Reztek Systems

Technology, Security, and More


A Tale of Two Presentations: nVidia and Intel

Within the past forty-eight hours, nVidia and Intel held their respective product announcements and demonstrations. The price-to-performance ratio for the RTX 3000-series GPUs returns to a consumer-friendly form. While the presentation was light on quantified performance numbers for specific benchmarks, the replacement tiering that was measured against the outgoing RTX 2000-series product line is believable due to manufacturing process improvement and associated product re-architecture. While concerns regarding power draw and consumption weren’t off base, the reference cooling solutions look to be incredibly capable at addressing the associated thermals for the 320W and 350W TDP designs of the RTX 3080 and RTX 3090 respectively. The announced software and technology additions, which consisted of nVidia Reflex, nVidia Broadcast, and nVidia RTX I/O, further highlight the company’s focus on an optimized experience and maximizing the value that one gets when staying within their ecosystem. The upcoming benchmarks for reference and partner cards will be of particular interest. The ASUS ROG Strix RTX 3090 eschews the space-optimized nVidia 12-pin power connector in favor of three 8-pin connections. While a 350W TDP may not be common in the consumer space from a historical perspective, this power target for the associated performance may still be optimized once the balance of cards are laid out on the table (pun intended). There’s been considerable silence from Team Red on Navi 2x in the PC space. Within the next sixty to ninety days, we’d expect the portfolio for Navi 2x to be communicated. Which segues perfectly to the third player, Intel, who will be sharing info on the DG1 solution in the near future.

Intel’s presentation today was effectively a non-starter. During their prior Architecture Day, announcements for SuperFin (insert Jon Masters singing here), Xe graphics, and additional processes helped provide insight into the company’s direction forward. The formal announcement for Tiger Lake answered very few real-world questions while massaging the messaging to fit an incomplete narrative. As a quick refresher, Steve Burke from GamersNexus had to pay Principled Technologies a visit to correct some egregious errors in the configurations for the AMD platform that competed against the Core i9-9900K. With the process improvements of SuperFin, Intel opted to increase the frequencies of Ice Lake-derived cores while integrating Xe-LP graphics. Costs are unknown, test conditions are unknown, and the alleged performance increases over AMD’s Ryzen mobile APUs reeks of bad marketing at a minimum. The potential for Steve or another person from the press to pay the benchmark execution team a visit isn’t outside off the realm of possibility based on the unproven percentage increases. Wait for the real-world benchmarks from trusted resources. The prior Project Athena effort for developing ultra-portable, power-optimized and performant laptops has now been saddled with the Intel Evo branding. The design objectives are admirable and beneficial, but the fine folks at Intel keep missing the mark in marketing. Announcing SuperFin during Shark Week was followed up with a naming convention that is more commonly associated with Samsung SSD’s and Mitsubishi’s discontinued performance compact car.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.