Zalman VF1000 LED Graphics Card Cooler
-
- Patron of SPCR
- Posts: 169
- Joined: Wed Sep 21, 2005 5:13 pm
- Location: South FL
I used a VF700 on my 7800GTX, and was happy with the stock cooler on my current 7900GTX for a long time.
That said, the Accelero S1, for $20, with a ~750RPM 120mm Yate-Loon strapped onto it, keeps the VGA card at 60 C tops at full load at 1680x1050. I'd really only consider this new Zalman for an HD2900 or 8800 series card.
That said, the Accelero S1, for $20, with a ~750RPM 120mm Yate-Loon strapped onto it, keeps the VGA card at 60 C tops at full load at 1680x1050. I'd really only consider this new Zalman for an HD2900 or 8800 series card.
I think that this sentence is not properly constructed. Not that my english is perfect, but mixing 5V and 12V in the same sentence with the single "18º". It may need to add a comma or rephrase it.Even at 5V, the VF1000 walks all over its predecessor at 12V by a staggering 18°C.
Of course I could take a look at the table, which is what I did.
-
- Friend of SPCR
- Posts: 34
- Joined: Sat Oct 16, 2004 3:21 pm
- Location: Socorro, New Mexico, USA
vf900 ok?
Seems like the results for the vf900 are inconsistant with earlier tests - it was a great cooler in last year's test.
I tried to calculate the difference in GPU power consumption - the net difference is ~25-30W but the CPU is different from the previous test. The rated TDP is ~10W higher in the new test rig processor.
Did you try remounting the vf900 to see if there was a thermal contact problem?
I'm curious here...
I tried to calculate the difference in GPU power consumption - the net difference is ~25-30W but the CPU is different from the previous test. The rated TDP is ~10W higher in the new test rig processor.
Did you try remounting the vf900 to see if there was a thermal contact problem?
I'm curious here...
I guess I have a little nitpicking with the article...mostly because it reminded me of other websites in some of the procedural approaches.
Specifically, when it comes to the "nah, we'll leave this stock thing on, and we'll bend this that way..." and then eventually the next review comes along and "we took all the stock hardware off, and followed the directions exactly..."
Leads to Red Delicous to Granny Smith comparisons Somewhat negates the point of the common test platform in my opinion I guess.
Just made me feel like I was reading an AT review for a moment. And that's not, you know, a *good* feeling
Anyways...nice to see all the reviews coming out
Why haven't Zalman just given up and started selling coolers with Nexus (or Yate Loon or anyone else quiet) fans yet?!?!
Specifically, when it comes to the "nah, we'll leave this stock thing on, and we'll bend this that way..." and then eventually the next review comes along and "we took all the stock hardware off, and followed the directions exactly..."
Leads to Red Delicous to Granny Smith comparisons Somewhat negates the point of the common test platform in my opinion I guess.
Just made me feel like I was reading an AT review for a moment. And that's not, you know, a *good* feeling
Anyways...nice to see all the reviews coming out
Why haven't Zalman just given up and started selling coolers with Nexus (or Yate Loon or anyone else quiet) fans yet?!?!
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
psiu --
Not sure what you're referring to, I didn't see anything inconsistent in our procedures. The same test case/system was used. The CPU is not identical, but the total power delivered via the AUX12V is just 6W higher than with the previous system -- 85W vs 91W. And this does not really play any part in our video card's cooling or power.
The graphics card is not the same one used in last year's GPU cooler reviews. Yes, it's been that long since we did a GPU cooler review. The 6800GT card died finally after > a years abuse in the lab. The 1950XTX seems like a good replacement because...
1) Its thermal envelope is much closer to the hottest of today's cards
2) It will thus push today's bigger & better aftermarket heatsinks more effectively; with a cooler GPU, the differences between coolers would be more obscured.
3) We happen to have one in the lab.
We didn't see the point of making a big deal of the change in vidcards. No you can't do a/b comparisons of the cooling performance of this HS against all the others from last year.... but why would you want or need to? It's obviously way better.
