3 ms·
This blog post is just an outline of the Nvidia PR. I don't know how anyone can claim it will "Push the Limit of Gaming Graphics" when there aren't even benchma
by css 8y ago
This blog post is just an outline of the Nvidia PR. I don't know how anyone can claim it will "Push the Limit of Gaming Graphics" when there aren't even benchmarks. No one knows how these cards will perform.
- Rotdhizon 8y agoThis x100. It's weird how so many sites, tomshardware being the most infamous of this recently, are trying to push the hype on these cards with so little information. Take a peep at the comments on this article below --- https://www.tomshardware.com/news/nvidia-rtx-gpus-worth-the-money,37689.html https://www.tomshardware.com/news/nvidia-rtx-gpus-worth-the-...
- css 8y agoIt's not unique to TomsHardware, though they are a prominent offender. Look up "reviews" for any tech product, almost every article will be mostly copy and paste from PR. This is on purpose: any marketing agency worth their salt will have relationships with major publishers to get content promoted and editorialized. It even happens with indie people with large social media followings. See the agency Social Native, for example. tl;dr everything is astroturfed
- Asooka 8y agoPeople with engineering samples do know, so maybe bloomberg have been in contact with one? Also see the wccftech leaks. These cards seem to perform close to nVidia's own estimates so far. This isn't to say that they aren't way pricier than the current 10x0 equivalents - you're paying twice the price for (generously) 150% the performance in gaming benchmarks. On the other hand, the ray-tracing core is quite important for anyone doing GPU rendering and the 2080(ti) should be a hit with architects, designers, animators etc - at least those who can't afford a Quadro (there is an entire order of magnitude difference in price). They also have tensor cores that can carry out low-precision matrix mathematics with good speed, which is important for neural networks. All these cores are also all on one chip and share memory, so the results of one can be used by the other with no costly memory copies. There are workloads where these cards will be twice as fast as the current ones. You'll have to wait for benchmarks and see if game studios and software vendors can adapt to all the new features present. I'll just point out that Microsoft unveiling a ray-tracing API for DirectX wasn't entirely just their own idea and showing it off not long before the RTX was announced is no coincidence.
- css 8y ago> These cards seem to perform close to nVidia's own estimates so far Considering they provided none to benchmark gaming, I am not sure what you mean here. > You'll have to wait for benchmarks and see if game studios and software vendors can adapt to all the new features present. The title is that the cards will "Push the Limit of Gaming Graphics" so this is the only relevant claim here. If it is anything like DX12, we won't see meaningful differences for a long time. Every game I have that "supports" DX12 (i.e. flagship Rise of the Tomb Raider) does not work with DX12 without artifacting problems [0][1]. [0] https://forums.geforce.com/default/topic/1026296/geforce-drivers/rise-of-the-tomb-raider-dx12-graphical-glitch-on-gtx-1070-fe-on-387-92-driver/ https://forums.geforce.com/default/topic/1026296/geforce-dri... [1] https://steamcommunity.com/app/391220/discussions/0/451852571441880358/#p1 https://steamcommunity.com/app/391220/discussions/0/45185257...