Overclocking and Performance Scaling

With outstanding performance at stock speed, we pushed the Arctic Cooling HSFs in our performance scaling tests to see if these air-coolers deliver similarly impressive performance when overclocking. As cooling solutions do a better job of keeping the CPU at a lower temperature, it is reasonable to expect the overclocking capabilities of the CPU will increase as cooler performance increases. In each test of a cooler, we measure the highest stable overclock of a standard X6800 processor under the following conditions:

CPU Multiplier: 11x (Stock 11x)
CPU Strap: 266 (stock) or 333 (overclocked)
CPU voltage: 1.5875V at Max overclock
NB Voltage: 1.70V
FSB Voltage: 1.40V
SB Voltage: 1.20V

Memory is set to Auto timings on the P35. This removes memory as any kind of impediment to the maximum stable overclock. As FSB is raised the linked memory speed increases in proportion. For this reason, we maintain memory speed as close as possible to CPU strap settings to prevent memory becoming an issue in CPU testing. This means 3.33GHz uses a 333 strap and runs 10x333, 3.73 uses a 333 strap and runs 11x339, and 3.90 GHz is a 333 strap and 354x11. The stock speed test is a 266 strap and 11x266. We may move in the future to a 3.0GHz stock speed with a 333 strap and 9x333 settings for complete consistency, but since we are not measuring raw computer performance in our cooling benchmarking this becomes a moot point. We use the same processor in all cooling tests to ensure comparable results.

Stable Overclock (MHz)

The Alpine 7 is rated at 90 watts heat dissipation, which means it should overclock a little better than the stock Intel cooler. That is exactly what we found in the overclocking tests, with a top overclock of 3.78GHz compared to the Intel top overclock of 3.73GHz.

The Freezer 7 Pro is rated at 130 watts dissipation and it lives up to that level of performance, topping out at 3.85GHz. This does not challenge our top coolers, which can dissipate more than 150W - or even 165W in the case of the Thermalright Ultra 120 eXtreme. Nonetheless, it is very competent performance for a cooler selling for less than $25. Many other coolers we have tested that sell for much more than these two coolers do not perform as well in overclocking.

Those who expected the Alpine 7 family or Freezer 7 Pro might give our top performing coolers a run for the top will likely be disappointed. However, they shouldn't be disappointed with either cooler's overclocking performance. It is very clear Arctic Cooling knows how to design and rate their coolers for performance. Both the Alpine 7 and the Freezer 7 Pro are fairly rated and outstanding performers in their respective price classes.

Performance Scaling

Performance scaling charts for the Arctic Cooling pair use a scale with zero on the bottom and 75C on the top. Keep in mind that the lower line is the best performance in this presentation of the data - lower temperatures represent better performance.

The Alpine 7 performance scaling at idle is impressive. It significantly outperforms the Intel retail cooler, which would be the closest competitor. Surprisingly it also outperforms most of the mid-range coolers we have tested to the point where it tops out at 3.78GHz. The Freezer 7 Pro is even better, with scaling close to the top guns to its max speed of 3.85GHz.

Comparing cooling efficiency of the Arctic Coolers under load conditions to the retail HSF and other recently retested CPU coolers doesn't change the picture much. Load testing can be very revealing of a cooler's efficiency. A basically flat line, particularly form 3.73GHz upward, indicates the cooler is still in its best cooling range. A rapidly rising line indicates a cooler is nearing the end of its ability to cool efficiently. Lines that parallel the best coolers over a range of values show the coolers provide similar cooling performance.

Under load test conditions, the limited performance of both these coolers is more noticeable. Top coolers we have tested exhibit a shallower slope, where the Alpine 7 and Freezer 7 Pro both have very steep curves in load performance scaling. Results at 2.93GHz and 3.33GHz are among the best test results so far, but the Alpine 7 drops rapidly into the entry/mid performance levels at 3.73GHz and its top of 3.78GHz. Keep in mind that this is a very low-priced cooler and you can get more excited about load scaling, because performance is still significantly better than the stock cooler.

The Freezer 7 Pro fares better under load, as might be expected of a small heatpipe tower. Load performance matches the top tier until around 3.5Ghz. Load scaling then drops into the mid-range area from 3.73GHz to the top for this cooler of 3.85GHz. This is still excellent performance for a cooler in this price range.

Noise Levels Final Words
Comments Locked


View All Comments

  • Etern205 - Tuesday, January 29, 2008 - link

    Sorry for the triple post.

