Review: The Best Air Coolers Face to Face!
- Thread starter Unkno
- Start date
yea...guess they are only comparing the best recent air coolers...oh well, i would like to have seen the si-120 too.
btw, did you notice all of the coolers are roughly the same size (huge), maybe that's why...lol
Well, it's not as big or as tall as these coolers, anyways, the xp-120 is a great cooler and you should stick with it instead of upgrading, unless you want silent, don't upgrade from a xp-120 to a ninja
I should be sorry to interrupt but can't hold back.
Well, guess i'll just throw-out the XP90C that should arrive Monday. Suprising how fast everything turns into junk. Gotta kick up the buying pace. The market demands it, don'tchano. ;-)
Sick joke: With my window build a skyscraper HS would look cool with a tiny 747 sticking
into the fin area near the top.
Well, guess i'll just throw-out the XP90C that should arrive Monday. Suprising how fast everything turns into junk. Gotta kick up the buying pace. The market demands it, don'tchano. ;-)
Sick joke: With my window build a skyscraper HS would look cool with a tiny 747 sticking
into the fin area near the top.
JEDIYoda
Lifer
- Jul 13, 2005
- 33,986
- 3,320
- 126
Also just for the record the I own the XP120 and the zalman 9500!!
As of this post the temps for both those coolers were approx the same--almost identicle!!
Also as I have stated many many times...
The Scythe Ninja is the current heavyweight Champion of the world!!
As of this post the temps for both those coolers were approx the same--almost identicle!!
Also as I have stated many many times...
The Scythe Ninja is the current heavyweight Champion of the world!!
you also have the ninja? According to this thread it's about the same as the 9500, which is near the same as the xp120...well, it all really depends on the case and other components such as fans.
BonzaiDuck
Lifer
- Jun 30, 2004
- 16,116
- 1,727
- 126
Y'all may tire of it, but I wanna dish it out anyway.
The reviewer's approach was to OC a Prescott 2.8 to 4.06. I'm not quite sure how he does this -- he probably runs the memory on a divider, but then again, the test-bed includes the latest generation of mobos and uses DDR2 modules.
But to the point . . . . his purpose in OC'ing the testbed was to run up the thermal power to around 130 watts. You could still test these things without OC'ing the processor, or by OC'ing it less -- knowing what the thermal power would be under those less strenuous circumstances.
This presents a slight predicament for me, because I haven't yet experimented with my socket-478 3.2E Prescott to push it beyond 3.5 Ghz and FSB at 1,000 Mhz. Or -- for that matter -- 3.61 Ghz and FSB at about 892 Mhz. So it might look -- or seem -- as though I'm not submitting the equipment to the same stress as the X-Bit reviewer. Yet, I can't find any other logic that makes sense -- you could test it with lower thermal power or higher thermal power, but the idle and load values in conjunction with the thermal wattage would accurately rate and rank-order the coolers the same way.
As I said in other posts, my idle-load spread with an SI-120 on my Socket-478 Prescott is about 5.5C degrees. If the thermal power is only what INtel reports on their web-site -- 103 watts -- and doesn't vary that much from a slightly higher thermal power of my moderately over-clocked processor, this represents an extremely -- ridiculously -- low thermal resistance -- approximately 0.05 C/W. It hardly makes sense. But it's "real." It may have something to do with my "sinking" the Mosfets and PLLs, or with my motherboard duct. But my test with the XP-120 was entirely consistent with the review-reports of its thermal resistance. My tests with the SI-120 -- I've already explained. I'm insanely happy with it, but what gives here?
If the thermal power were higher, the thermal resistance would be even lower -- and seem even more ridiculous. Various reviews I've read for the Prescott I'm using -- keep in mind that there are probably some thermal differences between the socket-478 version and the LGA775 version -- show users who didn't over-clock getting idle values in the mid-40's (C) and load values well above 50C. Yet at 70F (21C) room temperature, my idle value is about 90F or 32C and my load value is 100F or 38C. These measurements were taken running PRIME95 Large FFT "extreme heat, stress. . . ." test. Maybe I'd see something else with S&M, but I doubt it would matter that much.
If we simply accept the imperatives of thermal resistance, we can find other comparisons between the reviewed coolers and the SI-120. For instance, we discovered that the Scythe Ninja with a fan running around 950 rpm exhibited a thermal resistance of about 0.19 C/W while the SI-120 with fan running at about the same speed showed 0.18 C/W. This would indicate an advantage for the SI-120 -- even a slight one -- qualified only by the choice of the "quiet" fan speed.
