With no official phrase, I took a more in-depth take a look at the precise values and remembered our NVIDIA segregates Turing GPUs article, that was a part of the launch protection for the preliminary GeForce RTX unveil. In that article, we revealed that NVIDIA is creating two fashions for every GPU, which can be similar in each regard, apart from title and worth. If board companions need to construct a factory-overclocked card, they’ve to purchase the -A variant of the GPU, as a result of solely that’s allowed for use with an out of the field overclock. Manual overclocking by the customers works precisely the identical on each items.
For newer releases, NVIDIA deserted that idea once more, promoting only a single GPU for every SKU. But then why three machine IDs? Multiple variants of GeForce RTX Super playing cards are positively on the market, as we’ve acquired e-mails and seen feedback on our boards, from patrons questioning why their GPU “is different,” the way it impacts efficiency, or if it might even be a faux card.
In the above record, I reordered and grouped the entries within the screenshot to assist information the attention. While the primary SUPER entry in every group has a tool ID matching the NVIDIA Founders Edition, offered to us by NVIDIA for our evaluations, there are two extra entries for every SKU. If you’re taking an in depth take a look at the machine IDs, you may notice that every one simply differs by a price of 40 Hex from the one other fashions (that is what the pink arrows point out). Why would NVIDIA create these SKUs, based mostly on current fashions? It appears that in an effort to shield current stock in warehouses of their board companions, NVIDIA has created a technique to show current RTX 2070 non-Super playing cards into RTX 2060 Super; and RTX 2080 non-Super to RTX 2070 Super, respectively. This additionally explains why there’s two new IDs for every card: each non-A and -A GPUs might be transformed.
We think about that is completed by merely re-configuring the playing cards to disable some parts, however the technique employed might be a bit extra subtle than merely altering the video-BIOS and slapping on a brand new sticker. The cause why 40 Hex was chosen as a differentiator between each, is that it represents only a single bit distinction within the machine ID. Such a bit change is the smallest potential change, and we suspect it’s applied via a small resistor that will get soldered onto the PCB (or eliminated), which the GPU probes throughout its inside startup, and adjusts its reported machine ID accordingly. It’s for that reason that it could not be capable to flash an RTX 2060 Super again to an RTX 2070 offered the PCBs and ASIC match; or flashing an RTX 2070 Super to an RTX 2080.
Now in fact you are questioning what causes board companions must change current, completely wonderful playing cards sitting in warehouses, able to go, into slower (= cheaper) variants. Let’s take a look at the financials shortly. RTX 2080 at present sells for $640, flashing that into RTX 2070 SUPER, promoting for $520 makes little financial sense. Things do not look that significantly better for the $450 RTX 2070 to $400 RTX 2060 SUPER transition. Even if a board associate is having a tough time promoting RTX 2080 for $630, they may positively promote them for $600, and nonetheless make a greater revenue than turning it right into a $520 RTX 2070 SUPER. My guess is that NVIDIA desires to extend obtainable stock for SUPER, scale back stock of non-SUPER at…
