| HIS Radeon HD6950 IceQ-X Turbo-X Video Card |  |
| Reviews - Featured Reviews: Video Cards | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Written by Steven Iglesias-Hearst | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Thursday, 09 June 2011 | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Page 17 of 18
HIS HD6950 IceQ X Turbo X OverclockingBefore I start overclocking I like to get a little bit of information, firstly I like to establish operating temperatures and since we know these are nice we can quickly move on. Next I like to know what the voltage and clock limits are, so I fired up MSI's Afterburner utility. I established that vCore was adjustable between 0.800v and 1.299v and clock speeds were limited to 950MHz max on the GPU and 1350MHz (5.4GHz effective) maximum frequency for the memory. This is not a lot of range to move forward with but I'm sure I can squeeze every last drop of performance out of the HIS HD6950 IceQ X Turbo X. My preferred weapons are MSI Afterburner (v2.2.0 Beta 3) for fine tuning while using FurMark (v1.9.0) to heat the GPU.
Without needing to raise the vCore I was able to push the GPU to its 950MHz limit. Increasing the memory speed to 1350MHz (5.4GHz effective) also required very little effort at all. I am really impressed by the capabilities of the already overclocked HIS HD6950 IceQ X Turbo X video card, I'm sure there is still some overclock headroom but 950MHz/1350MHz is the highest Afterburner will allow.
Armed with a 70MHz GPU overclock (150MHz over reference design) and a 50MHz memory overclock (100MHz over reference design), we went back to the bench and ran through the entire test suite. Overall we saw an average 6.77% increase in scores (at 1920x1080 resolution. We ran the fan speed at 75% to ensure the card stayed nice and cool during the overclocked benchmarking run as this card can get very hot if the fan speed is not adjusted accordingly. We also re ran temperature tests at the overclocked speeds at a slightly higher ambient temperature of 26°C. The IceQ X cooler on the HIS HD6950 IceQ X Turbo X once again did not fail to please, at idle the GPU sat at 44°C (29% fan speed). Pushing the temperature up with FurMark saw the GPU load temperature rise to 78°C (51% fan speed). The auto fan speed is quite relaxed so next I tested at 75% fan speed and the temperature dropped to 68°C. Cranking the fan on manual to 100% saw the temperature drop to 63°C That's all of the testing over, in the next section I will deliver my final thoughts and conclusion.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
*-
Comments
I'm glad to be different from the masses, I do a lot of research before I buy things and I like to get the best deals possible.
Your comparison charts spoke of something, to me at least. All the cards in the comparison would be pleasing to own. Not a real "dog" amongst them. I'm keeping in mind that gaming performance is a primary consideration. But even without games, these cards are all excellent to have in an HTPC or demanding video environment.
Oh and one more thing to watch when buying HIS graphics cards is the Fan wont last long if it's anything like the one on my HIS HD5770 (which it looks to be) not even 6 months old and it's already noisy as on start-up takes about 10~15mins of running before it gets to some semblance of quiet.. You all know the sound that grinding labored sound a fan starts to make when it's just about to fail
try it at 1920x1080 with lower AA and see how it goes
For the benchs I seen of one of those cards fully unlocked and with some good extra OC put in, its performance is way better than 570.
Unfortunately, these low-profile manufacturers are so for a reason, they don't have any feature to stand themselves ought of several sharks like PowerColor, Sapphire, XFX, ASUS and of course MSI.
In the end you only need that super high speed fan on TF3 if you're OCing it close or beyond a 1ghz core frequency.
But please note this are just suppositions I've made base on looking at both coolers and reading testimonies. For example, notice that HIS cooler, unlike TF3, hasn't got the back opened, allowing to suck in the air coming from front fans.
P.S.: For instance, because of that feature alone, I'm thinking on making a mod to my case so I can mount a Noctua P12 fan close to HD bays, to pressure air into the Graphics card cooler through the back of the cooler (the 6 pin connectors are above, so the intake from that position is very optimized)
And you completely ignored what I said about the noise per watt of heat spread. Without keeping this in mind you're just comparing apples with oranges...