My real question is why -- in a review so recent -- did they exclude the SI-120? Everyone knows about ThermalRight; the XP-120 was a big hit last year -- that's why I've got one of those in another machine now -- a legacy from a year ago. It's not as though the company is "obscure" or nobody has heard of the product.
As for the other ThermalRight models -- like the XP90-C -- that's still a pretty damn good cooler, and it was something of a precursor for the SI-120, except that it's made entirely of copper -- correct me if I'm wrong. The TR for the XP90-C is something between 0.15 and 0.16 C/W. Just a shade less effective than the SI-120.
But to reiterate earlier insinuations -- it doesn't much matter if you over-clock the test-bed. Lower clock-speed -- lower thermal power; higher clock-speed -- higher thermal power. If you know the thermal power and you measure idle and load temperatures, you still have your index of performance. It won't change between a test-bed run at stock speed and a test-bed that is over-clocked.
The reviewer's approach was to OC a Prescott 2.8 to 4.06. I'm not quite sure how he does this -- he probably runs the memory on a divider, but then again, the test-bed includes the latest generation of mobos and uses DDR2 modules.
But to the point . . . . his purpose in OC'ing the testbed was to run up the thermal power to around 130 watts. You could still test these things without OC'ing the processor, or by OC'ing it less -- knowing what the thermal power would be under those less strenuous circumstances.
This presents a slight predicament for me, because I haven't yet experimented with my socket-478 3.2E Prescott to push it beyond 3.5 Ghz and FSB at 1,000 Mhz. Or -- for that matter -- 3.61 Ghz and FSB at about 892 Mhz. So it might look -- or seem -- as though I'm not submitting the equipment to the same stress as the X-Bit reviewer. Yet, I can't find any other logic that makes sense -- you could test it with lower thermal power or higher thermal power, but the idle and load values in conjunction with the thermal wattage would accurately rate and rank-order the coolers the same way.
As I said in other posts, my idle-load spread with an SI-120 on my Socket-478 Prescott is about 5.5C degrees. If the thermal power is only what INtel reports on their web-site -- 103 watts -- and doesn't vary that much from a slightly higher thermal power of my moderately over-clocked processor, this represents an extremely -- ridiculously -- low thermal resistance -- approximately 0.05 C/W. It hardly makes sense. But it's "real." It may have something to do with my "sinking" the Mosfets and PLLs, or with my motherboard duct. But my test with the XP-120 was entirely consistent with the review-reports of its thermal resistance. My tests with the SI-120 -- I've already explained. I'm insanely happy with it, but what gives here?
If the thermal power were higher, the thermal resistance would be even lower -- and seem even more ridiculous. Various reviews I've read for the Prescott I'm using -- keep in mind that there are probably some thermal differences between the socket-478 version and the LGA775 version -- show users who didn't over-clock getting idle values in the mid-40's (C) and load values well above 50C. Yet at 70F (21C) room temperature, my idle value is about 90F or 32C and my load value is 100F or 38C. These measurements were taken running PRIME95 Large FFT "extreme heat, stress. . . ." test. Maybe I'd see something else with S&M, but I doubt it would matter that much.
If we simply accept the imperatives of thermal resistance, we can find other comparisons between the reviewed coolers and the SI-120. For instance, we discovered that the Scythe Ninja with a fan running around 950 rpm exhibited a thermal resistance of about 0.19 C/W while the SI-120 with fan running at about the same speed showed 0.18 C/W. This would indicate an advantage for the SI-120 -- even a slight one -- qualified only by the choice of the "quiet" fan speed.
My real question is why -- in a review so recent -- did they exclude the SI-120? Everyone knows about ThermalRight; the XP-120 was a big hit last year -- that's why I've got one of those in another machine now -- a legacy from a year ago. It's not as though the company is "obscure" or nobody has heard of the product.
As for the other ThermalRight models -- like the XP90-C -- that's still a pretty damn good cooler, and it was something of a precursor for the SI-120, except that it's made entirely of copper -- correct me if I'm wrong. The TR for the XP90-C is something between 0.15 and 0.16 C/W. Just a shade less effective than the SI-120.
But to reiterate earlier insinuations -- it doesn't much matter if you over-clock the test-bed. Lower clock-speed -- lower thermal power; higher clock-speed -- higher thermal power. If you know the thermal power and you measure idle and load temperatures, you still have your index of performance. It won't change between a test-bed run at stock speed and a test-bed that is over-clocked.
JEDIYoda
Lifer
- Jul 13, 2005
- 33,986
- 3,320
- 126
okay I will say this only once..hehee...I agree!!
