Ansau

Lol people are so dumb to believe whatever they see on the internet? These benchmarks have been proven to be false no matter how official they are: - The 480 is nowhere near performing that low. Go search trusted reviews and you can see in the same game and same settings gettiong like 20 fps more. - Still no true frequencies of the 1060. So far, we have a Firestrike comparison where the 1060 at 1911MHz being 8% slower than a 480 at 1350MHz, or at 2126MHz being only 16% faster than 480 at stock.

Well, in any case, included performance/price, the rx 480 is still better than the 970. But performance/price doesn't contradict the fact that in this thread people took the numbers they wanted to make Polaris vs Maxwell efficiency biased to Nvidia.

You cherry-picked some testing that showed higher power consumption. Curiously, these sites are some of those having 480 with over-consumption in the pcie. You must also be very bad at looking through internet. Sites like anandtech, pcper, techspot, sweclockers, purepc, techreport, computerbase or pclab show the rx 480 having very similar power consumption than the 970. And if you want to compare efficiency through nº of transistors, you should count those disabled in the 970. As we don't know how many transistors are included in a SM, only full chips must be counted. Moreover, comparing the effeciency at the performance should be done with a more extended workload scenario, not just dx11 games where Maxwell is designed for. Computing, dx12 games, mining... should also be counted, and AMD tends to be better.

Wowow, how much bias in the Polaris vs Maxwell... First of all, you should be doing the comparison with 980 vs 480, both full chips, and not with the 970 which is a cut version. Then, most reviews show the power consumption of the rx 480 is around the same, if not a bit lower than the 970, which means those 166W of the 480, but 154W of the 970 aren't fair. If we put the power consumption of the 480 at 150W: Polaris: - 2.78W per 10.000.000 transistors. - 36.000.000 transistors/watt. Maxwell: - 3.17W per 10.000.000 transistors. - 31.515.151 transistors/watt. Polaris is about 15% more power efficient than Maxwell. And that's with Maxwell at reference frequencies (1127/1216 Mhz). And Pascal is about 10% more power efficient than Polaris (Again comparing full chips 1080 vs 480).

A bit too optimistic. GTX 1080 is about 80% faster than 980. Also, the 1080 competes with much faster gddr5x vram. The 1060 will presumably be half a 1080 (1280 CUDA vs 2560). The 1080 has seen a +10% price increase, while the 1070 is at +15%. GTX 1060 has more chances to be between 970 and 980 at 230-250$ than what you claim.

Well, good luck with prioritizing color scheme over product quality.

Well, that Nitro cooler will be overkill, suficient for the most extreme OC possible. In the 380 series it is one of the best among all partners and capable to keep a 180W gpu with an OC of about 25% at sub 80ºC without needing fans to be at 100%. I only hope Sapphire puts bearing ball fans to rest for good. About the aesthetics, not the best but personally better than some horrors found in the new GTX 1000...

