AMD Shanghai/Deneb Review Thread

[Uncle Jimbo]

You know it's gotta be good if Anand likes it.

IoC mentioned that the production parts had minor improvements from the ES parts - Anandtech seems to confirn that.

In fact, the latest retail steppings that AMD displayed this past week showed significant improvements in overclocking headroom compared to the press samples we utilized. Our 940 topped out at 3.9GHz, which is not bad, but after reviewing AMD’s results and seeing some early retail numbers on the forums, the expectation level for air-cooling is now set to the 4.1GHz range with the 920 hitting 3.8~3.9GHz on the right motherboard.

[Uncle Jimbo]

Are people really happy with ph2? I mean not even slightly disappointed? Read the comments on anandtech and others and it seems that most are disappointed, It still lacks clock for clock against all Intel quads and the overclocking and power consumption dont live up to the hype.

I'm sure people who were hoping for an i7 killer are disappointed - but those with more realistic expectations (that AMD would produce a part which is price competitive with equivalent C2D parts, and offers good overall platform advantage for OC) should be pleased with the results.

The advantage for buyers is big, too - wider choice and lower costs for all but the extreme upper end of the market.

I doubt Intel will take pricing action on i7, but they will probably improve Penryn value, and may also extend the life of those parts, which is a good thing IMO.

I don't think we can say yet how the power will play out. C2D had great power management which could be improved without much work, where Phenom was a hog for sure. Phenom II might be in the same power range as C2D - time will tell. AM3 may make a difference too...

[knopflerbruce]

Are people really happy with ph2? I mean not even slightly disappointed? Read the comments on anandtech and others and it seems that most are disappointed, It still lacks clock for clock against all Intel quads and the overclocking and power consumption dont live up to the hype.

I guess the "power consumption" depends on how you measure it. SOmeone used a temp probe and some elementary physics to get a pretty decent number

There are a whole lot of results out there now, and eve if it's not i7 killers they've released I'd say AMD did an A-W-E-S-O-M-E job - these chips are HUGE improvements over Agena, and that's the only fair comparison. The fact that they can compete with Intel quads is also a good sign.

I don't really see why Anandtech's conclusion should be so important, either. it seems obvious to me that that's just some Intel fan's opinion. "If Intel chooses to do this and that"... well, that's not the case TODAY, so its' a useless comment. The fact is that the Deneb platform gives you alot of performance for your $$$.

[TL1000S]

I agree with "Uncle" and "Knopfler".

PHII shoud be compared to PH1, and as such it is a great step in the right direction.
920/940 *do* compare great with current Intel C2Q (Q6600/Q9300/Q9400).

For anybody wanting to build/buy a new quad-system there is no reason to prefere/choose Intel over AMD.

For those already having an AM2+ rig the extra cost of purchasing a 920 as an upgrade is a matter of taste really, and perhaps support the "Green team" (both AMD and environment.. as in saved energy).

[gallag]

So the new top end AMD competing with the q9300-q9400 is enough? I remember that when people suggested this months ago they got a serious flaming, People expected more. To be fair for the price deneb is ok but I was hopping for some competition in the high end, I find it hard to get excited about a product that would have been good 1.5 years ago.

[Uncle Jimbo]

I guess the "power consumption" depends on how you measure it. SOmeone used a temp probe and some elementary physics to get a pretty decent number

There are a whole lot of results out there now, and eve if it's not i7 killers they've released I'd say AMD did an A-W-E-S-O-M-E job - these chips are HUGE improvements over Agena, and that's the only fair comparison. The fact that they can compete with Intel quads is also a good sign.

I don't really see why Anandtech's conclusion should be so important, either. it seems obvious to me that that's just some Intel fan's opinion. "If Intel chooses to do this and that"... well, that's not the case TODAY, so its' a useless comment. The fact is that the Deneb platform gives you alot of performance for your $$$.

I agree - this seems to be a very good part.

Anandtech had an interesting chart on OC - and they pointed out that the production parts would probably move Phenom II 940 to 4.1G or more.

But even using their ES they got the following, which was done using STOCK cooler, and shows the Phenom II 940 looking pretty good:

Processor ............ Stock Voltage . Overvolt. % Increase . Vcore
AMD Phenom II 940 .. 3.2GHz ......... 3.9GHz ...... 30% ..... 1.52V
AMD Phenom 9950 BE 3.03GHz ....... 3.38GHz ..... 30% ..... 1.45V
Intel Core i7-920 ..... 3.83GHz ....... 4.0GHz ....... 50% ..... 1.35V
Intel Core 2 Q9550 ... 3.48GHz ....... 3.91GHz ..... 38%...... 1.35V

[roofsniper]

lol, You gotta be kidding. It will be closer to double that.

seems like a joke but jimbo does know what he is talking about. he is way above all of us in that field.

