Core i7 920 rev D0 stepping 5 :)

Printable View

Show 100 post(s) from this thread on one page

03-07-2009, 04:40 PM
tomb18

Quote:

Originally Posted by jcniest5

I7 is still in high demand (especially if you sell yours for lower than the retail/new price out there). It will be SOLD almost instantly. You can get some serious cash to help you buy a D0 one.

I wish! I've had my 920 on sale for a month at $100 less than the retail rate and there have been no takers!

Let's face it...the only people buying now are overclockers...and would you want to buy one from someone selling one this soon? Wouldn't you think its a bad clocker?
03-07-2009, 04:41 PM
DerekT

No, I agree. Questioning everything is valid in determining resultant outcomes. However, taking ones personal views (without any determining proofs) and extrapolating to a negative outcome is mere conjecture and opinion and has NO place in the questioning of actual and provable outcomes.

Quote:

Originally Posted by ReverendMaynard

wow, let it go. You need to check out some Bill Hicks stand up and learn how questioning everything is valid.
03-07-2009, 05:52 PM
jcniest5

Quote:

Originally Posted by tomb18

I wish! I've had my 920 on sale for a month at $100 less than the retail rate and there have been no takers!

Let's face it...the only people buying now are overclockers...and would you want to buy one from someone selling one this soon? Wouldn't you think its a bad clocker?

How much did you have yours on sale? Don't tell me you had yours on sale for $220 because Micro Center sell theirs (New) for $229.99. Are you sure you had yours for $100 less than their retail stores? I had mine on sale for $200 and it was gone within seconds. You have to sell yours for less than the cheapest price from many online vendors.
03-08-2009, 12:32 AM
Blackbird_CaD_

For me it will be a lottery if i get D0 or C0 stepping :( since i buy from a internet store (and they build the PC for me)
03-08-2009, 01:56 AM
Pt1t

I tried too i7 920 D0 ... Max Bclk on this board is also 222Mhz like C1 cpu.
To be prime stable @ 4Ghz under intel cooler i needed 1,5V Vcore ... but MAx temp was ~65° under 8 Threads Load.

Here is also my max 3dMark 2k6 on AIR.

This CPU needs a lot of Vcore.

http://www.pt1t.eu/public/X58/920D0es/920D0_2k6.PNG
03-08-2009, 02:12 AM
PcCI2iminal

Quote:

Originally Posted by Solarfall

impressive scores so far criminal..any plans for subzero ?? would be interesting to see how far you can go :toast:

the LN2 is in my room but im a little busy :) ,just give me one more day :D

Quote:

Originally Posted by Bobbylite

I didnt say they never clock better, I said they never clock like retail. Getting my butt handed to me in this thread, I feel sorry for all of you getting your hopes up. Obviously I hope they clock better. Im stuck at the dreaded 222 bclk, I doubt d0 will reslove this though.

http://img13.imageshack.us/img13/9177/940prime.jpg

i dont think this ES is a good Core i7,all my retails runs at 3600Mhz with the same or with lower vcore.
03-08-2009, 02:20 AM
PcCI2iminal

Quote:

Originally Posted by unclewebb

Starting with Core i7, Intel writes TJMax information into each processor. Intel actually calls this information TJ Target and actual TJMax can be a few degrees different from that number, typically higher. TJMax is not an exact science so reported temperatures will never be exact.

The Core i7 C0 sensors are far better than the previous 45nm Core 2 sensors. They don't seem to have much slope error and TJMax might be off by a couple of degrees but nothing major.

If you don't think that Core Temp is reading the TJMax information correctly then you could try RealTemp.
http://www.fileden.com/files/2008/3/...alTempBeta.zip

You could also use a program like MSR Tool to read the model specific register 0x1A2 bits[23:16] where Core i7 stores the TJ Target information.

http://img6.imageshack.us/img6/4218/corei7msr.png

My Q6600 doesn't show anything here but a C0 Core i7 should show 64 in these two digits because 64 hex = 100 decimal which is TJMax for all the early Core i7 that I have seen.
If 5C is in that location then Core Temp is right; TJ Target is 0x5C = 92.

I don't think that Intel would deliberately low ball TJ Target so software automatically reports lower temperatures.

