Two very unequal sisters
The fact that the Founders Edition of Nvidia's GeForce RTX 2080 Ti would have particularly good chips is not only an urban legend, but complete nonsense. Clock-adjusted and provided with the same BIOS, e.g. the board of my press sample from Nvidia water-cooled also no better than e.g. the KFA2/Galax RTX 2080 Ti OC directly from the assembly line, wherethe hong Kong card is only the "entry" model, i.e. the cheapest of all RTX 2080 Ti of this manufacturer.
But I was really only interested in the pure chip comparison. For this purpose, I secured the BIOS of the KFA2/Galax RTX 2080 Ti OC and first installed the firmware of the Founders Edition. Both cards ran in the same performance window, within the framework of normal measurement tolerances. Golden Sample for curious journalists? I think Master Jensen certainly has other problems than to have the cards specially preselected for the hunter-gatherers of wild pixels.
The five reference boards tested were all somehow the same, whether press samples or retail. And yet the board of the KFA2/Galax RTX 2080 Ti OC is a little more equal than the same, which of course was immediately an opportunity for me to hook in at this point and above all also in terms of measurement.
First of all, I sat above the individual test of the KFA2/Galax RTX 2080 Ti OC (which is yet to come this week) and found that both boards are absolutely identical and also equipped, but the firmware in the BIOS of the non-Nvidia card has it really in it. Instead of a power target of 260 watts and a maximum power limit of 320 watts, the BIOS of the KFA2/Galax RTX 2080 Ti OC approves a whopping 320 watts of power target and incredibly high 380 watts as the maximum limit.
While the board partner cards could be flashed without any problems, this attempt with the FE failed both on Windows and under DOS due to the different hardware ID. It didn't matter if the write protection was deactivated by parameter and/or the device driver was completely uninstalled beforehand. No chance of nothing.
Stable overclocking up to 2145 MHz, but not with the FE
In less demanding games like Doom, the watered KFA2/Galax RTX 2080 Ti OC even manages 2170 MHz stable over runtime when the GPU can be operated significantly below 40°C. But Doom is neither a demanding game nor a true benchmark. I tested both Witcher 3 in Ultra-HD (DX11) and Forca Motorsport 7 (DX12) and recorded a stable 2145 MHz even after over 30 minutes. For this purpose, I set the Power Target to 126% for the KFA2/Galax RTX 2080 Ti OC, which is roughly equivalent to the power limit of 380 watts stored in the BIOS.
With maximum Power Target, the Founders Edition managed to alternate with only 2025 to 2040 MHz, which it was able to maintain over a long period of time with its significantly lower power limit stored in the firmware. That alone is reason enough to look for one of the cheaper board partner cards with reference board, which have an increased power limit. Because even from the point of view of the overclocker it is worth it, even if the efficiency fanatic in me already gets goosebumps.
But if you can spend more than 1200 euros on such a card, you will certainly have the coal left over for the required socket gold. Even if I deviate somewhat from my principles when making such a statement. Let's look at the two maps again in a direct benchmark comparison:
Scaling with the clock increase still works reasonably, amazingly enough. We will see how far this hammer then penetrates to the power station. First, a very similar output under DirectX 12:
Coolable? With water already!
I tested the map with a prototype of a fullcover water block, which has also been completely reworked in the meantime, but nevertheless already performs quite passably. The time-lapse video (1:10) shows us very nicely that even this piece of experimental metal already has everything loosely under control.
But let's turn around quickly and get to the essentials, which is what this extraordinary fun gives us even more in certain incidental costs…
Kommentieren