And no, there was no issue with the VF900 being poorly mounted -- the mounting system is identical to that of the VF1000, and it's very difficult to do it wrong -- even if you haven't done dozens of vidcard cooler installs/uninstalls.
Bending the thing a bit was a snap decision and its implementation took a couple seconds. Larry said, hey this touches, and I said, let's see, and after one quick look, gave it a firm gentle nudge; it moved maybe 1/8' to 1/4". Problem solved. If you think that's an unacceptable procedure for a review of interest only to DIYers, well... think again!
Not sure what you're referring to, I didn't see anything inconsistent in our procedures. The same test case/system was used. The CPU is not identical, but the total power delivered via the AUX12V is just 6W higher than with the previous system -- 85W vs 91W. And this does not really play any part in our video card's cooling or power.
The graphics card is not the same one used in last year's GPU cooler reviews. Yes, it's been that long since we did a GPU cooler review. The 6800GT card died finally after > a years abuse in the lab. The 1950XTX seems like a good replacement because...
1) Its thermal envelope is much closer to the hottest of today's cards
2) It will thus push today's bigger & better aftermarket heatsinks more effectively; with a cooler GPU, the differences between coolers would be more obscured.
3) We happen to have one in the lab.
We didn't see the point of making a big deal of the change in vidcards. No you can't do a/b comparisons of the cooling performance of this HS against all the others from last year.... but why would you want or need to? It's obviously way better.
And no, there was no issue with the VF900 being poorly mounted -- the mounting system is identical to that of the VF1000, and it's very difficult to do it wrong -- even if you haven't done dozens of vidcard cooler installs/uninstalls.
Bending the thing a bit was a snap decision and its implementation took a couple seconds. Larry said, hey this touches, and I said, let's see, and after one quick look, gave it a firm gentle nudge; it moved maybe 1/8' to 1/4". Problem solved. If you think that's an unacceptable procedure for a review of interest only to DIYers, well... think again!
Last edited by MikeC on Tue Nov 06, 2007 10:03 am, edited 2 times in total.
Indeed, one would think so. Mike, how about some informal (or formal, whatever) testing of case temps. Find a couple of GFX cards that exhaust hot air out the back and compare the case temps to those same cards with the Zalman VF1000. I'd be keen to find out how much it really helps when the GFX card exhausts all its hot air externally. Say, the X1950XTX and something from the GeForce 8800 GTS/GTX line, in an Antec P180/182 case?MoJo wrote:Needs a shroud to exhaust air out the back.
The problem is the fan is in the wrong place. It will blow air back into the case. It needs to be at one end, so it sucks air from inside the case and blows it out.Sandwich wrote:Indeed, one would think so. Mike, how about some informal (or formal, whatever) testing of case temps. Find a couple of GFX cards that exhaust hot air out the back and compare the case temps to those same cards with the Zalman VF1000. I'd be keen to find out how much it really helps when the GFX card exhausts all its hot air externally. Say, the X1950XTX and something from the GeForce 8800 GTS/GTX line, in an Antec P180/182 case?
Maybe some sort of DIY effort could be made.
To any aftermarket heatsink manufacturers out there, please consider the idea floated by several people (including myself) on SPCR. A three-slot cooling solution, similar to the stock cooler but taller. Fan at the rear end of the card (furthest inside the case), preferably a standard 80/92mm unit but use your own if you must. Shrouded so all hot air goes out the back. The fan would probably have to be at an angle, or could just be flat and extend down a bit (still wouldn't be as tall as a HR-03 I think).
That's just my point - we know it has to make some sort of difference, but let's find out just how much of a difference it makes.MoJo wrote: The problem is the fan is in the wrong place. It will blow air back into the case. It needs to be at one end, so it sucks air from inside the case and blows it out.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
There's no easy way to duct this style of fan/VGA cooler to exhaust out o the case. The results would be entirely specific to the particular system and setup, as well; it wouldn't be useful in a broad way. Most people in this forum know that the total thermal load on the system is reflected by its power consumption, which does not change with VGA cooling. The question you ask is not about this VGA cooler at all, but about system cooling.Sandwich wrote:That's just my point - we know it has to make some sort of difference, but let's find out just how much of a difference it makes.MoJo wrote: The problem is the fan is in the wrong place. It will blow air back into the case. It needs to be at one end, so it sucks air from inside the case and blows it out.