Correct - the 'I' you are measuring may or may not be going to the CPU, and likewise you don't know what 'E' actually gets supplied internally, but there is no question about the heat load.

Here's a good write-up on heatsink characterization by some guys who actually know what they are doing:
http://www.frostytech.com/testmethod_mk2.cfm

lol and everyone tells me that frosty tech sucks.

Are people really happy with ph2? I mean not even slightly disappointed? Read the comments on anandtech and others and it seems that most are disappointed, It still lacks clock for clock against all Intel quads and the overclocking and power consumption dont live up to the hype.

the thing here is that most of us are enthusiasts. we buy the parts because we like them. for more mainstream users they might look at the graphs and think differently and choose intel over amd but most of us here don't need absolute performance. most people just like tweaking and messing around with the systems to get more performance. so deneb is a dream for amd users. im really happy with the power consumption since it blows the original phenom out of the water and i bet running it at 3.5ghz would produce the same or less power than my phenom 9600 does at 2.3. it does stay really close to the core 2's power consumption even with the imc and large l3 cache too. and the fact that cool and quiet can reduce the clocks down to 800 mhz and the below 1V is just nice for me.

[TL1000S]

the thing here is that most of us are enthusiasts. we buy the parts because we like them. for more mainstream users they might look at the graphs and think differently and choose intel over amd but most of us here don't need absolute performance. most people just like tweaking and messing around with the systems to get more performance. so deneb is a dream for amd users. im really happy with the power consumption since it blows the original phenom out of the water and i bet running it at 3.5ghz would produce the same or less power than my phenom 9600 does at 2.3. it does stay really close to the core 2's power consumption even with the imc and large l3 cache too. and the fact that cool and quiet can reduce the clocks down to 800 mhz and the below 1V is just nice for me.

What he said.

Could not agree more.

[Mats]

Any HTT overclocking reviews?
I'd like to see a comparison of a multiplier and a HTT overclocked 940.
Benchmarks (even though I don't expect much difference), temps and power consumption.

[gallag]

uncle jimbo, Now that the reviews are out and power consumption is known do you still think that "the estimate of under 100W at 4G seems reasonable."

[Lightman]

uncle jimbo, Now that the reviews are out and power consumption is known do you still think that "the estimate of under 100W at 4G seems reasonable."

Well, it probably is too low by 20-50%, but it won't be 200W as suggested earlier.
67W 4core Prime @3GHz is not bad. (and that's before PWM, radiator is dissipating 5-25% less due to PWM efficiency)

[roofsniper]

uncle jimbo, Now that the reviews are out and power consumption is known do you still think that "the estimate of under 100W at 4G seems reasonable."

well for the cpu alone it might be correct. we are seeing under 200W for the entire system load and as lightman said the pwms use a lot of power. i wish someday people would design some good ones. or get something better than passive cooling on them.

[Uncle Jimbo]

uncle jimbo, Now that the reviews are out and power consumption is known do you still think that "the estimate of under 100W at 4G seems reasonable."

Let me put this in perspective -
I did a fair amount of testing on a C2D E8600, which is IMO a great combination of speed and low power. I was able to set it up to deliver 4G with idle power of 30W and load power around 80W at 1.26 vCore. Total power to the board was 220W... so RAM and chipset used nearly twice as much power as the processor itself.

Because of improved power management, i7 has lower idle power, but runs up at load. A 965 at stock clock and voltage was under 25W at idle, but over 140W at load. Power to the board on the 965 was over 290W at load.

I have tested the ASUS M3A79-T Deluxe MB, which uses 790FX. I have found that with a 4830 graphics card, with 4GB of DDR2, and the processor load subtracted, it draws around 90W with a CPU intensive load and everything at stock voltage. That can go up a lot with increased voltage on RAM and chipset - up to 120W or more.

So with a number of phenom II screen shots showing total load at 170W, 180W and such at speeds near 4G, it does seem reasonable that the processor is drawing less than 100W of the total.

But all of that is guesswork until I get one of the Phenom II chips and test it myself. There is not enough good data available from other testers to do more than a ball-park estimate.

[Uncle Jimbo]

lol and everyone tells me that frosty tech sucks.

I can't comment on their other review work, but their heatsink characterization method is perfect.

[gallag]

thanks jimbo, I look forward to your results.

[roofsniper]

I can't comment on their other review work, but their heatsink characterization method is perfect.

yea thats what i use them for. i think many people look at it and see that they put the TRU above the TRUE and just disregard them. thanks for the info on this.

[The PyroPath]

Hmm at least idle it uses the same as my 9500 ;-) can't do much more atm because my PSU isn't too shabby... and still trying to find out why some things act a bit weird (got a Scythe Musashi with 2 fans, one runs @1250 while the other runs @2050, and my all new artic freezer extreme is going fast/slow/fast/slow/etc and give unreliable RPM numbers, while my fancontroller works flawless with all other PWM fans....

