AMD Phenom II 920 & 940 Full Review [UPDATED with more tests]

[justapost]

No, total system consumption is still the best way to measure CPUs, now that Phenom and i7 have seperate power rails that don't just use +12V alone (IIRC).

For the PII it's still valid. But you are right it does not make sense with i7 any more.

@roofsniper: thx for details, prime95 varies alot depending on FFT size. povray benchmark generates an good consistent load and it's very similar to CB10 load. for worst case i always ask for CoreDamage or Core2MaxPerf.

[roofsniper]

For the PII it's still valid. But you are right it does not make sense with i7 any more.

@roofsniper: thx for details, prime95 varies alot depending on FFT size. povray benchmark generates an good consistent load and it's very similar to CB10 load. for worst case i always ask for CoreDamage or Core2MaxPerf.

i have no clue if he used prime 95 or not but i know most reviews do. its just a guess. im guessing the power consumption is legit but i think it would get better power consumption with the more popular 790fx.

[Dante80]

Either way I will wait for more professional/in-depth reviews.

In what way do you think this review is not professional enough? The guys are busting their balls even on New Years eve to give us leechers numbers to crunch...



O_o...what a powerhog this 780!...

[roofsniper]

well it would be nice if there are more reviews just so you can confirm things and things will be different in different reviews. as you can see there on idle alone the 780a uses 40W more and that will expand out with load. you can't look at the power consumption on this review alone and will need more reviews with different configurations.

[Dante80]

Of course it would be nice to have more reviews. The more, the better.

And you are right about platform power consumption, the more different setups we have, the better (p45 instead of only X48, 790fx and gx numbers, etc etc).

Also, different platforms may give different results. For example, you may get better numbers at stock for phII if you tweak your system more (this comparison is made on an un-tweaked vista os), put better memory in, oc only the nb etc etc. Same goes of course for all the intel setups too.

[justapost]

In what way do you think this review is not professional enough? The guys are busting their balls even on New Years eve to give us leechers numbers to crunch...

From an complete power consumption I'd expect detailed system setup descriptions including used voltage meter, tools used to generate load, software running in background (if), ambient temps, idle and load temps, all bios settings whom affect load, all important voltages used, maybe even an efficiency diagramm of the psu used. Did I miss something?
Really appreciate their hard work, but these infos are required for comparisons.

[annton]

Phenom ii gaming performance is good. With it's oc headroom, it will have very good price/performance ratio when oc.

[sundancerx]

Phenom ii gaming performance is good. With it's oc headroom, it will have very good price/performance ratio when oc.

not to mention a lot cooler and lower power consumption

[bldegle2]

BenchZoner-----

"First of all, the decent DDR3 kits can do DDR3-1400 5-5-4-15, and some can even go up to DDR3-1500 - DDR3-1550 at 5-5-5-15."

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

all that information is nice...sorry for the misunderstanding....I was basing my reply based on a review of all the DDR3 modules presently available @the Egg, the best timings I saw were 6-6-6-20, many were 7-7-7-xx, and several were 9-9-9-xx...even ddr2 rated @1200 comes with a 5-5-5-15/18 timings........DDR3 with take your wallet right now.....

i am sure it will get better as the transition continues, I will research DDR3 some more regarding the timings you mentioned....

thanks...

laterzzzz.................

[OverClocker_gr]

For the power consumption tests the mobo i used is M3A79-T so,forget about 780a

For the 3D tests, Crysis

CPU load,Cinebench/orthos

[roofsniper]

For the power consumption tests the mobo i used is M3A79-T so,forget about 780a

For the 3D tests, Crysis

CPU load,Cinebench/orthos

wait so which benchmarks used the 79T? cause in the system specifications page it only says foxconn destroyer.

[OverClocker_gr]

i said for power consumption i used 79T.


And for the clock to clock 79T also used

[sundancerx]

i said for power consumption i used 79T.


And for the clock to clock 79T also used

hope you goin to include power consumption test on agena vs. deneb test at 3.0ghz ctc comparison? id like see how mch imprvement deneb made over agena on this metric.