Certainly, a VGA cooler that exhausts out of the case like the Arctic Cooling products of old and the stock cooler on our test video card should help the overall thermal load on the system, but our test results shows just +3C worst-case change in CPU temperature when the VF1000 is used. This suggests that with efficient enough transfer of the GPU heat into the surrounding air, in our test system, the 120mm case exhaust fan at 7V provides enough heat exhaust.
Why, in our case, would we want to bother with a duct for the VGA? There would be no significant benefit.
On the other hand, if you made a case/VGA card mod to make such a duct, I'm sure lots of folks here would be interested to know about it. But imo, you'd be better off to start with a fanless cooler like the VNF100 or AC Accelero S1/2.
I found that, when using a HR-03 Plus on an 8800 Ultra, with the CPU idle and GPU running ATITool the CPU temperature rose by 7-8C compared to the stock cooler with enhanced shroud. That's in a P182 with fans at 10V, and an extra Kama Bay added.
Perhaps you only get a +3C difference because your test case had better airflow than a P182, and your unmodified stock cooler would still have been throwing a fair bit of heat into the case. I use a paper shroud and electrical tape to ensure all hot air goes out the back.
In a high end system where the CPU is already running close to acceptable thermal limits, an 8C rise in temps (maybe more under load, unfortunately I didn't test that condition) is very significant.
Perhaps you only get a +3C difference because your test case had better airflow than a P182, and your unmodified stock cooler would still have been throwing a fair bit of heat into the case. I use a paper shroud and electrical tape to ensure all hot air goes out the back.
In a high end system where the CPU is already running close to acceptable thermal limits, an 8C rise in temps (maybe more under load, unfortunately I didn't test that condition) is very significant.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Your post strengthens my point -- this kind of analysis is not about the VGA cooler, it's about specific system thermals.MoJo wrote:...In a high end system where the CPU is already running close to acceptable thermal limits, an 8C rise in temps (maybe more under load, unfortunately I didn't test that condition) is very significant.
Indeed, hence why I said (note the bolded text):MikeC wrote:Your post strengthens my point -- this kind of analysis is not about the VGA cooler, it's about specific system thermals.
Sandwich wrote:Mike, how about some informal (or formal, whatever) testing of case temps. Find a couple of GFX cards that exhaust hot air out the back and compare the case temps to those same cards with the Zalman VF1000. I'd be keen to find out how much it really helps when the GFX card exhausts all its hot air externally. Say, the X1950XTX and something from the GeForce 8800 GTS/GTX line, in an Antec P180/182 case?
Besides difference in case temperatures, there's got to be a difference in GFX card cooler performance between ducted and non-ducted coolers, even if the presence or absence of the duct were the only difference. After all, a non-ducted cooler will heat up case temperature more than a ducted cooler, thus increasing the "base" temperature of the air flowing into the GFX cooler in the first place.
-
- Posts: 51
- Joined: Thu Aug 09, 2007 8:28 pm
- Location: Burbank, CA
Surely system thermals are incredibly important though, especially on SPCR where the main battle now is heat management.MikeC wrote:Your post strengthens my point -- this kind of analysis is not about the VGA cooler, it's about specific system thermals.MoJo wrote:...In a high end system where the CPU is already running close to acceptable thermal limits, an 8C rise in temps (maybe more under load, unfortunately I didn't test that condition) is very significant.
The P18x is a very popular case, and although it is particularly restricted a lot of people are running cases with fairly low airflow in order to reduce noise. Anyone looking to buy this cooler must consider the impact it may have on their system as a whole.
