Cooler Master GeminII: Performer or Poser?
by Wesley Fink on April 30, 2007 2:00 AM EST- Posted in
- Cases/Cooling/PSUs
Scaling of Cooling Performance
While the noise levels varied widely among the four fan configurations tested with the Cooler Master GeminII, the actual cooling performance was very similar across all fan configurations. For that reason, scaling performance will only be charted with the Noctua dual fan configuration, which exhibited the lowest noise with cooling efficiency comparable to the high-output fan configurations. The lower the line for a cooler (representing lower temperatures) the more efficient that cooler is at cooling.
At 2.93GHz the retail HSF is running at 41C, compared to 30C with a typical GeminII dual-fan configuration. This is not the best performance seen in our tests at idle, but it is competitive with the top coolers tested so far. This is a delta of 11C. The delta becomes greater as the overclock increases. At 3.73GHz the idle with the retail fan is 56C compared to the Gemini II at 39C - a delta of 17C. The cooling performance of the Gemini II is much better than the Intel retail cooler at idle, but the Gemini II does not reach the same cooling levels seen in the Thermalright coolers or the Tuniq 120.
Cooling efficiency of the GeminII under load conditions was compared to the retail HSF and other recently tested CPU coolers. Load testing can be very revealing of a cooler's efficiency. A basically flat line, particularly from 3.73GHz upward, indicates the cooler is still in its best cooling range. A line that is increasing rapidly indicates a cooler nearing the end of its ability to cool efficiently. Lines which parallel the best coolers over a range of values indicate similar efficiency slopes.
The GeminII is very efficient in cooling in the 2.93GHz to 3.73GHz overclock range. It continues on to a highest overclock of only 3.83GHz, which is well short of the 3.90GHz to 3.94GHz most of the top coolers in our testing have achieved. However, the sharp slope of the line from 3.73GHz to 3.83GHz indicates the GeminII is rapidly losing its ability to cool effectively at these higher overclocks. This is surprising, considerably the huge size of the GeminII combined with dual fans. Basically the Gemini II is performing no better than some of the $20 to $30 coolers we have tested.
Compare the GeminII, for example, to the lines for the Thermalright Ultra 120 Extreme and the Monsoon II Lite hybrid air/TEC cooler. Both these coolers have very flat stress performance lines as the overclock rises above 3.73GHz, which suggests they are still efficiently cooling and the CPU has run out of overclock headroom rather than the cooler reaching the limit of what it can effectively cool.
As stated many times, the overclocking abilities of the CPU will vary at the top, depending on the CPU. This particular CPU does higher FSB speeds than any X6800 we have tested, but the 3.9GHz top speed with the top tier heatpipe towers is pretty average among the X6800 processors we have tested with Tuniq cooling. A few of the other processors tested with the best air coolers reach just over 4 GHz, but the range has been 3.8 to 4.0GHz. Stock cooling generally tops out 200 to 400 MHz lower, depending on the CPU, on the processors tested in our lab. The 3.83 GHz with the Gemini II is average at best; we would expect a premium-priced CPU cooler to perform better.
While the noise levels varied widely among the four fan configurations tested with the Cooler Master GeminII, the actual cooling performance was very similar across all fan configurations. For that reason, scaling performance will only be charted with the Noctua dual fan configuration, which exhibited the lowest noise with cooling efficiency comparable to the high-output fan configurations. The lower the line for a cooler (representing lower temperatures) the more efficient that cooler is at cooling.
 |
| Click to enlarge |
At 2.93GHz the retail HSF is running at 41C, compared to 30C with a typical GeminII dual-fan configuration. This is not the best performance seen in our tests at idle, but it is competitive with the top coolers tested so far. This is a delta of 11C. The delta becomes greater as the overclock increases. At 3.73GHz the idle with the retail fan is 56C compared to the Gemini II at 39C - a delta of 17C. The cooling performance of the Gemini II is much better than the Intel retail cooler at idle, but the Gemini II does not reach the same cooling levels seen in the Thermalright coolers or the Tuniq 120.
Cooling efficiency of the GeminII under load conditions was compared to the retail HSF and other recently tested CPU coolers. Load testing can be very revealing of a cooler's efficiency. A basically flat line, particularly from 3.73GHz upward, indicates the cooler is still in its best cooling range. A line that is increasing rapidly indicates a cooler nearing the end of its ability to cool efficiently. Lines which parallel the best coolers over a range of values indicate similar efficiency slopes.
 |
| Click to enlarge |
The GeminII is very efficient in cooling in the 2.93GHz to 3.73GHz overclock range. It continues on to a highest overclock of only 3.83GHz, which is well short of the 3.90GHz to 3.94GHz most of the top coolers in our testing have achieved. However, the sharp slope of the line from 3.73GHz to 3.83GHz indicates the GeminII is rapidly losing its ability to cool effectively at these higher overclocks. This is surprising, considerably the huge size of the GeminII combined with dual fans. Basically the Gemini II is performing no better than some of the $20 to $30 coolers we have tested.
Compare the GeminII, for example, to the lines for the Thermalright Ultra 120 Extreme and the Monsoon II Lite hybrid air/TEC cooler. Both these coolers have very flat stress performance lines as the overclock rises above 3.73GHz, which suggests they are still efficiently cooling and the CPU has run out of overclock headroom rather than the cooler reaching the limit of what it can effectively cool.