PcCI2iminal: If you have time can you try running one of the above programs to clear this up. Thanks for sharing your new toy with us. :up:

12?

room temp @ 23c

http://img3.imageshack.us/img3/2443/...mpcoretemp.jpg
03-08-2009, 02:27 AM
PcCI2iminal

Quote:

Originally Posted by Pt1t

I tried too i7 920 D0 ... Max Bclk on this board is also 222Mhz like C1 cpu.

920 D0 + Tpower "X58A"
nice Thomas :)

Quote:

To be prime stable @ 4Ghz under intel cooler i needed 1,5V Vcore ... but MAx temp was ~65° under 8 Threads Load.

65° with original heatsink?
the temps must be wrong like mine

Quote:

Here is also my max 3dMark 2k6 on AIR.

This CPU needs a lot of Vcore.

go cascade ,go sub zero :D
03-08-2009, 03:58 AM
Pt1t

Quote:

Originally Posted by PcCI2iminal

920 D0 + Tpower "X58A"
nice Thomas :)

I flashed the TpowerX58 with wrong Bios :p:
03-08-2009, 04:58 AM
Infa

Im not flaming anyone or so, but why are all using outdated temp progs? :) 0.99.5 latest coretemp, showinig right temp?
and Realtemp 3,07 ? It's up to you, but shows the right tempk instead of wrong?

and i allways like teasers :) keep it up!
03-08-2009, 05:28 AM
Pt1t

Max CPUZ validation on this motherboard.

Vcore 1,60V
Evaporator temp -10°
ColdBug -65°

http://valid.canardpc.com/show_oc.php?id=522857
http://valid.canardpc.com/cache/screenshot/522871.png
03-08-2009, 05:37 AM
eva2000

not bad sounds like more bclk limited than cpu raw mhz limited ?
03-08-2009, 05:45 AM
Pt1t

Quote:

Originally Posted by eva2000

not bad sounds like more bclk limited than cpu raw mhz limited ?

Tpower X58 got bclk limited to 222.

I will try on p6t6. And for frequency my 920 D0 is really bad and it want too much Vcore.

Quote:

Originally Posted by Infa

Im not flaming anyone or so, but why are all using outdated temp progs? :) 0.99.5 latest coretemp, showinig right temp?
and Realtemp 3,07 ? It's up to you, but shows the right tempk instead of wrong?

and i allways like teasers :) keep it up!

The 99.5 Gives exactly the same temperature.
03-08-2009, 05:49 AM
msgclb

1 Attachment(s)

Quote:

Originally Posted by PcCI2iminal

12?

room temp @ 23c

http://img3.imageshack.us/img3/2443/...mpcoretemp.jpg

You have to enter 0x1A2 in the MSR Number box and click the Read MSR button.

Attachment 96057
03-08-2009, 06:04 AM
Pt1t

I compared C1 (3841Axxx) and D0 (3842Bxxx) on same board at 4GHZ, 200*20, VTT 1,4V.

Under Intel stock cooler the C1 takes prime95 stable with 1,32V and the D0 needed 1,50V. Temperatures under load are almost the same.

:)
03-08-2009, 06:47 AM
PcCI2iminal

Quote:

Originally Posted by msgclb

You have to enter 0x1A2 in the MSR Number box and click the Read MSR button.

Attachment 96057

thanks :)

http://img183.imageshack.us/img183/1...mprealtemp.jpg

Quote:

Originally Posted by Pt1t

Max CPUZ validation on this motherboard.

Vcore 1,60V
Evaporator temp -10°
ColdBug -65°

http://valid.canardpc.com/show_oc.php?id=522857
http://valid.canardpc.com/cache/screenshot/522871.png

nice Thomas :D but cold bug @ -65c :down:
anyway, is good to know about that low cold bug cz tomorrow i will try some LN2 :D

Quote:

Originally Posted by eva2000

not bad sounds like more bclk limited than cpu raw mhz limited ?

we have the same bclock wall but these two samples are not balanced, mine needs less vcore to work at the same clock as Thomas.