I can see your argument that the system you tested is a reasonable silent system to use as a baseline, but to be honest such systems are not really of much concern any more. Cases, fans and heatsinks and advanced to the point where it's fairly easy to assemble a silent, cool system now. The difficulty comes with higher powered systems, and so I would suggest that pushing review hardware to the limits might be a worthwhile exercise.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Really? That's news to me and probably to 99% of PC users in the world who have no idea how quiet a PC can be. Your knowledge after reading SPCR and hanging in these forums for nearly 5 years is hardly reflective of the general PC user population.MoJo wrote:I can see your argument that the system you tested is a reasonable silent system to use as a baseline, but to be honest such systems are not really of much concern any more. Cases, fans and heatsinks and advanced to the point where it's fairly easy to assemble a silent, cool system now. The difficulty comes with higher powered systems, and so I would suggest that pushing review hardware to the limits might be a worthwhile exercise.
Anyway, I beg to differ that our system is low power, which you suggest. The main power suckers in the system --
Intel D950 processor -- TDP 130W: Very few current gen processors pull this kind of power. A C2D E6850 is rated for 65W TDP; the vast majority of Intel CPUs are at this level. The exceptions are invariably more expensive -- a C2D 6600 is rated at 95W, and the Qx6850 and upcoming QX9650 are 130W. Among AMD processors, 65~89W is the norm, and none are rated higher than 125W.
ATI X1950XTX -- real power draw ~125W: The 3rd graph on this Xbit-labs page compares the DC power draw of this vidcard against various 8800 models, including a GTX purportedly to be more power than an Ultra. At 125W, it still looks to be one of the the most power hungry vidcards around.
OK, so it doesn't have 4 Raptor drives or 4gb of RAM, but our Antec NeoPower PSU is less efficient than most these days -- it never gets to 80%. All in all, it's NOT a low power system. This system is still cooled with just one slow spinning 120mm fan. If people are having trouble make their gaming systems quiet, they need to look at what they're doing different from this system.
PS -- OK, we also don't have a second vidcard, but I'll be damned if we ever review a dual vidcard system. That concept goes against every instinct I have -- technical, political, social, economic, etc. You're on your own with a dual vidcard system.
Finally a new video card cooler review fra spcr, and I'm pleased to see that the cooler performs so well acoustically.
I wouldn't dare buy this cooler (or any other cooling solution with a fan) before spcr can vouch for it's noise because it seems that most reviewers have a completely differet noise tolerance than myself and the rest of spcr
So now it's just a matter of time before the noisiest component (GeForce 8800GTS) in my system gets the silent treatment.
I wouldn't dare buy this cooler (or any other cooling solution with a fan) before spcr can vouch for it's noise because it seems that most reviewers have a completely differet noise tolerance than myself and the rest of spcr
So now it's just a matter of time before the noisiest component (GeForce 8800GTS) in my system gets the silent treatment.
I agree with some of what you said, MikeC, but take issue on a couple of points.
My argument about medium power systems not being a major concern any more was not made because of experience, but rather because medium power systems are not hard to cool quietly. This cooler and a Ninja is easily enough with a good case. So while I understand your point that not everyone reading SPCR knows about silent computing, in terms of reviewing products for enthusiasts who want higher end kit it makes sense to really stress them.
To that end I would suggest running CPUBurn and ATITool at the same time. Or Prime95/rthdribl, perhaps. It's certainly an artificial test but proves the system will be stable under any conditions.
The X1950XTX is not a bad choice of card for testing, although it does use GDDR4 which generates less heat than GDDR3, and RAM cooling can be an issue. Probably not in this case, to be fair.
You mention you tested with a Pentium D950 with a TDP of 130W. That's a good match for quad core CPUs and thus an would be excellent choice. I say "would" because the article states that you used a D930, with a TDP of 95W. Still high, admittedly.
I stand by my original argument though. If you did use a D950 then your system should have a fairly similar thermal output to one of mine - a QX6700 and 8800 Ultra. I got a 7-8C rise with a HR-03 + 92mm 10V fan with both CPU and GPU stressed. It looks like you only tested with ATITool and the CPU at idle. In those conditions, I only get a 3C rise too.