[qurious63ss]

Correct - the 'I' you are measuring may or may not be going to the CPU, and likewise you don't know what 'E' actually gets supplied internally, but there is no question about the heat load.

Here's a good write-up on heatsink characterization by some guys who actually know what they are doing:
http://www.frostytech.com/testmethod_mk2.cfm

So you measure temp at the base of heatsink? So that means the better your heatskink is at dissipating the heat, the lower the temp at the base of the heatsink should be, correct? So how does heat translate into the PIE formula? Power is still a function of current and voltage so how would lowering the temp of the heatsink affect power draw?

[ajaidev]

After reading almost all the PhII reviews on the net i came to at least one conclusion about the OC tests carried out by the reviews that most of them were utter arse holes at the job.

Some sites were BIAS enough to use the default AIR HSF and overclock and then rate the max OC at 3.4-3.5Ghz what the hell is that suppose to be.

Others upped the Ghz but not the Voltage enough or touching the HTT with an excuse that something was apparently wrong with their sample.

Then there were the great peple who say we used Win 64 for the test and in their i7 review they used win 32 bit wtf.

Seems that the whole pc community is ruled by Intel or something too many simple mistakes made by pro people. Also the fact that in most games the PhII's had min FPS around I7 920's min FPS rate was not mentioned in the conclusion...

[The PyroPath]

I've something weird, running the stresstest of AMD overdrive my Phenom II core temp goes to 63 degrees celsius and settles there. Much higher then in the reviews... also my motherboard reports a 55 degrees celsius temperature(which is measured the old fashion way. All at idle and on C&Q it's about 38...

Is my HSF mounted wrong, are their numbers incorrect or am I just crazy?

My CPU cooler is the Freezer Extreme. Not a bad cooler by any means...

[qurious63ss]

I've something weird, running the stresstest of AMD overdrive my Phenom II core temp goes to 63 degrees celsius and settles there. Much higher then in the reviews... also my motherboard reports a 55 degrees celsius temperature(which is measured the old fashion way. All at idle and on C&Q it's about 38...

Is my HSF mounted wrong, are their numbers incorrect or am I just crazy?

My CPU cooler is the Freezer Extreme. Not a bad cooler by any means...

Start with remounting heatsink and if that doesn't work try using an other program to read temps and possibly look at BIOS rev for your motherboard.

[Uncle Jimbo]

So you measure temp at the base of heatsink? So that means the better your heatskink is at dissipating the heat, the lower the temp at the base of the heatsink should be, correct? So how does heat translate into the PIE formula? Power is still a function of current and voltage so how would lowering the temp of the heatsink affect power draw?

A CPU is not a resistive load, and actual draw varies from one microsecond to the next. There is also no way to determine 'I' for a CPU unless you instrument the VRM. So it's not practical for most people to use P=I*E since they can only guess at 'I'. We do know that a CPU does no mechanical work, so all of the input power is turned into heat. Basically the heatsink method gives that power directly.

Since we know (or can determine) the thermal resistance of the heatsink, we can determine the power dissipated by applying the thermal resistance Tr (in C/W) where C is the difference between the base and ambient temperature. The formula is Crise / Tr = Wload. The actual temperature at the base, along with ambient, is only used to determine the rise.

Lowering the temperature at the base will have some effect on total power, but not very much unless the change is large. It's always best to use a big cooler, but the goal is to get rid of the heat, not reduce power.

If for some reason you want to calculate amps input, and you know power (from the heat load) and E (from the Vcore) you could use I=P/E to find amps... but unless you are designing a VRM, I'm not sure why you would do that.

[Zucker2k]

Snip

Uncle Jimbo what do you make of this?

I am certain that the reason so many overclocks by reviewers were so low was the sheer lack of experience with the systems. I was afraid of blowing up my sample before the launch date so I stuck with a conservative 1.45vCPU which yielded a paltry 3800MHz. Now that the article is live, I'm blasting 1.55vCPU through this chip at 4200MHz and climbing

The biggest issue I have though is the lack of temperature readings, I'm currently using a thermalprobe wedged up against the IHS of the actual CPU to get a ballpark figure :-\ It's reading around 55C at idle while the thermal diodes on the chip still read 19C

From here: http://www.xtremesystems.org/forums/...=213691&page=5

[roofsniper]

jimbo doesn't some of the power get burned off as sound? not much but a little? thats why the computer makes noise even when all fans are off?

[Uncle Jimbo]

Uncle Jimbo what do you make of this?

From here: http://www.xtremesystems.org/forums/...=213691&page=5

Not sure what to say about this... he doesn't give a lot of detail, and the numbers make no sense