Corsair H50 Water Cooling Kit

Gaidin109 said:
goes to show how little you know then doesnt it. I will do the h50 again if you want.
realtemp used the cpuz engine anyway.
Click to expand...

I'm not bothered whether you do it over or not - but if you want your results to be taken seriously then you should expect to have to do them in a methodical fashion.

And if they both use the CPUz engine, then one is broken becasue CPUz says one thing and RealTemp says another. When you start OCCT is specifically says NOT to use it with COUz running as they both access the same data and give spurious results. This differential between reality and programs reported results has been known for years.

If you want to have a pop and call me ignorant, then fine, but all I'm doing is looking at your results critically and saying that it doesn't pass scrutiny. Pray you never get audited by a professional if you think what you've posted is acceptable evidence or if you think anyone else on here is being tough on you.
 
Just wondering - when will OC bring H50 in stock ? It been taken so long times to wait ? In fact, I haven't pre-order it yet but I rather to wait until OC stock it.
 
WJA96 said:
I'm not bothered whether you do it over or not - but if you want your results to be taken seriously then you should expect to have to do them in a methodical fashion.

And if they both use the CPUz engine, then one is broken becasue CPUz says one thing and RealTemp says another. When you start OCCT is specifically says NOT to use it with COUz running as they both access the same data and give spurious results. This differential between reality and programs reported results has been known for years.

If you want to have a pop and call me ignorant, then fine, but all I'm doing is looking at your results critically and saying that it doesn't pass scrutiny. Pray you never get audited by a professional if you think what you've posted is acceptable evidence or if you think anyone else on here is being tough on you.
Click to expand...


But you are still wrong nevertheless.

Remember Load Line Calibration is off, so the cpu follows the intel spec, so will step down as the load decreases. idle voltage is higher, but load voltage is variable, dependant on the load. Besides I was offering an opinion not a review.
I'm not religious, so praying isnt something I would do, also I dont think being audited by a "professional" is relevant especially considering my occupation. The original poster asked if anyone had used both. I had, and gave what info I had. I do not have the Megahalems any longer to do a definitive test, otherwise I would.
Yes I know the prime test was for 10mins and is not a test of stability, but does give some idea of what temps to expect.


i74ghz50oc.jpg
 
Last edited:
Gaidin109 said:
But you are still wrong nevertheless.

Remember Load Line Calibration is off, so the cpu follows the intel spec, so will step down as the load decreases. idle voltage is higher, but load voltage is variable, dependant on the load. Besides I was offering an opinion not a review.
I'm not religious, so praying isnt something I would do, also I dont think being audited by a "professional" is relevant especially considering my occupation. The original poster asked if anyone had used both. I had, and gave what info I had. I do not have the Megahalems any longer to do a definitive test, otherwise I would.
Yes I know the prime test was for 10mins and is not a test of stability, but does give some idea of what temps to expect.
Click to expand...

LLC has absolutely nothing to do with varying the multiplier. The screenshot shows the system loaded over 99% so all the cores should be running at the full 21x multiplier. They're not. They;'re running at the 16x multiplier and you're only doing 3GHz, not 4GHz.

I have absolutely no idea why you keep posting 'prime' shots where the system is running cooler than the maximum recorded temperature. 4 minutes before the screenshot was taken the CPUs were all running 2-4C higher than at the time of the screenshot. Why?

I have no idea why you are so defensive about all this. All anyone (including me) is pointing out is that you are making claims that are not backed up by the evidence pictured.

And if I'm so wrong how come CPUz and RealTemp are suddenly saying the same thing?

There is still something odd about these readings. Why do you have such a massive differential in core temperatures? They are less than 1cm apart yet you have an 8C differential?
 
Last edited:
You posted a screen of the mega on air 200x20 with Loadline on. What i would like is the same OC you used then to be used with this h50 set up so we can compare. You can call diff BLCK speeds and voltages fair.
 
WJA96 said:
I have absolutely no idea why you keep posting 'prime' shots where the system is running cooler than the maximum recorded temperature.

I have no idea why you are so defensive about all this. All anyone (including me) is pointing out is that you are making claims that are not backed up by the evidence pictured.

And if I'm so wrong how come CPUz and RealTemp are suddenly saying the same thing?

There is still something odd about these readings.
Click to expand...


You are not listening...

when you turn off LLC, you are no longer forcing a constant voltage to the cpu, thus the voltage will vary depending on load. Prime does not give 100% load all the time, it also differs, hovering around 95%-99%. When the processor draws current it also produces heat if this the voltage is variable so is the temp produced. this is why 68c is shown when I snipped the shot. the temp never rose above 72c during the 10mins, but the temp was around 68c-70c most of that time.
What is wrong with the evidence, you can see plainly what is running, how long its been running, what the idle and max temps are.
To be honest I couldnt care less, I just dont suffer fools gladly thats all.

And Bradley as I said originally the temp difference is not worth worring about, its a couple of degrees if that. I dont want to enable LLC on my system due to the risk of voltage spikes. I was just giving my opinion thats all, and to be honest didnt expect to get slagged off for it.
The reason I changed from the fenrir to megahalems was temperature and the difference was about 10c under load, the reason I changed from the megahalems to the Corsair was due to room and less weight stress on the Mobo. I want to increase to 12GB soon and needed to get to the dimm socket that was inaccessable with the megahalems.
Bulldog could probably do a specific test as I think he has both the Megahalems and the Corsair.

I see from your Sig that your running 4.2 at 1.23v, so I dont think you have much to worry about. I have an IBT vhigh 16thread 10 pass test on 4.2OC with the H50 if your interested. Temps maxed out at 80c with a 32c case temp. I'm sure someone with pick this to pieces too. but it will give you an idea.