    Somehow the link code does not work so...

  • strikeback03 - Wednesday, January 30, 2008 - link

    Great, an LED fan direct from Intel.

    and it still uses push pins, though since it weighs about the same as the previous stock cooler I guess that is expected.
  • sparkuss - Tuesday, January 29, 2008 - link

    Is there any reason you don't include with the new testbed/database your custom water cooling setup that you've been using to chart the X38/X48?

    I realize it may be extreme but it feels missing if only to show what that "next" level of cooling means in relation to the "top rated performers". I guess I'd also be remiss in not asking for at least one of the new Peltier/Water compact combo coolers in the mix just for those reference lines on the graphs.

    If you only want to limit results to "available/ready to buy" I understand.

    I'm still looking at all options for my next "technology-leap" (AMD 4000+ 939) system build and being able to see if investing in the extreme is worth the results would help with some of the choices.

  • strikeback03 - Wednesday, January 30, 2008 - link

    I'd guess Wesley is in a different part of the country/world than Raja or whoever has the extreme cooling setup.
  • mindless1 - Tuesday, January 29, 2008 - link

    Quote:"low noise, longer-life fan bearings and elastomeric fan mounts. This is expensive engineering."

    These aren't really longer-life bearings. Anyone can make a bearing and claim that in an ideal environment it will have really long life. That's shady marketing, the finished fan, as implemented, is not that ideal environment for several reasons such as thrust level, imbalance, ambient temp. Frankly I find the bearing on my Freezer 64 Pro to be below average compared to my major (fan manufacturer not PC parts relabeler) brands. I would rate them as well if not better than many of the crude fans one would find on cheap heatsinks at least, and the fan imbalance being offset by the rubber mounts does help.

    As for the elastomeric fan mounts, no this is not expensive engineering. Maybe a penny a piece, no more expensive than screws to hold a fan on. Perhaps we could say the unique fan frame design cost a slight bit extra though when in volume the cost may be less than you'd think, particularly if not manufactured by a major label. Upon examination of the fan bearing anyone with a trained eye can easily see these are not premium sleeve bearings by any stretch, and they are a lot short on lubricant, you should expect them not to be so quiet within the life of the system. Relube the bearing periodically for best results.

    While my comments seem (are) negative, overall these coolers are a great value, but we do need to be objective in recognizing the cons as well as the pros. I'm not aware of better value for the money so they are still 'sinks to seriously consider except for attempts at extreme overclocking.
  • Wesley Fink - Tuesday, January 29, 2008 - link

    The Engineering is expensive - not necessarily the parts. What we were trying to say here is that these kinds of solutions are usually reserved for higher-priced coolers, and not often seen on coolers selling for such a low price.

    As for bearing life, most coolers in this price range don't even rate fan bearing life. The expected "life" of the fan on the under $15 Alpine 7 is 400,000 hours and the bearing is a Fluid Dynamic Bearing - like the Scythe Sflex 120mm fan which is $20 for the fan alone. These are both impressive specs for any cooler fan - especially one that sells with the complete cooler for less than $15.
  • mindless1 - Tuesday, January 29, 2008 - link

    The life rating of 400K hours is nonsense.

    Their typical fan sells for $6, and that with a bit of profit built in. Their bearings are not special, just the marketing is.


    A minimum price is a better proof than an inflated one, as even the generic junk out there selling for $2 is also marked up 250% or more through relabelers.
  • forgotmypassword - Tuesday, January 29, 2008 - link

    Your under load test is WAY TOO GENTLE. Same AC Freezer 7 can hardly keep my E6550 @ 2.8GHz Core Duo under 75C under 100% load... Compare it to your 41C
  • RamarC - Tuesday, January 29, 2008 - link

    if your e6550 is hitting 75c you either don't have the freezer 7 mounted properly or your bios/fan control is configured to allow that level of heat.
  • mindless1 - Tuesday, January 29, 2008 - link

    You have no evidence of that. A higher ambient temp, worse case ventilation, and/or higher vcore can cause this.

    However, some mountings don't seem to put as much pressure on the 'sink, it can be mounted as "properly" as possible and still this (and especially along with a combination of aforementioned factors) could result in that temp.

    The real question is WHY someone would have allowed their CPU to get this hot instead of reducing the o'c or taking whatever other measures are necessary.

Log in

Don't have an account? Sign up now