AMD Shanghai/Deneb Review Thread

[informal]

source: http://www.xtremesystems.org/forums/...167502&page=18

~6% increase in Vcore and 20% increase in clocks

I think we will see no benchmarks from sxs112 due to NDA,but it would be nice to hear some general experience he has with the chip .

[leoy]

http://oktabit.foracamp.gr/content/f...enom-ii-x4-940
It is in greek....

[Hornet331]

so many cpu-z screens, but no numbers.

[gallag]

http://oktabit.foracamp.gr/content/f...enom-ii-x4-940
It is in greek....

Probably a very early sample, Voltage is too high.

[dread77]

OKTABIT POWER!!! I smell leaking screenshots with excellent overclocking results coming up...

[dread77]

Probably a very early sample, Voltage is too high.

I can assure you that is not an early sample and it overclocks GREAT with air...

[PetNorth]

Probably a very early sample, Voltage is too high.

Not really. Thats stock voltage (it is the same than Shanghai, 1.35v http://products.amd.com/en-us/Optero...f8=&f9=&f10=4& )

[informal]

Probably a very early sample, Voltage is too high.

Nope,default Vcore for Shanghai is the same 1.35V.And you will be amazed how "much" power it draws at default,both idle and load . The "higher" than usual voltage seems to be pretty normal for AMD's version of 45nm(SOI) process.And that's why you will see 1.6V for air cooling being pretty normal for 3.8-4Ghz OCs.And temps will be quite surprising .
BTW,the retail C2 stepping is in the CPuz shot.

@ dread77

I hope you can post more . It's getting hot in here .
Oktabit ftw ^^

[PetNorth]

I can assure you that is not an early sample and it's and it overclocks GREAT with air...

waiting for results

[Calmatory]

Nope,default Vcore for Shanghai is the same 1.35V.And you will be amazed how "much" power it draws at default,both idle and load . The "higher" than usual voltage seems to be pretty normal for AMD's version of 45nm(SOI) process.And that's why you will see 1.6V for air cooling being pretty normal for 3.8-4Ghz OCs.And temps will be quite surprising .
BTW,the retail C2 stepping is in the CPuz shot.

@ dread77

I hope you can post more . It's getting hot in here .
Oktabit ftw ^^

Can't be cool with 125W TDP.

[informal]

Can't be cool with 125W TDP.

Oh rly ? I guess it depends what is cool for you?
125W is TDP,a requirement for cooling device,it doesn't mean it draws that much power under full load.Look at 140W 9950's,they draw less then 125W 9850s,around ~100W(and those 9850s are rated the same way..). Keep in mind this is a major improvement process wise versus 65nm,current leakage is cut down considerably and parts cope great with excess vcore.Temps are great too,on air.
You'll have to wait for the first results i guess.

[leoy]

My 9950(140w) idled at 1.5vcore at 25 degrees and had a max of 45-50's with true dual fan. Surprised...., and the temps should be real cause the heatsink was not that hot, my new 9950 (125) w idles at 35 and maxes at 60 with only 1.4v,With same setup... So..., nothing is impossible.

[gallag]

Nope,default Vcore for Shanghai is the same 1.35V.And you will be amazed how "much" power it draws at default,both idle and load . The "higher" than usual voltage seems to be pretty normal for AMD's version of 45nm(SOI) process.And that's why you will see 1.6V for air cooling being pretty normal for 3.8-4Ghz OCs.And temps will be quite surprising .
BTW,the retail C2 stepping is in the CPuz shot.

Interesting, If that pans out to be true then colour me impressed. Thx for the info.

[Calmatory]

Oh rly ? I guess it depends what is cool for you?
125W is TDP,a requirement for cooling device,it doesn't mean it draws that much power under full load.Look at 140W 9950's,they draw less then 125W 9850s,around ~100W(and those 9850s are rated the same way..). Keep in mind this is a major improvement process wise versus 65nm,current leakage is cut down considerably and parts cope great with excess vcore.Temps are great too,on air.
You'll have to wait for the first results i guess.

My only concern is not the temperature of the core, it is usually just up to the cooler, but also the power it draws and how much it heats the rest of the case.

