|
|
 |
| Author: |
Vu Le |
| Date: |
2006.04.09 |
| Topic: |
Hardware |
| Provider: |
Multiple |
| Manufacturer: |
Multiple |
|
|
|
Overclocking
I actually had very high expectations for the 805 and it didn’t
disappoint I was truly impressed by how much headroom the 805 has. The inadequate
FSB may be a big reason for this. My only disappointment was that the multiplier
was fixed at 20. With that aside, I dove right into it. Right off the bat, I
increased the FSB to 700mhz, netting me 3.5ghz. Since it booted up fine, I went
ahead and increased it to 800mhz. It failed to start, so I increased the voltage
and with minimal effort got it stable at 3.9ghz @1.475v (although Asus PC Probe
and CPU-Z report lower readings than the bios). I’m certain I can hit
4ghz, but the temperature trade offs were not worth it.
*note: The Pentium D 805 was stable at 3.6ghz @ stock voltage.
I applied the same method to the AMD X2 3800+. The end result was 2.5ghz @ 1.4v.
I was able to hit 2.6ghz, but with an insane amount of voltage. The heat and
risks were certainly not worth it.
These were not cherry picked processors. They were purchased out
of my own pocket, so it’s not out of the question that you can expect
the same results with an aftermarket heatsink and fan. Others on the web have
been reporting 3.9ghz with just the stock cooling. Due to time constraints,
I was not able to test with the stock coolers.
Temperature
I was a bit worried about the temperatures of this "Bandaid Prescott."
However, the temperatures were surprising to say the least. Is this right? The
miniature space heater idling at 30c and 41c under load? The lower frequency
and front side bus can attribute to the lower temps. However, once the 805 was
overclocked, the temperatures ramped up. At 3.9ghz, the chip idled at 36c and
57c under load - not too bad. The X2 also surprised me. At stock and overclocked
speeds, the idle and load were exactly the same (34c idle, 54c load)! As a side
note, with the 805 at 3.825ghz, the idle and load were identical to the X2.
Power Consumption
Using my Kill-A-Watt power meter, I measured the power consumption
of the X2 system and the Pentium D system at idle and load. To make it even,
I removed the Geforce 7800GT from the X2 system and installed it on the Pentium
D system.
Power
Consumption - Idle watts in blue, Load watts in grey |
The overclocked Pentium D is one hungry processor. As you can see, AMD clearly
has the advantage when it comes to your electricity bill, but how much of an advantage?
If my math serves me correctly, the chart below is representative of the cost
of ownership. I'm going by the national average of $0.09 a kilowatt per hour.
Usage is 8 hours per day. Let's face it, no one is going to run their computer
at full load or leave it idle all the time, so, I'm going to average out the idle
and load power consumption. I won't bore you with the calculations, but just know
that it's right :).
Cost of ownership |
Intel Pentium D 805 (stock) |
AMD X2 3800+ (stock) |
Pentium D 805 @ 3.9ghz |
AMD X2 3800+ @ 2.5ghz |
| Cost of processor |
$132 |
$295 |
$132 |
$295 |
| Cost per month |
$3.80 |
$2.65 |
$5.09 |
$3.21 |
| Cost per year |
$45.60 |
$31.80 |
$61.08 |
$38.52 |
So what does this mean? From a stock perspective, in order for the cost of ownership
to be the same, it would take 11 years and 8 months. From an
overclocked perspective, it would take 7 years and 2 months for
the cost of ownership to be the same. In my opinion, the costs of ownership is
irrelevant as I don't keep my hardware for more than 5 years, and that's stretching
it. To each his own.
*Load temps and power consumption were measured with Prime 95 running on both
cores. Idle temps and power consumption were measured after 15 minutes of inactivity.
Asus PC Probe was used to monitor the temperatures. Testing was done in an open
air environment. Ambient temps were 22C.
*The AMD X2 uses a Zalman CNPS9500 for cooling. The Pentium D 805 uses a Scythe
Ninja 1000 w/Silverstone FN121 for cooling. For those that think the X2 uses an
inferior heatsink, you are wrong from a performance standpoint. They perform the
same with the Zalman at full fan speed, although the combination of the Ninja
and FN121 fan is much quieter and also cheaper.
That's it for the details, let's get it on.
|
|
|
