousand + xx
4. T = Turbo
4. Z = Black Edition (and Turbo)
5. WF = 95W TPD
5. FB = 125W TPD
6. K = AM3 CPU
7. 6 = Number of cores
8. D = 6 MB L3 and 512 kB L2 per core
9. GR = Revision E0

According this the 960's OPN could be:

HD960TWFK4DGR

or

HD960ZWFK4DGR

if it will be Black Edition.

...but I still haven't find anything about the 960.

[FlanK3r]

cool! winrar up 3000kb/s at default, Cinebench 32-bit 14 000 etc...14 days waiting for me ...

[Dresdenboy]

Something is no clear for me. Many people says that turbo core kicks in when 3 cores are idle. I think this has to be clarified. It is not the same 3 idle cores than 3 active cores or threads. Because if that is the case and the 960T being just a thuban with 2 disabled cores, What will happen to turbo core in this model? Just one thread/core will be turbo core enabled?

I made a calculation to find out, how this might work. See
http://citavia.blog.de/2010/04/12/tu...cking-8353947/

The problem is, that the turbo mode voltage will be applied to all cores if turbo mode is engaged.

With Zosma 2 cores need to be idle to allow the 2 other cores to engage turbo mode. This would also stay below TDP. But that is based on voltages as known from Thuban.

[Manicdan]

cool! winrar up 3000kb/s at default, Cinebench 32-bit 14 000 etc...14 days waiting for me ...

for some giggles i got cinebench 10 and ran the 32bit mode. i have my PII 940 at 16x210 for 2 cores, and 14x210 for the other 2 cores. running DDR2 840mhz Cas4

my single cpu score was 2948, near identical, except thats at 3360mhz, while the 1090T did that at 3200mhz, lol. i guess the ram really does make a nice difference.

and my multi cpu score was 10,008. theirs being 14,381 is quite a sick number. using the same settings i did 64bit and got exactly 25% increase. i hope this means we see the 1090T break 18k at stock in 64bit

[FlanK3r]

OC 1090T at x64 get 20 000 easily . I know it, cause my 4 GHz X4 965 got 16 500+

[flippin_waffles]

So intels version of Turbo CORE can't accelerate 3 cores at time at max turbo? Only a single core IF the environmental conditions are right. I've no idea whatsoever but that seems to be the case yes? If so, is it the superior design, or SOI process(or both) that allows AMD to achieve such a good turbo?

[Manicdan]

OC 1090T at x64 get 20 000 easily . I know it, cause my 4 GHz X4 965 got 16 500+

with how good core effiency seems to be, 16500 + 45% = 24,000 then factor in it will probably OC better than yours, we can assume MANY people will be in the 25,000pts CB-10 64bit club

[ajaidev]

I did not include core 5 and 6 because it same as the 4 core for the Thuban.

[charged3800z24]

Look sgood, I am excited, I can't wait to try one of these CPU's out. I found a place that sells dry ice near me as well. I dunno, if the wife will let me buy one tho....we'll see. keep the info coming.

Simon, I saw your Q&A at the legion, but I didn't get signed up fast enough to post. Still looking for the video of the interview.