Besides, I wonder why 125W TDP if the thing goes well below 110-115 W etc. Sure the lower temps with the stock cooler are better, but then again I'd rather have $10 cheaper product with 10 C higher temperature. Also lower TDP usually means lower power draw which means less excess heat and lower power bills, and it "saves climate"!

I do understand that the 45 nm with immersion lithography is a godsend for AMD when compared to their 65 nm process, and that leakage has been reduced greatly, but that does not help to convince me that the high TDP wouldn't mean high power draw, which means a lot of heat which means hot chip.

[dread77]

Nope,default Vcore for Shanghai is the same 1.35V.And you will be amazed how "much" power it draws at default,both idle and load . The "higher" than usual voltage seems to be pretty normal for AMD's version of 45nm(SOI) process.And that's why you will see 1.6V for air cooling being pretty normal for 3.8-4Ghz OCs.And temps will be quite surprising .
BTW,the retail C2 stepping is in the CPuz shot.

@ dread77

I hope you can post more . It's getting hot in here .
Oktabit ftw ^^

In mentioned OKTABIT screenshot the VCore is set in auto, and the motherboard is an ASUS M3A79-T. I can't say but most probably it can be stable with less VCore.
Most probably OKTABIT will be able to upload a couple of results just as an indication of what PhenomII is capable.

[[XC] Lead Head]

My only concern is not the temperature of the core, it is usually just up to the cooler, but also the power it draws and how much it heats the rest of the case.

Besides, I wonder why 125W TDP if the thing goes well below 110-115 W etc. Sure the lower temps with the stock cooler are better, but then again I'd rather have $10 cheaper product with 10 C higher temperature. Also lower TDP usually means lower power draw which means less excess heat and lower power bills, and it "saves climate"!

I do understand that the 45 nm with immersion lithography is a godsend for AMD when compared to their 65 nm process, and that leakage has been reduced greatly, but that does not help to convince me that the high TDP wouldn't mean high power draw, which means a lot of heat which means hot chip.

TDP is not really that important. It is mainly a specification for Heatsink designers and motherboard manufactures for a worst case scenario type of thing. Like a future CPU that may have ridiculous clocks/voltages

Put it this way, the TDP for my 939 Opteron 165 is 110 Watts. My whole system only was drawing 180-200 Watts of AC power full CPU and 3D Load, and my power supply is only around 72% efficient. That is 130 to 140 Watts DC. That is with the Opteron Over-Volted to around 1.45 volts, @ 2.6GHz and an ATI x1800 GTO.

[Calmatory]

I still wonder why 125 W TDP if e.g. 95 W was sufficient.

[informal]

I still wonder why 125 W TDP if e.g. 95 W was sufficient.

You can ask them why they rated 9950s with 140W rating at first,while it really pulled ~100W in full load .That rating obviously is for reference only.BTW, the 2.8Ghz 920 part has the same rating of 125W,it must be "hot" too?
We'll have to wait and see what power draw these thing really have,another month and that will no longer be the unknown characteristic.

In mentioned OKTABIT screenshot the VCore is set in auto, and the motherboard is an ASUS M3A79-T. I can't say but most probably it can be stable with less VCore.
Most probably OKTABIT will be able to upload a couple of results just as an indication of what PhenomII is capable.

I'm looking forward to it

[foch3 -USA-]

Maybe they are leaving room for the increase in overclocking head room.

[ownage]

I found this on another forum. Where does it come from?, does somebody know?

Also look at the II after Phenom, while early samples didn't have that. (old ones, 'Phenom tm')

[Eson]

http://www.xtremesystems.org/forums/...7&postcount=36 ?

[ownage]

Oh, lol, didn't knew it came from XS
Thanks Eson.

[G0ldBr1ck]

whats with the corners of the heat spreader? are they raised?

[prostoman]

whats with the corners of the heat spreader? are they raised?

To me it looks like marks from a cooler. Agree anyone?

My QX have some marks like that, only closer to each other.

[charged3800z24]

To me it looks like marks from a cooler. Agree anyone?

My QX have some marks like that, only closer to each other.

they are going downward.. they are polished edges..not raised..