My Phenom II benchmarks

Table21

Limp Gawd
Joined
Oct 17, 2006
Messages
293
Sorry for making another thread, not sure if it is the right thing to do but anyways, here is what I have for a system to test things on.

Cooler -OCZ vendetta 2
CPU - Phenom II 940
Ram - 2x2gb Mushkin redline
mobo - MSI K9A2 plat
video - 2x 4870, 4x4850 (installed right now is 1x 4870)
H/D - WD5000AAKS
PSU - Corsair HX1000
O/S vista ultimate 64bit

I have super pi runs done now and will do some 3dmark 06 in a few min

Hpic.jpg
 
You said you had a Phenom II 945, which is AM3, this clearly shows that it's the AM2+ (Phenom II 940)

However, i'm still happy that someone has one, if only you got something to compare this to.. MOAR BENCHMARKS!
 
If you've got a current gen Phenom can you post a bench of the comparison?
I'd like to know what i'd be stepping up to if I bought one.
Also,
TEMPS PLEASE (On stock cooler) :(
 
sorry I am not really an amd guy I have only a 5000be thats been sort of out of use for a while.
 
You said you had a Phenom II 945, which is AM3, this clearly shows that it's the AM2+ (Phenom II 940)

However, i'm still happy that someone has one, if only you got something to compare this to.. MOAR BENCHMARKS!

actually it shows its a Socket AM2+ (940 pins) CPU thats because its in an AM2+ Mobo, i would only assume it would show AM3 when its in an AM3 mobo
 
actually it shows its a Socket AM2+ (940 pins) CPU thats because its in an AM2+ Mobo, i would only assume it would show AM3 when its in an AM3 mobo

It shows AM2+ because it is an AM2+...

That chip has an unlocked multi... There are no unlocked multi's
schedualed for AM3 (only the AM2+ 940) with the exception of
the X3 720 (Tricore)... The roadmap only goes to the 3rd Qtr
though so there may be some after that, or AMD may add chips
to the linup out of the blue...

The AM3 945 isn't scheduled for release until 2nd Qtr (april, may
or june) so I doubt we'd see any early handout samples until feb.
the earliest... You may see some early samples of the AM3 925
in jan. though as that chip releases in feb... ;)

30b.jpg
30c.jpg

Source: http://www.techpowerup.com/76417/Phenom_II_Final_Nomenclature_and_Launch-Schedule_Revealed.html
 
Wow. This appears to be right up there with the Core 2 Quads. 3.8 on air too... let's pray AMD gets this out soon.
 
It would be nice to be able to build an economical AMD Quad core system that can play games as well as my year old Q6600. I was always an AMD guy because I like the underdogs so if the Phenom II can hang at a decent price I will be ordering up a MB and 940 as soon as they are released. The rest of the system is sitting here on my workbench just waiting for a worthy heart.

Coolermaster 590 case.
4 x 120MM case fans
OCZ StealthXStream 600W PSU
WD 320GB 7200.1 HD SATA
Samsung SH-223B DVD SATA
Saphire HD4850X2

I would like to keep it pure AMD so I am rooting for decent if not top shelf gaming performance from the Phenom II.
 
Wow. This appears to be right up there with the Core 2 Quads. 3.8 on air too... let's pray AMD gets this out soon.

If its anything like the C2Q chips though, thats too much voltage for 24/7. Wonder what the safe voltage is on these things?
 
having had a bit more time to play with it it does not seem to change things if I drop my vcore down to 1.4, I am running into some problems adj the NB right now, which seems to be why I am stuck at 3.9GHz
 
looks decent, the stock scores surprised me a bit, just under 1k from what i score.

Keep the results coming!
 
i could show some comparisons between a brisbane be-2350 at 3ghz if that would help at all for comparisons if some one wanted me too.. and ill talk to another guy thats using a phenom 9750 Black edition and have him overclock it to 3ghz to show a comparison between that and the phenom II.. just pm me if you want me to and ill do it..

nvm i was bored.. so heres the comparison to a AMD X2 BE-2350 @ 3045mhz w/ single 8800GT (overclocked numbers are in the screen shot)

2ykavd5.jpg


i know my signiture says i have a AMD Athlon X2 6400+ but i ran the same tests on that one and my BE-2350 actually scored better in every test.. im leaving complete blame on the crappy nvidia chipset on my m2n-sli deluxe motherboard. Also my best time on the Super PI is actually 30.765 seconds for 1M..
 
It shows AM2+ because it is an AM2+...

