AMD Shanghai/Deneb Review Thread

[Uncle Jimbo]

At stock perhaps, but the updated Hwbox review suggests Phenom II's power usage grows at a incredible rate when overclocked, to a point where it uses more power than an equally clocked i7 at 3.7GHz.

This would explain why AMD couldn't break 4GHz on air cooling.

original hwbox numbers were all very strange - not sure they have any idea what they are doing.

Intel added thermal envelope protection to i7 because of concern about local hot spots and all the big OC turn it off, because otherwise it shuts the chip down on 4 cores at around 3.5G. Intel also allows full parking of 2 cores to let the other 2 use the thermal capacity, which gives more headroom. The thermal envelope envelope is targeted at 130W.

Looking at screen shots here and on other forums, there seems to be no rapid increase in power on phenom II, at least up to about 4.2G clocks.

It is not reasonable to compare different architectures based on core voltage - power draw with voltage differs even among members of the same family. The best way is to instrument the VRM but that is not easily done. A method that is nearly as good is to use heat rise of a known heat sink. If you have a .1 C/W sink and show a 10C rise, that is a pretty good indicator that the load is 100W. I have several sinks with embedded thermistors at the base. With a controlled ambient air supply that method can be pretty accurate.

Until retail phenom II parts are available, I won't be able to test any by myself, but I will characterize one as soon as I can get it. I am expecting a significantly lower power envelope than i7.

Also, as several posters have pointed out, internal temps are higher than the heatsink contact, often 10C or more. Lower power levels reduce those hot spots - one of the reasons laptop CPUs may run 70C or 80C in 'normal' use. Still, I'm a lot happier with the measured heatsink interface at 50C or less.

[roofsniper]

At stock perhaps, but the updated Hwbox review suggests Phenom II's power usage grows at a incredible rate when overclocked, to a point where it uses more power than an equally clocked i7 at 3.7GHz.

This would explain why AMD couldn't break 4GHz on air cooling.

research before posting. i believe the first leak ever about phenom II was when coolaler reached 4ghz on air and we have seen many others at 4.5ghz on air too. also if you would look at the chips of many people clocked here you would see that even at 3.7ghz it still stays pretty cool. look at all the screens charged and honda guy have posted. many people are reaching 3.5, 3.6ghz on stock volts too so it really isn't that power hungry. i have no clue where hwbox got all of their numbers from but the power consumption seems really far off compared to every other screen we have seen for phenom II so far.

[Caveman787]

Umm I have nott seen these 4.5ghz on air for phenom II

If you can link to them that'd be nice because so far the highest I've seen on air is 4.2ghz.

[Zucker2k]

research before posting. i believe the first leak ever about phenom II was when coolaler reached 4ghz on air and we have seen many others at 4.5ghz on air too. also if you would look at the chips of many people clocked here you would see that even at 3.7ghz it still stays pretty cool. look at all the screens charged and honda guy have posted. many people are reaching 3.5, 3.6ghz on stock volts too so it really isn't that power hungry. i have no clue where hwbox got all of their numbers from but the power consumption seems really far off compared to every other screen we have seen for phenom II so far.

Stop showing your ignorance! My God, how many times have you posted on this page alone! No two chips are the same. In any case, what was Coolaler's chip? ES. Maybe that's all you need to know. Man, slow down!

[Raiderman]

Stop showing your ignorance! My God, how many times have you posted on this page alone! No two chips are the same. In any case, what was Coolaler's chip? ES. Maybe that's all you need to know. Man, slow down!

Name calling is not the way to get a point accross. Slow down, take a deep breath, it usually helps before inserting a foot.

[informal]

Name calling is not the way to get a point accross. Slow down, take a deep breath, it usually helps before inserting a foot.

Talking to him is like talking to a wall,no effect.But it's good you tried,maybe it works this time

As for Deneb's power draw,I'm on the same page as iocedmyself,Uncle Jimbo and roofsniper. Deneb should draw noticeably less power at stock and still less power OCed/overvolted than i7 system.I'd expect overall that it will be equal/lower than Yorkfield systems if Deneb is paired with newest 790GX boards.

[Zucker2k]

Talking to him is like talking to a wall,no effect.But it's good you tried,maybe it works this time

The only wall I see is you; at least that's how dumb you sound with your entrenched position. You can add your buddy, roofsniper to that list. I know it hurts, but this is hardware so get over it.

[roofsniper]

The only wall I see is you; at least that's how dumb you sound with your entrenched position. You can add your buddy, roofsniper to that list. I know it hurts, but this is hardware so get over it.

woa woa chill out dude. just because your sad that intel isn't best at everything doesn't mean you have to get all angry about it. don't you remember when you said this earlier? or have you already forgotten? its pretty funny how people here are just pointing out the obvious and you get so angry about it.

snip.
Awww, you're breaking my heart. Can we keep personal feelings out of this thread? Anyway, I'm curious; why don't you like me?

[Throwed]

The only wall I see is you; at least that's how dumb you sound with your entrenched position. You can add your buddy, roofsniper to that list. I know it hurts, but this is hardware so get over it.

Dude, relax. It's not that serious. I honestly don't understand why you aren't banned yet.

[justapost]

Stop showing your ignorance! My God, how many times have you posted on this page alone! No two chips are the same. In any case, what was Coolaler's chip? ES. Maybe that's all you need to know. Man, slow down!

Well Coolaler has access to alot of chips to pick the good ones. Nothing wrong with that.

Has Overclock_gr posted specs used for power consumption anywhere? SOmethin like a screenshot wit everest cpu-z aod and so on showing all the settings he used. They must have done such an screenshot, they spended so much time running all the benchmarks, would be sad if they missed that one.

[accord99]

research before posting. i believe the first leak ever about phenom II was when coolaler reached 4ghz on air and we have seen many others at 4.5ghz on air too.

None of which were shown to be truly stable.

also if you would look at the chips of many people clocked here you would see that even at 3.7ghz it still stays pretty cool. look at all the screens charged and honda guy have posted. many people are reaching 3.5, 3.6ghz on stock volts too so it really isn't that power hungry. i have no clue where hwbox got all of their numbers from but the power consumption seems really far off compared to every other screen we have seen for phenom II so far.

Except we've never seen any other power consumption measurements for Phenom II so far. Given hwbox's results, it's far more likely that Phenom II temp sensors are inaccurate, or at best not directly comparable to Intel sensors readings.

[roofsniper]

None of which were shown to be truly stable.


Except we've never seen any other power consumption measurements for Phenom II so far. Given hwbox's results, it's far more likely that Phenom II temp sensors are inaccurate, or at best not directly comparable to Intel sensors readings.

coolaler said that they were stable, im not sure about the other 4.5ghz ones but i doubt any tests were run. and considering that phenom II's temp sensors are the same as phenom I's i doubt they would be inaccurate. now the am3 ones are a different type of sensor so they might be.

[accord99]

coolaler said that they were stable, im not sure about the other 4.5ghz ones but i doubt any tests were run. and considering that phenom II's temp sensors are the same as phenom I's i doubt they would be inaccurate. now the am3 ones are a different type of sensor so they might be.

The Phenom I's aren't that accurate, or at least they don't measure the same thing that Intel temperature sensors do.

http://forum.xcpus.com/amd/9811-phen...-behavior.html

[Throwed]

Is there a way to ignore a users post?.....In a way that their post never pop up as long as you are logged in?