42IBT10pass.jpg
 
Last edited:
Gaidin109 said:
You are not listening...
Click to expand...

No, but I'm looking and reading very carefully.

Gaidin109 said:
when you turn off LLC, you are no longer forcing a constant voltage to the cpu, thus the voltage will vary depending on load. Prime does not give 100% load all the time, it also differs, hovering around 95%-99%. When the processor draws current it also produces heat if this the voltage is variable so is the temp produced. this is why 68c is shown when I snipped the shot. the temp never rose above 72c during the 10mins, but the temp was around 68c-70c most of that time.
Click to expand...

What does any of that have to do with the multiplier dropping to 16x?

Gaidin109 said:
What is wrong with the evidence, you can see plainly what is running, how long its been running, what the idle and max temps are.
Click to expand...

Yes, except it's running at 3GHz in one test and 4GHz in the other. And you still haven't explained why the temperatures are so variable?

Gaidin109 said:
To be honest I couldnt care less, I just dont suffer fools gladly thats all.
Click to expand...

Uh huh... So now I'm a fool? Why? Because I queried some thoroughly spurious results? Fine, I'm a fool, but at least I still have some credibility.

I certainly don't think this H50 thing is anywhere near as good as it's being made out to be.
 
Im not considering either a mega or this corsair im just interested in the difference between the 2. I understand your worried about voltage spikes but your a fair way below the kinda voltage that will damage your rig for one stress test.

Anyway what software did you use to stress during this run?

Capture1.jpg


If you want to then it would be interesting to see the same OC settings used with the same program but with the H50 for a fair comparison.

This should show once and for all if the H50 is any better than decent air.
 
WJA92 is far from a fool, all you've done is started defending odd test methodology with personal insults.

I'm sure Gaidin can see why posting adjacent screenshots showing the Corsair to be far better than the megalahems to the uninformed reader is misleading. The conditions are different so the test result means little, but it will be taken as conclusive evidence for one over the other. Still, I thank you for posting the screenshots.

I'm also skeptical of this cooler. If a 240 radiator in a dedicated wc system is considered the minimum to cool an i7, then why does the h50 with roughly half that surface area do so well? Either the results on the h50 are misleading, or we don't need a 240 radiator for the i7.
 
Last edited:
JonJ678 said:
WJA92 is far from a fool, all you've done is started defending odd test methodology with personal insults.

I'm sure Gaidin can see why posting adjacent screenshots showing the Corsair to be far better than the megalahems to the uninformed reader is misleading. The conditions are different so the test result means little, but it will be taken as conclusive evidence for one over the other. Still, I thank you for posting the screenshots.

I'm also skeptical of this cooler. If a 240 radiator in a dedicated wc system is considered the minimum to cool an i7, then why does the h50 with roughly half that surface area do so well? Either the results on the h50 are misleading, or we don't need a 240 radiator for the i7.
Click to expand...

It was never supposed to be conclusive in the first place. As for WJA92, he is obviously just taking the mickey or he is foolish if he thinks that putting your cpu under load gives a constant temperature. It does not. Also the multiplier does not just go from 21 to 20 to 12, it drops down incrementally as the processor load drops and when the screenshot was taken it was at 16x thats all. As I'm typing this, the cpu multiplier is changing regularly as the PC is working away. The temps wont be comparable anyway due to ambient fluctuations.
I said it was just to give some idea of the difference and I also said at the beginning that the difference was negligible anyway.
Even when I took a screenshot of everything as it was working, he still said it was crap, so what am I to think.
I'm tired of all this crap anyway. Back to Anandtech forums.
 
Last edited:
Gaidin102 said:
As for WJA92, he is obviously just taking the mickey or he is foolish if he thinks that putting your cpu under load gives a constant temperature. It does not.
Click to expand...

Well, mine all equilibrate under load and once they get up to the equlibration temperature they pretty much stay there. And I never see big variations in CPU Core temperatures either. Certainly not more than 2 degrees across the cores.

Gaidin102 said:
Also the multiplier does not just go from 21 to 20 to 12, it drops down incrementally as the processor load drops and when the screenshot was taken it was at 16x thats all. As I'm typing this, the cpu multiplier is changing regularly as the PC is working away.
Click to expand...

One of my former business partners was an Intel Salesman in a previous life. After I read this, I asked him what he thought and his undertanding is that Speedstep is on or off. If it's on, it's 12x on the i920, if it's off it's 20 or 21 depending on whether or not Turbo mode is enabled. And it still doesn't explain why CPUz showed one thing while RealTemp showed something else.


Gaidin102 said:
The temps wont be comparable anyway due to ambient fluctuations.
I said it was just to give some idea of the difference and I also said at the beginning that the difference was negligible anyway.
Click to expand...

But your screenshots don't show that - they show what appear to be big differences.

Gaidin102 said:
Even when I took a screenshot of everything as it was working, he still said it was crap, so what am I to think.
Click to expand...

Some screenshots don't look right, that's what you're supposed to think. And I never said it was crap, I just said they were odd.

Gaidin102 said:
I'm tired of all this crap anyway. Back to Anandtech forums.
Click to expand...

OK. Well, I've been a member on there for about 6 years, so I'll see you around.
 
Regarding multiplier, I dont know what windows does but linux is letting me select any multiplier from 12 up to 20 and switches up one at a time under load if I ask it to. Windows might do the same thing *shrug*
 
JonJ678 said:
Regarding multiplier, I dont know what windows does but linux is letting me select any multiplier from 12 up to 20 and switches up one at a time under load if I ask it to. Windows might do the same thing *shrug*
Click to expand...

But it doesn't do it by itself - it lets you select multipliers in software, but under 99.7% load, unless you've selected another multiplier manually, it will be running at 20 or 21x.
 
You must log in or register to reply here.
Back
Top Bottom