Actually, I do have one idea why there might be a discrepancy. Core 2 chips are notorious for having badly concave heatspreaders. A 10C improvement from lapping does not seem to be too optimistic for a quad core processor. Perhaps your Pentium D just makes better thermal contact.
My argument about medium power systems not being a major concern any more was not made because of experience, but rather because medium power systems are not hard to cool quietly. This cooler and a Ninja is easily enough with a good case. So while I understand your point that not everyone reading SPCR knows about silent computing, in terms of reviewing products for enthusiasts who want higher end kit it makes sense to really stress them.
To that end I would suggest running CPUBurn and ATITool at the same time. Or Prime95/rthdribl, perhaps. It's certainly an artificial test but proves the system will be stable under any conditions.
The X1950XTX is not a bad choice of card for testing, although it does use GDDR4 which generates less heat than GDDR3, and RAM cooling can be an issue. Probably not in this case, to be fair.
You mention you tested with a Pentium D950 with a TDP of 130W. That's a good match for quad core CPUs and thus an would be excellent choice. I say "would" because the article states that you used a D930, with a TDP of 95W. Still high, admittedly.
I stand by my original argument though. If you did use a D950 then your system should have a fairly similar thermal output to one of mine - a QX6700 and 8800 Ultra. I got a 7-8C rise with a HR-03 + 92mm 10V fan with both CPU and GPU stressed. It looks like you only tested with ATITool and the CPU at idle. In those conditions, I only get a 3C rise too.
Actually, I do have one idea why there might be a discrepancy. Core 2 chips are notorious for having badly concave heatspreaders. A 10C improvement from lapping does not seem to be too optimistic for a quad core processor. Perhaps your Pentium D just makes better thermal contact.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
MoJo --
It is the D930, not D950. I got it wrong; both the HSF test system and the vidcard test systems were on and switched via KVM -- I obviously read data off the wrong PC.
But as mentioned in that post, the actual power draw into the AUX12V plug is 91W, higher than average by today's standards.
Anyway, while I understand your interest & concern about testing under very hot conditions, I have to say it is not really relevant to VGA cooler testing.
We've set up a fairly hot test bed that is cooled very quietly using very little airflow. It's not meant to replicate a poorly cooled, noisy, system -- that defeats the purpose of our test bed -- which is to check how well a VGA cooler can cool AND with what kind of noise. The reviews are meant to compare the coolers against each other for quiet cooling, not give some "ultimate" thermal torture test. IMO, if you have a system that presents such a test, that's your own misguided business. (Just a fun poke; no insult.)
At the end of the day, does this review tell us how the VF1000 does at similarly quiet levels against other VGA coolers we've tested? I think so. I think it's very clear, the thing handily bests everything else we're tested so far. (tho naturally, there are other new competitors that could do as good a job... hopefully we'll get to them in the near future, too)
Finally, I have to ask why you'd subject your system to a CPU + GPU stress test? Isn't the point of a gaming system to... play games? And if it does well playing games, isn't that good enough?
PS -- I admit I've just been presented with a game that I find rather addictive -- C&C3 "tiberium wars" -- reminds me of "Dune" days way back in the 90s. An ATI 1800GTO card with our review sample of the Aerocase Condorwas installed so I could play this game -- it works nicely, as when temps are monitored mid-game, they're all modest. I've never tried to push the system artificially. I don't see the point; I already know it does what I need it to do...
It is the D930, not D950. I got it wrong; both the HSF test system and the vidcard test systems were on and switched via KVM -- I obviously read data off the wrong PC.
But as mentioned in that post, the actual power draw into the AUX12V plug is 91W, higher than average by today's standards.
Anyway, while I understand your interest & concern about testing under very hot conditions, I have to say it is not really relevant to VGA cooler testing.
We've set up a fairly hot test bed that is cooled very quietly using very little airflow. It's not meant to replicate a poorly cooled, noisy, system -- that defeats the purpose of our test bed -- which is to check how well a VGA cooler can cool AND with what kind of noise. The reviews are meant to compare the coolers against each other for quiet cooling, not give some "ultimate" thermal torture test. IMO, if you have a system that presents such a test, that's your own misguided business. (Just a fun poke; no insult.)
