New Nvidia Turing RTX 2000-series key features detailed ahead of review embargo lift
A couple of days before Nvidia lifts the review embargo for the new RTX 2000 GPUs on September 14, the guys over at Videocardz unveiled a short summary of the Turing architecture white paper, which details all the new technologies that facilitate a theoretical 50% performance boost over the Pascal chips.
The list of key features includes:
• new INT32 cores that enable concurrent floating point and non-floating point instruction execution (may lead to 36% additional throughput for floating point operations)
• improved shading algorithms: mesh shading (allows more objects per scene), variable rate shading (limits shaders when not in field of view), texture-space sharing (shading results stored in memory for faster access), multi-view rendering (single pass rendering for multi views)
• Turing memory compression – 50% increased bandwidth compared to Pascal
• DisplayPort 1.4a (2x 8K @60 Hz) + NVENC (encode H.265 stream at 8K/30 FPS) + NVDEC (decodes HEV YUV444 10/12b HDR, H.264 8K and VP9 10/12 HDR)
• two 8X gen 2 NVLink (RTX 2080 Ti only; RTX 2080 features 1 NVLink), discontinued 3/4-way SLI support
The first detailed reviews of the Turing gaming GPUs should be published over the weekend and we will finally find out if it is worth it to buy into the ray tracing hype, or maybe it would be a smarter move to grab the GTX 1080 at extra-discounted prices.
Source(s)