Your guesses and suppositions might be plausible in the absence of any concrete data, but reviewer Steven Iglesias-Hearst did in fact point you to the concrete data that informed his assessment, data that he derived from (1) Full Reviews of the two cards that (2) He conducted personally, 17 days apart, using (3) A single Temperature Testing methodology (FurMark 1.9.0 Torture Test) and using (4) The exact same 'Lancool PC-K63'/'Intel P55'/'i5 760' System for each Review.
It?s rare that so many variables (Reviewer, Methodology, Case and System Cooling Config) are kept constant when comparing two cards, so his assessment of the relative merits of the two Cards should probably carry more weight than random testimonials. His results:
HIS 6950 IceQ X - Ambient Temp: 24 / Fan: 100% / FurMark Load Temp: 61 / Sound: ?Quite Bearable? Noise is a non-issue?
MSI 6950 Twin Frozr III -- Ambient Temp: 28 / Fan: 100% / FurMark Load Temp: 59 / Sound: ?Noise level is too much??
Granted, (5) Ambient temperatures are different (24 versus 28 degrees), but if (6) the Twin Frozr Fan is creating significantly more noise than the IceQ X, while (7) achieving approximately the same Load Temperature in FurMark (59 versus 61 degrees), then (8) your speculation -- ?You would need a lower speed [on the Twin Frozr 3] to match the [HIS IceQ X], and at that lower fan speed, it would make less noise than the top HIS heat spread? ? is not really consistent with the data.
Note also that in the test, the HIS Card was running at a higher default clock speed (880 MHz, versus 850 MHz for the MSI Card).
To be sure, we could ask Steven Iglesias-Hearst to (9) quantify (in decibels) ?quite bearable? and ?too much?, and also to (10) demonstrate the effect of a 4 degree difference on Ambient Temperature on FurMark Load Temps, but I think the data he has already provided is more than enough to support his assessment that the HIS Cooler is ?better? overall.
In the absence of such survey data, the overclocking results in Hearst?s review -- Street Fight IV 152.76 FPS @ 950/1350 (max OC) for the HIS IceQ X, versus Street Fighter IV 143.30 FPS for the MSI Twin Frozr III @ 900/1325 (max OC) -- give us the most helpful picture of the relative power/performance of the two cards.
Those OC numbers (900/1325) aren't the mirror of that MSI's true potential, since I've read dozens of high OCing numbers, being 975mhz on the core what most got, and a few lucky individuals got 1000mhz+ (cant remember memory speed). With such numbers I believe MSI's card would get similar results as HIS', and but with more durable and stable OC with its well-known military class components.
Here's win or win situation in which specific personal requirements make a difference (OC roof, silence, durability), but overall they seem to be in same level (opposite of my previous statements)
I end up loosing the right time for purchase with HD6xxx, so for HD7xxx I'll keep an open mind about this 2 coolers.
Offtopic: Does anyone think the thermal paste that comes with the card (with all brands) is as good as Arctic Silver 5 or a similar one? Never got the chance to compare then and now results with same ambient temperature, reason why I'm asking.
It's very hard to say without doing actual testing, but in my honest opinion you will not gain more that 2-3 degrees Celsius if you are lucky.
idle temps before AS5 were 36~37c after AS5 and curing time 27c
load temps before AS5 63c after AS5 52c (using F@H to load GPU's)
Because over time thermal paste gets dry and loses some of their heat transference capabilities, reason why if you already used the card for some time it's an unfair comparison.
I got the same results back then, after changing the 2 year old thermal paste of my HD4850, reason why I guess you did the same.
to flash my card is one that is the 840mhz version
thanks before hand
I need this card is at 880MHz core gpu bios
can use this bios
my original bios is the 840mhz version?