At the end of the day, does this review tell us how the VF1000 does at similarly quiet levels against other VGA coolers we've tested? I think so. I think it's very clear, the thing handily bests everything else we're tested so far. (tho naturally, there are other new competitors that could do as good a job... hopefully we'll get to them in the near future, too)
Finally, I have to ask why you'd subject your system to a CPU + GPU stress test? Isn't the point of a gaming system to... play games? And if it does well playing games, isn't that good enough?
PS -- I admit I've just been presented with a game that I find rather addictive -- C&C3 "tiberium wars" -- reminds me of "Dune" days way back in the 90s. An ATI 1800GTO card with our review sample of the Aerocase Condorwas installed so I could play this game -- it works nicely, as when temps are monitored mid-game, they're all modest. I've never tried to push the system artificially. I don't see the point; I already know it does what I need it to do...
-
- Friend of SPCR
- Posts: 1439
- Joined: Tue Dec 14, 2004 4:06 pm
- Location: New Hampshire, US
- Contact:
I can't speak for Mojo, but I've played games that do stress both cores of my CPU. Wouldn't I want to know that my system can handle the stress of CPU+GPU load?MikeC wrote:
Finally, I have to ask why you'd subject your system to a CPU + GPU stress test? Isn't the point of a gaming system to... play games? And if it does well playing games, isn't that good enough?
I dont doubt this Zalman is far better than most stock coolers.
But what I definately dont like, is Zalman fans in general, they are way too noisy for my taste. And I especially dont like that they aint easy to change to another.
I was very dissapointed by how noisy the stock fan on the VF-900 is, even at 5 volt. I was also dissapointed by the SPCR test of the VF-900, which gave me the impression the VF-900 fan was quiet.
I doubt I ever try a cooler again, which aint passive, or at least support standard fans for easy fan swap.
The good thing about the VF-1000, is that it'll probably satisfy many, many users around the world.
But what I definately dont like, is Zalman fans in general, they are way too noisy for my taste. And I especially dont like that they aint easy to change to another.
I was very dissapointed by how noisy the stock fan on the VF-900 is, even at 5 volt. I was also dissapointed by the SPCR test of the VF-900, which gave me the impression the VF-900 fan was quiet.
I doubt I ever try a cooler again, which aint passive, or at least support standard fans for easy fan swap.
The good thing about the VF-1000, is that it'll probably satisfy many, many users around the world.
I totally agree with you Thomas and I too cant understand why Zalman doesnt make their heatsink with removable fans.
They could just bundle the heatsink with one of their own fans from the factory and people that doesnt like the fan could just throw it away and enjoy a great heatsink with the fan of their choice.
They could just bundle the heatsink with one of their own fans from the factory and people that doesnt like the fan could just throw it away and enjoy a great heatsink with the fan of their choice.
I think you misunderstand, that was not my suggestion. I want a powerful but quiet computer. It's possible, just not easy.MikeC wrote:We've set up a fairly hot test bed that is cooled very quietly using very little airflow. It's not meant to replicate a poorly cooled, noisy, system
I understand you don't like systems like that yourself, and prefer to go for a mid-range system that is easier to cool. I wasn't suggesting that you shouldn't review hardware in such systems, just that it's also worth testing in the most extreme conditions too. If you read the forums quite a few people are trying to build high end quiet systems, and SPCR is in a unique position to provide us information that few other sites can.
It doesn't tell us how the VF1000 compares to the stock cooler in a fully stressed system. You didn't, for example, test with 3DMark or a modern game either, so your test doesn't actually say anything about gaming performance. Therein lies the rub - as I mentioned I got an 8C rise in CPU temps under full load of CPU+GPU. Most games won't hit that (although Supreme Commander comes close) but even so I observed a 6C rise with 3DMark after an hour of combined CPU+GPU tests.At the end of the day, does this review tell us how the VF1000 does at similarly quiet levels against other VGA coolers we've tested?
...