As stated many times, the overclocking abilities of the CPU will vary at the top, depending on the CPU. This particular CPU does higher FSB speeds than any X6800 we have tested, but the 3.9GHz top speed with the top tier heatpipe towers is pretty average among the X6800 processors we have tested with Tuniq cooling. A few of the other processors tested with the best air coolers reach just over 4 GHz, but the range has been 3.8 to 4.0GHz. Stock cooling generally tops out 200 to 400 MHz lower, depending on the CPU, on the processors tested in our lab. The 3.83 GHz with the Gemini II is average at best; we would expect a premium-priced CPU cooler to perform better.
Facebook
Twitter
RSS
44 Comments
View All Comments
yyrkoon - Wednesday, May 2, 2007 - link
err, make that the second to the last page. Sorry.dm - Monday, April 30, 2007 - link
Nice review Wesley. I have owned this cooler and it does have its own merit. While it failed to beat the competition, what I can share is that I used this fan on an overclocked QX6700@3.2GHz at stock vCore and it performed great. If anyone would like to use it to achieve 3.8GHz+++, I suggest they get a quad core instead, say an X3210 and overclock it to 3GHz and use this cooler with a silent fan. With such cooling power, and since I have tested it personally, this cooler is great and should perform at par with competing ones.My only gripe is the price. The Ninja RevB I owned can cool the same quad core at same voltage but with much lower price tag.
joex444 - Monday, April 30, 2007 - link
This thing just got pwned. It's obviously a poser. All the extra fin space did nothing, and a 2nd fan proved to be a novelty.LaGUNaMAN - Wednesday, May 2, 2007 - link
Thanks for the review. I was waiting for this a long time. Definitely the best Gemini II review out there. (^^,)evident - Monday, April 30, 2007 - link
is it me or does thermalright xp/ultra/whatever 120 series own the competition constantly? every review i read this cooler dominates the market. I'm happy to own an xp-120 and that it's not getting topped anytime soon it seems :)stromgald - Monday, April 30, 2007 - link
Um, the XP-120 has been topped by the Ultra-120 (which is a completely different design than the XP), Scythe Ninja, and Zalman CNPS9500 to name a few. The Ninja barely edges it out, and the Ultra-120 beats the XP-120 by a good margin. The Zalman beats it in °C rise per W of processor heat, but is also noisier. Thermalright's good, but it's not like the competition is that far off at all.P.S. Thermalright doesn't have a 120 'series'. It's more like the XP series (XP-120, XP-90), SI series (SI-128, SI-97), and Ultra Series (Ultra-120, Ultra-90). The number is related to fan size, the name is the overall design/series.
Wesley Fink - Monday, April 30, 2007 - link
Thermalright has advised that the Ultra-120 eXtreme issues with the Socket 775 adapter have been resolved and the eXtrme is now in full production. For those who asked, the IFX-14 production has been delayed and Thermalright is now looking at an end of May release for the IFX-14.We will do an update on the performance of the production Ultra-120 eXtreme which is on its way to the labs.
TA152H - Monday, April 30, 2007 - link
Not to state the obvious, but shouldn't you have tested this thing with no fans? I mean, you get a heatsink with no fans at all, that says it can run with no fans, and kind of sucks when used with fans. Hmmmm, would kind of hit me like an airborne brick that I should probably test it without any fans at all, and see if it works. Cooler Master in my experience has been a pretty good company, and it makes me wonder why they made a cooler like this one that sucks so bad. Except maybe it doesn't, and for whatever peculiar reason is exceptional at running with no fans vis-a-vis other heat sinks with no fans. Of course, it could suck at that as well, but it's so obvious a question, I'm surprised no attempt was made to answer it. It could also be potentially useful. A sucky cooler like this is of little or no interest to people with fans, but if it cooled as well as a stock heat sink/fan, without the fan, that would be attractive to a much greater audience.I'm also surprised they are still making these massive beasts. With the death of the horrible P7, you'd think they would start making more reasonable sized heat sinks - unless they needed it this big to work fanless.
Wesley Fink - Monday, April 30, 2007 - link
We did try to test with no fans, but temperatures immediately ran to 70C under load at stock X6800 speeds and the system rebooted. I was not comfortable continuing a "no-fan" test under those conditions. perhaps this should have been mentioned in the review, but the results already pointed to less than stellar cooling efficiency compared to the Thermalrights, for example, which actually did fanless cooling within acceptable temperature ranges.As we have also stated several times, our current test bed is not well setup for testing fanless coolers. THe new test bed, which will be introduced with the 120mm fan roundup, is much better in that regard and will better support reliable, comparable results with fanless cooling.
TA152H - Monday, April 30, 2007 - link
OK, thanks.I think you should mention that in the article, because that was the first thing that popped into my head. "Can I use this thing fanless at stock clock speeds?"
With regards to the poor cooling, with some stuff it's bad to extrapolate data, even if it's logical. More specifically, it could be possible that one did better with fans, and another without, so I resist making assumptions on stuff like that. Put another way, if you were a designer, and I told you to make two heat sinks, one that was designed to work with a fan, and one that was designed to work without, would they end up being different? I would think so, but really I don't know enough to be sure. Consequently, I think stuff like that should be spelled out instead of assumed. In this case, it sucked at everything, but that may not always be the case.
Anyway, thanks for your response. I wouldn't want to ruin my processor either, so I can understand why you'd stop testing right away. I would have done the same.