Well, first we need to understand what does it mean "degradation". Degradation is the process where the cpu needs slightly more voltage to maintain the same frequency as it wears. It is mainly produced by electromigration, the process where there is a flux of mass transport of the metal produced by the momentum transfer of the electrons and atoms. High rates of electromigration can lead to failure. The flux is measured by: As we can see, higher current density (j) and temperature (T) accelerates the process. It is mistaken that the voltage has a direct impact in this process, but it doesn't. However, higher voltages will produce higher current which will increase the current density. Also, due to the Joule effect, higher voltages will produce more heat, but this can be countered with more cooling. It has also been found a much practical way to measure the electromigration using the Black's Law, where MTTF is the "Mean Time To Failure": MTTF is normally mentioned in the warranty period and, in the case of Intel, boxed cpus are rated for a minimum of 3 years, or around 26.000 hours. This is the minimum amount of hours Intel certifies the cpu last working at 100% non-stop. Looking at previous equation we can see that big increasings of temperature and voltage are required to affect in a noticeable way the lifespan of a chip. Even when overclocking, these are rather small, as voltages are usually just 20-30% higher and we have deltas of temperatures of only 10-15ºC. We can see it in the relation of voltage increase over time of Intel 45nm chips (back at old Core 2 Duo). At the worse case scenario predicted by intel, there's only an increase of 0.125v about 5 years later (~45.000 hours): So, degradation doesn't slow down any chip at all. And in any realistic scenario of use, degradation happens at such an extremely slow rate, we can assume no death-time for consumer use. Mainly because a cpu is far from being used all the time at 100%. But also because voltage and temperatures changes at which any silicon component is working are relatively low. In fact, to significally accelerate the degradation of a cpu or gpu, you need to intentionally go nuts with voltages for 24/7 use without caring about temperatures. Some sources: http://www.anandtech.com/show/2468/6 http://www.overclockers.com/forums/showthread.php/723980-Truth-about-CPU-degradation https://en.wikipedia.org/wiki/Electromigration http://www.eetimes.com/document.asp?doc_id=1275855

Sorry but this is one of the dumbest things I've ever read in the Internet. Any cpu or gpu will last you for several years, so many that you'll replace it for the lack of performance way before it dies or it becomes unstable.

When AMD releases Zen we will talk about it. Until then, AMD has nothing on the table. They put thier highest solution against one of the lowest from intel and it barely wins. Put the highest intel model and let's see. Oh sorry, I didn't remember some people are fooled by rebranding. FX-8800p and FX-9800p are exactly the same chip, but the second has a couple of more features that aren't enough to solve the issues...

i7 6650U only cost 22$ more than the i7 6500U, and then you have the i5 6260U that costs 89$ less, both having the HD 540, which twice the amount of EUs and EDRAM. It is well known that the issue doesn't come from bad cooling solutions, but the fact that when the cpu and gpu are both working at the same time, they throw more than just 15W, so if configurated at that TDP the APU downclocks. Prime95 doesn't show throttling issues.

They aren't that different. Same lithography, same amount of transistors, same specs, similar clock range... The only differences introduced are improved I/O capabilities (which have nothing to do with heat output), and improvements in terms of voltage regulator and adaptive frequency-voltage curve, but only allow a 10% improvement of frequency/voltage ratio over Carrizo 1.0. But more importantly, Carrizo was seeing throttling in the cpu of 1.3GHz from a base of 2.1GHz, and gpu of 300MHz from a maximum of 800MHz. These aren't tiny throttling that can be fixed with some tweaks that bring only a 10% improvement, also the new 9800P has high cpu clocks, and let's not forget that throttling is experimented in actual implementations, rather that theoretical values provided by AMD in a laboratory/ideal situation. I already said both are based on Excavator. The power efficiency boost is only a 10% over Carrizo, not 35%. And sorry, but APUs have always been targeted at gaming in the entry level market. As a matter of fact, the comparison they make against the i7 6500u is exclusively at gaming.

Both are based on Carrizo. In fact, the FX-9800P is just a FX-8800P with different frequencies. FX-8800P: 28nm 2m/4t Excavator at 2.1/3.4GHz. R7 gpu with 512 shader cores at up to 800MHz. FX-9800P: 28nm 2m/4t Excavator at 2.7/3.6GHz. R7 gpu with 512 shader cores at up to 758MHz. Btw, Intel created Iris Pro for more gaming oriented solutions, while those cpus with the normal HD 520 are targeted towards more office-multimedia tasks requirements. It is a bit unfair to compare one of your gaming solutions against an office-multimedia solution from the competition, and leave the room saying yours wins in gaming. Of course it will...

15W they say, throttling they hide. http://www.gaminglaptopsjunky.com/amd-carrizo-fx-8800p-benchmarks-15w/

There are far better examples: Once you switch to live tiles based experience, both Android and iOS feel so clunky and rudimentary it is very hard to go back to them, despite all the little things of W10M.