Tpower X58 has a option called QPI Link Speed and i can choose full speed(default) or slow mode to reduce the QPI link, I think this option is for reducing the multiplier and when the slow mode is activated the QPI link will work with a lower frequency than default but with no real gain on bl:banana::banana::banana::banana: ....the max is 222Mhz .

http://img365.imageshack.us/img365/6...lockpi32mj.jpg
03-08-2009, 07:01 AM
The Watcher

so does that mean the D0 not much different from the old C1 then.
03-08-2009, 07:02 AM
eva2000

Quote:

Originally Posted by The Watcher

so does that mean the D0 not much different from the old C1 then.

clock wise i think there's the same variation from what i've seen i.e. this thread

but temps definitely are cooler on D0
03-08-2009, 08:09 AM
Rammsteiner

Quote:

Originally Posted by eva2000

clock wise i think there's the same variation from what i've seen i.e. this thread

but temps definitely are cooler on D0

To be honest, temps seem to be the only good thing. What's the point of low temps but 1.5Vcore for a 24/7 OC. If you would pay attention to Intels Volts table you're basicly not lmited by your board and CPU, but by Vcore:rolleyes:

Of course there are different D0 CPU's, just like C0/C1, but thus far I cant really say I missed a lot from D0 compared to my chip, which seems to have made its own hole in the ozon layer to clock even higher. It's seriously unstoppable:cool:
03-08-2009, 09:11 AM
Zucker2k

Ramm, good to see your having good results on the *blue* sid; :p: impressive clocks :up:
03-08-2009, 09:32 AM
Rammsteiner

Meh, it ain't stable yet. I get a BSOD after a little while while running Intelbrun Test and Prime95 Small FFT. Now playing with different Voltages.

Guess it's a minor thingy since it ran above mentioned combo for 1.5h or so, and it does put quite a strain on the system. Maybe it's QPI:rolleyes:

EDIT Actually considering to set Blck back to 200Mhz and just live with a 4.2Ghz 24/7 OC.
03-08-2009, 11:54 AM
unclewebb

PcCI2iminal: Thanks for reading MSR 0x1A2 and posting that. 0x5C = 92 so Intel did set TJ Target to 92 on your CPU and Core Temp and RealTemp are both reading that information correctly. Pt1t has a 920 ES and CoreTemp shows that his TJ Target is set to 93.

The question now is did Intel lower TJMax or did they only lower the TJ Target value? It's possible that the actual TJMax is still 100C.

I hate to be cynical but if all Intel did was lower the TJ Target value in this register then all software will report your core temperature lower than it actually is.
We'll have to wait until the retail CPUs come out to find out what's really going on.

The best way to test for this is to lower your vcore and MHz as low as they can go and compare your reported core temperature to your room temperature near your open case.
rge did some testing here.
http://www.xtremesystems.org/forums/...postcount=2429

If your reported core temperature during this test is the same or lower than your air temperature when air cooled then I'd be suspicious about these new TJ Target values.
03-08-2009, 12:29 PM
roller11

1 Attachment(s)

When/where will these D0 chips go retail... not that I would replace my current C0 with a D0:)
03-08-2009, 01:55 PM
Splave

you using a slow mode or something roller? Why is QPI link so low
03-08-2009, 02:47 PM
PcCI2iminal

Quote:

Originally Posted by unclewebb

PcCI2iminal: Thanks for reading MSR 0x1A2 and posting that. 0x5C = 92 so Intel did set TJ Target to 92 on your CPU and Core Temp and RealTemp are both reading that information correctly. Pt1t has a 920 ES and CoreTemp shows that his TJ Target is set to 93.

The question now is did Intel lower TJMax or did they only lower the TJ Target value? It's possible that the actual TJMax is still 100C.

I hate to be cynical but if all Intel did was lower the TJ Target value in this register then all software will report your core temperature lower than it actually is.
We'll have to wait until the retail CPUs come out to find out what's really going on.

The best way to test for this is to lower your vcore and MHz as low as they can go and compare your reported core temperature to your room temperature near your open case.
rge did some testing here.
http://www.xtremesystems.org/forums/...postcount=2429

If your reported core temperature during this test is the same or lower than your air temperature when air cooled then I'd be suspicious about these new TJ Target values.

Ive checked the temps after the first boot at default clock and vcore and the lowest showed by coretemp was near to my room temp,if i remember well under that condition the delta is 1c or 0c
anyway after finishing the LN2 run i will do that agai for you ;)

Show 100 post(s) from this thread on one page