[OverClocker_gr]

http://www.xtremesystems.org/forums/...&postcount=125

[PetNorth]

For the power consumption tests the mobo i used is M3A79-T so,forget about 780a

For the 3D tests, Crysis

CPU load,Cinebench/orthos

Thats interesting cause it's my current mobo (with Ph 9850), and I've noted that if a set vcore to Auto (or manually to 1.30v, 9850 default vcore), this mobo overvolts it in load to 1.344v. I've measured power consumption in these circumstances, and means ~12w more comparing it with a real 1.30v, exactly 1.296v (manually to 1.265v), or ~10w more with 1.31v (manually to 1.275v).
Have you checked if overvolts PhII too?
What about Intel rigs?

[sundancerx]

http://www.xtremesystems.org/forums/...&postcount=125

anyway, good effort.

[demonkevy666]

deneb's power is fine.

you don't know the efficiency of power supply he used.

effciency can be form anywhere from 60%-88% (88%) being the highest "rating" I've seen

[Hornet331]

Thats interesting cause it's my current mobo (with Ph 9850), and I've noted that if a set vcore to Auto (or manually to 1.30v, 9850 default vcore), this mobo overvolts it in load to 1.344v. I've measured power consumption in these circumstances, and means ~12w more comparing it with a real 1.30v, exactly 1.296v (manually to 1.265v), or ~10w more with 1.31v (manually to 1.275v).
Have you checked if overvolts PhII too?
What about Intel rigs?

Interesting, first time i heared that mobos/cpu overvolt under load, usually they have a drop.

[VoodooProphetII]

Oh yes... my Asus M3A79-T Deluxe overvolts to 1.32 to 1.33 when I set the voltage in the bios to 1.3, and it would slightly increase to about 1.34 under load. That's when running a Phenom 9950. Even when I overclock the processor to 3.3Ghz and set 1.4Vcore in the bios, it will always be about 0.02 higher when idle and 0.04 higher under load.

[donitsi]

With some bioses my ASUS M3A mobo overvolts a bit
Weird

[leoy]

Mine set @ 1.3875 real from cpu-z and aod 1.424 under load.
All m3a79-t overvolt at all volts.

[OverClocker_gr]

Thats interesting cause it's my current mobo (with Ph 9850), and I've noted that if a set vcore to Auto (or manually to 1.30v, 9850 default vcore), this mobo overvolts it in load to 1.344v. I've measured power consumption in these circumstances, and means ~12w more comparing it with a real 1.30v, exactly 1.296v (manually to 1.265v), or ~10w more with 1.31v (manually to 1.275v).
Have you checked if overvolts PhII too?
What about Intel rigs?

On asus when you see at cpu-z that on load it overvolts about 0.022v or so, its actual no overvolt and no vdroop.

I ve measured that on 4 asus(commando,maximus formula,Evolution,m3a79-t) with a DMM.

[justapost]

AsRock AOD790GX shows huge overvolts ~0.06V with modest overclocks. I measured vcore with an DMM on the voltage regulator vcore sensor pin and there it was ~0.02-0.03V. Only boards with low overvolt are those from MSI.

Thanks for the details overclocker_gr. Jsut to be sure, you ran two instances of orthos together with the CB10 multi core test. Correct?
Have you had C1E enabled in the bios?

[PetNorth]

Interesting, first time i heared that mobos/cpu overvolt under load, usually they have a drop.

that's what happening with my Intel rig (P5B + Q6600). For example, I set vcore to Auto, and under load, undervolts to 1.25-1.26v (vcore default is 1.30v). If I set it manually to 1.30v, undervolts ~1.26v too... etc

But on the other hand, with ASUS AMD mobos occur the contrary usually, overvolts CPU like is the case with this M3A79.

These issues are crucial to keep in mind specially when power consumption tests are done, to obtain accurate results.

[OverClocker_gr]

Thanks for the details overclocker_gr. Jsut to be sure, you ran two instances of orthos together with the CB10 multi core test. Correct?
Have you had C1E enabled in the bios?

4,nop