Finally, I have to ask why you'd subject your system to a CPU + GPU stress test? Isn't the point of a gaming system to... play games? And if it does well playing games, isn't that good enough?
I'm not saying it's a bad review, just that it would have been very easy to run a few extra tests while you were there which would have been invaluable to some of us.
Really, thanks for SPCR.
-
- Site Admin
- Posts: 12285
- Joined: Sun Aug 11, 2002 3:26 pm
- Location: Vancouver, BC, Canada
- Contact:
Mojo,
I do understand what you're saying. But it's always the case that there's no way to satisfy all interested parties in a single review, and for extreme gamers, they'll just have to extrapolate. It's what I do all the time when I read reviews that don't quite cover everything I want to know.
I still differ with you on your comment that our cooler test system is mid-range. This is is not true, thermally. For a lean, clean, single CPU/GPU system, the system is just one little step down from the hottest.
Having said that, I've reviewed the video card and VGA cooler test procedures that we've used over the past couple years, and I see that the procedures Larry's used are not quite as demanding thermally as what Devon used to do. The main difference is that in the past, CPUBurn was run in the background to ensure constant 100% CPU utilization while the GPU was also being stressed.
We'll go back to running CPUBurnx2 or Prime95x2 while also running ATI tool or rthdrbl simultaneously. You mentioned 3DMak06 -- this does not give us as high a load on the system (typically 10% lower than what I just mentioned), and its demand on the system fluctuates.
We may repeat the testing on the VF1000/900/ATI stock cooler just to see whether any differences come up.
I do understand what you're saying. But it's always the case that there's no way to satisfy all interested parties in a single review, and for extreme gamers, they'll just have to extrapolate. It's what I do all the time when I read reviews that don't quite cover everything I want to know.
I still differ with you on your comment that our cooler test system is mid-range. This is is not true, thermally. For a lean, clean, single CPU/GPU system, the system is just one little step down from the hottest.
Having said that, I've reviewed the video card and VGA cooler test procedures that we've used over the past couple years, and I see that the procedures Larry's used are not quite as demanding thermally as what Devon used to do. The main difference is that in the past, CPUBurn was run in the background to ensure constant 100% CPU utilization while the GPU was also being stressed.
We'll go back to running CPUBurnx2 or Prime95x2 while also running ATI tool or rthdrbl simultaneously. You mentioned 3DMak06 -- this does not give us as high a load on the system (typically 10% lower than what I just mentioned), and its demand on the system fluctuates.
We may repeat the testing on the VF1000/900/ATI stock cooler just to see whether any differences come up.
Yeah, sorry, I don't know why I said that. Distracted I think.MikeC wrote:I still differ with you on your comment that our cooler test system is mid-range. This is is not true, thermally. For a lean, clean, single CPU/GPU system, the system is just one little step down from the hottest.
Good to know. As for 3DMark, I know it doesn't push as hard as more synthetic tests but it's a good indication of likely performance in games. It only takes a couple of hours to run in a loop (disabling the CPU only and theoretical tests of course).Having said that, I've reviewed the video card and VGA cooler test procedures that we've used over the past couple years, and I see that the procedures Larry's used are not quite as demanding thermally as what Devon used to do. The main difference is that in the past, CPUBurn was run in the background to ensure constant 100% CPU utilization while the GPU was also being stressed.
If you do decide to re-test the VF1000, I would be very interested to see the results. My feeling is that you won't get as big an increase in CPU temps as I did. I am tempted to try the HR-03 again once I have lapped the CPU to see how much of a difference that makes.
-
- Patron of SPCR
- Posts: 445
- Joined: Mon Apr 03, 2006 10:23 am
- Location: Pacific Northwest
The fan is removable. The VF900 fan removes with screws, while the VF1000 fan can be unclipped (tricky to remove though).AuraAllan wrote:I totally agree with you Thomas and I too cant understand why Zalman doesnt make their heatsink with removable fans.
They could just bundle the heatsink with one of their own fans from the factory and people that doesnt like the fan could just throw it away and enjoy a great heatsink with the fan of their choice.