Overclocking
When you consider that our 2.6GHz Phenom needed more voltage just to be stable at 2.6GHz, we didn't have a great feeling about overclocking. We found that upping the HyperTransport was a virtual no-no - we got 205MHz out of it but above that and things crashed regardless of extra voltage applied.
We were therefore limited to multiplier overclocking, which is great for the multiplier unlocked samples we get, but not so good for the general public. If AMD does release a "black edition" of the 9600 like has been hinted, then that makes things a lot better for you.
Getting back to this chip though, we could barely get it to 2.7-2.8GHz "stable" at 1.5-1.55V CPU Vcore. As it is, it was barely stable at 2.6GHz with elevated voltage so this is hardily surprising. Let's hope that AMD sorts its process technology out before it releases the Phenom 9900, because our experience with it has been less than pleasant from an end user perspective.
It was stable enough to get reliable results out of it, but it wasn't particularly stable for prolonged periods of time.
Power Consumption
-
Core 2 Duo E6750 (2x2.67GHz, 1333MHz FSB)
-
Core 2 Duo E6850 (2x3.00GHz, 1333MHz FSB)
-
Core 2 Extreme QX9650 (4x3.00GHz, 1333MHz FSB)
-
Phenom 9500 ES (4x2.2GHz, 2.0GHz HTT)
-
Phenom 9600 ES (4x2.3GHz, 2.0GHz HTT)
-
Phenom 9700 (4x2.4GHz, 2.0GHz HTT)
-
Core 2 Quad Q6600 (4x2.40GHz, 1066MHz FSB)
-
Core 2 Quad Q6700 (4x2.67GHz, 1066MHz FSB)
-
Athlon 64 X2 6000+ (2x3.00GHz, 1.0GHz HTT)
-
Core 2 Extreme QX6850 (4x3.00GHz, 1333MHz FSB)
-
Athlon 64 X2 6400+ (2x3.20GHz, 1.0GHz HTT)
-
Phenom 9900 (4x2.6GHz, 2.0GHz HTT)
0
50
100
150
200
250
300
350
Watts (lower is better)
Despite all the results above being recorded without any form of power management technology used, we tried to enable Cool'n'Quiet in the BIOS to see the effect, but there was no noticeable drop in core frequency or voltage when we checked in CPU-Z in OS even after it had installed the driver.
The Phenom chips are all built on the newer 65nm process - the same as all current Core 2 processors except the QX9650, which uses the new 45nm process. Compared to the older Athlon 64 X2s, it's still an improvement on the 90nm SOI process and from the results you can see you get a quad-core for less than a dual-core.
So while the 2.2-2.4GHz Phenoms have a higher idle power consumption, they still turn up with similar power use to a similarly clocked Core 2 Quad at load. With that said, we found that the 2.6GHz Phenom 9900 sample we've got, which states a "1.3V" operating voltage on the heatspreader, needed at least 1.4V to keep it stable - this extra voltage has obviously affected its power consumption results somewhat.
Want to comment? Please log in.