That chip has an unlocked multi... There are no unlocked multi's
schedualed for AM3 (only the AM2+ 940) with the exception of
the X3 720 (Tricore)... The roadmap only goes to the 3rd Qtr
though so there may be some after that, or AMD may add chips
to the linup out of the blue...

The AM3 945 isn't scheduled for release until 2nd Qtr (april, may
or june) so I doubt we'd see any early handout samples until feb.
the earliest... You may see some early samples of the AM3 925
in jan. though as that chip releases in feb... ;)

I was under the impression that engineering samples had their multipliers unlocked regardless of how the retail version is locked.
 
I was under the impression that engineering samples had their multipliers unlocked regardless of how the retail version is locked.

Some may, some may not...

Engineering samples can be just about anything as they are made
for testing... Many times they may even fall short of the performance
of the retail poduct...

With an ES anything can happen depending on what it was
made to test... ;)
 
Ah I see... Well I guess then the cpu could still be an AM3 version. I have no clue how CPU-Z determines the socket though, it could just detect that the cpu is being used on an AM2+ socket.
 
werent they supposed to be released on the 20th and begining of January for the AM3 version thats 5-10% faster????

I am going with AMD, its time to let gone of my aging 6000+. wanted to give intel some love but price to performance is again too much for an intel system.

I am going to love supporting AMD anyway, they need it.
 
Looks good, hopefully we can get this out soon to help stimulate competition.
 
Default vcore for PhII is 1.35v
So, 1.45-1.50v on air is perfectly safe 24/7.

Looks like it. I think people put too much emphasis on the actual voltage number in relation to other CPUs. That is simply a number in a desirable range for each CPU specifically, not in general for all CPUs everywhere. The real number to care about would be the effective wattage (ie- power) given to the CPU. Let's say a CPU needs 70W to operate. At 1V that CPU would draw 70A. At 2V the CPU would only need to draw 35A, which is generally much easier and safer for the power regulation equipment to handle. Of course there are a ton of other factors to account for like what happens to the electrons at a certain voltage once they get to the CPU, etc, but that's a very basic idea of what's going on. I think AMD designed the CPUs to run at "higher" voltages in order to reduce the amperage needs on the voltage regulators, which happened to be a pretty big problem with some of the higher clocked Phenom I's. Probably also has to do with backwards compatibility of the chips to go into AM2+ mobos with power circuitry designed for that spec. Whatever the case, the chips seem to like having that kind of voltage available to them when they overclock and there aren't any rampant complaints of dying mobos or astronomical temperatures (yet).
 
At 1V that CPU would draw 70A. At 2V the CPU would only need to draw 35A, which is generally much easier and safer for the power regulation equipment to handle.

Ummmm....my understanding is that if you change the voltage, you change the power consumption of the processor. It doesn't draw the same amount of watts at all voltage settings.

It may draw more current at lower voltages, but not in direct 1-1 proportion.
 
Ohm's law states that amps = volts / resistance, so that if the resistance (the load) remains constant,
an increase in voltage will actually result in an increase in current flow. And power = volts x amps, so
a doubling of the voltages results in 4 times more power consumption.
 
Hmmm while these preliminary results look somewhat promising for the Phenom II, the comments on the HardOCP front page seem to hint that there isn't much to get excited about. I wonder what is up with that? Maybe they have had time to run some more extensive benchmarks and they are not impressed with the results?
 
Yep, no way it draws less current with more voltage, even if you consider the capacitive and inductive properties of the silicon junctions. Digging deeper, the current density is equal to the conductivity times the electric field. Electric field is simply Voltage over the distance the E-field travels. So, higher E-field = higher current density crossing a load of the same length, leading to higher current.

Just because a CPU is designed for 70W doesn't mean it will drop down to 35 A if you push it to 2 V.
 
Ohm's law states that amps = volts / resistance, so that if the resistance (the load) remains constant,
an increase in voltage will actually result in an increase in current flow. And power = volts x amps, so
a doubling of the voltages results in 4 times more power consumption.

Right so in going to a smaller process I would assume resistance has gone up (smaller path for the electrons to flow), meaning a decrease in overall amperage with a similar voltage (1.35v). So in designing a new chip you would have the possibilities of lower power draw and current with the same voltage and increased resistance, or current stays the same, power voltage and resistance go down, correct? I'd have to assume AMD went more towards the former. Perhaps current leakage was a major problem, or they couldn't get resistance down any farther with the materials?
 
Back
Top