No I don`t have the Ninja....
But I have never seen a bad review or heard a discouraging word!!
I know a few reviews where they placed the Ninja above the Big Typhoon and other heatsinks fo shere low temps!!
Sure all the temps were within a few Centigrade of each other but the Ninja still an impressive heatsink!!
No I don`t have the Ninja....
But I have never seen a bad review or heard a discouraging word!!
I know a few reviews where they placed the Ninja above the Big Typhoon and other heatsinks fo shere low temps!!
Sure all the temps were within a few Centigrade of each other but the Ninja still an impressive heatsink!!
BonzaiDuck
Lifer
- Jun 30, 2004
- 16,116
- 1,727
- 126
I found this article by Citarella -- same guy who reviewed the SI-120 at OverClockers.com in August to show that it's thermal resistance or "thermal efficiency" was 0.14 C/W. I still have to finish reading it, but I would say there is much revealing and useful information here:
OverClocking's impact on CPU life
The article turned up in a web-query to find a download site for the S & M test program. I want to see if PRIME95 is really adequate to the task anymore, and I'm willing to give it a shot.
But the query didn't get me what I was looking for. Anyone know where I can get it?
OverClocking's impact on CPU life
The article turned up in a web-query to find a download site for the S & M test program. I want to see if PRIME95 is really adequate to the task anymore, and I'm willing to give it a shot.
But the query didn't get me what I was looking for. Anyone know where I can get it?
BonzaiDuck
Lifer
- Jun 30, 2004
- 16,116
- 1,727
- 126
OK. I found it. It was really at the top of the list, but I had overlooked it:
S&M Beta version 1.7.5
It's a Russian program. First, let me say this in respect for the Russian people, and my friend Yulia, whom I've taken steps to recommend for a graduate scholarship at an American school. This in no way is meant as a "slight" or insult.
When you do research on the Kennedy assassination, you study intelligence history. What you learn is that the KGB were . . . . if not "pussies" . . . . then "conservative." The real bad-a** M*****F*****s were the CIA "grand-standers" stationed in South Florida's Zenith Technologies Building. Who killed Jack? THEY DID!!
And so, I have to pass judgment on the impact of this program -- S&M -- on my system. It is a wonderful, extremely elegant program. You can set CPU usage to any level you want before running it. It has these wonderful little real-time graphs. It does memory tests, and it loads up the CPU under hyperthreading to whatever level you set it.
I set mine first to 75%, and then to 95%. At the 95% level, it fluctuates between 90 and 100%.
And my system? It's a bad-a** m*****f*****!!!! At a room-ambient of 67F, the CPU load temperature doesn't even make it to 100F or 38C.
Thank you, ThermalRight!!
S&M Beta version 1.7.5
It's a Russian program. First, let me say this in respect for the Russian people, and my friend Yulia, whom I've taken steps to recommend for a graduate scholarship at an American school. This in no way is meant as a "slight" or insult.
When you do research on the Kennedy assassination, you study intelligence history. What you learn is that the KGB were . . . . if not "pussies" . . . . then "conservative." The real bad-a** M*****F*****s were the CIA "grand-standers" stationed in South Florida's Zenith Technologies Building. Who killed Jack? THEY DID!!
And so, I have to pass judgment on the impact of this program -- S&M -- on my system. It is a wonderful, extremely elegant program. You can set CPU usage to any level you want before running it. It has these wonderful little real-time graphs. It does memory tests, and it loads up the CPU under hyperthreading to whatever level you set it.
I set mine first to 75%, and then to 95%. At the 95% level, it fluctuates between 90 and 100%.
And my system? It's a bad-a** m*****f*****!!!! At a room-ambient of 67F, the CPU load temperature doesn't even make it to 100F or 38C.
Thank you, ThermalRight!!
BonzaiDuck
Lifer
- Jun 30, 2004
- 16,116
- 1,727
- 126
I was always reticent about accessing sites with the suffix ".ru"
Maybe this is the knee-jerk conditioning I have growing up during the Cold-War era. You worry that if you download something from such a site, that a worm, virus or trojan-horse will sneak into your computer, paint the screen Red with a hammer and sickle, and suck your precious bodily fluids dry.
:laugh:
Maybe this is the knee-jerk conditioning I have growing up during the Cold-War era. You worry that if you download something from such a site, that a worm, virus or trojan-horse will sneak into your computer, paint the screen Red with a hammer and sickle, and suck your precious bodily fluids dry.
:laugh:
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 19K
-
-
-
Facebook
Twitter