People in NVIDIA pronounce the “Ti” differently

Pronunciation can sometimes be a subjective matter (is Asus pronounced “A-Soos” or A-Sus”?) but you’d expect there to be some sort of consensus within the actual company. NVIDIA has a “Ti” at the end of the model number for its high-end GPUs, but people seem confused about how it’s pronounced; people either pronounce it as “tie” or as separate letters. The Verge outlined how the senior vice president at the company, Jeff Fisher, pronounces it as “tie.” He said it during the company’s keynote introducing the latest GPUs including the RTX 3090 Ti. On the other hand, the company’s founder, Jensen Huang, goes the other route and pronounces it as two separate letters. While it’s not quite the biggest scandal in tech, it’s still rather odd that the company can’t settle on a pronunciation.

Also, NVIDIA announced a new GPU, the RTX 3090Ti

Jeff Fisher held a keynote earlier as part of the company’s CES 2022 event where he introduced us to a new player in the GPU market. Fisher held up the RTX 3090 Ti and then hit us with a torrent of specs. Based on the name, you can tell that it’s the company’s most powerful GPU that’s coming to the market. Unfortunately, he didn’t tell us much more about the GPU other than the specs. It’s capable of 40 shader teraflops, 78 RT teraflops, 320 tensor teraflops, 24GB of 21Gb/s G6X memory. Based on the numbers alone, it seems like it’s going to be a really powerful GPU.

In other NVIDIA news: NVIDIA is launching new versions of its RTX 2050 GPU

Even though we’re talking about new GPUs, there are some old ones that are making headlines. Recently, NVIDIA announced that it’s launching a revised version of its RTX 2050 GPU. The company isn’t just taking old units and putting them back on store shelves, however. These newer RTX 2050s will have some notable improvements. It will share a few key specs with the higher-end RTX 3050. The new RTX 2050 will have the same Ampere GA107 architecture, GDDR6 memory, and 2,048 CUDA cores as the RTX 3050. While this is nice, the new RTX 2050 will retain its 64-bit memory bus. These new GPUs will launch early this year.