XLR8 Gaming EPIC-X RGB DDR4 4200MHz, 4400MHz, and 4600MHz Desktop Memory While pricing is still unknown, availability will be through Amazon, Best Buy, or directly through PNY’s online sales. Each kit features RGB lighting, CAS 18 or 19 Latency, and a limited lifetime warranty. The Epic-X RGB will feature 2×8 GB (16 GB) ram kits ranging from 4000 MHz up to 4600 MHz. This is the first of the RTX 40-series to use the older style of power connector, though reportedly only the reference clocked versions have been given the option to switch from a 16-pin connector.PNY has just announced the addition of the EPIC-X RGB series into its already successful XLR8 Desktop memory. There is one key difference, that's mostly key because of recent fiery issues, and that's the use of an 8-pin PCIe power connector instead of the 16-pin version Nvidia still favours for its cards. ![]() It's a subtle wee graphics card that will happily chew through your games without demanding undue attention. That does mean it draws less power, but the simple dual-fan cooler isn't quite as effective as the Founders' shroud, though it's not exactly hot. In all honesty, though, it's so small as to be largely insignificant in the grand scheme of gaming things generally just a few frames per second here or there. The Founders card actually runs around 100MHz quicker than the chip inside the PNY card, and that translates into a small frame rate lead. The Founders cards still remain the best versions of Nvidia's reference GPUs-they look the best, and they perform the best, even if they're not specifically overclocked. ❌ You're into card aesthetics: The PNY reference RTX 4070 is a basic card, with a black plastic shroud, and is certainly not a flashy card you're going to want to show off in a presentation case.Īs a reference card, the PNY RTX 4070 is capable of delivering similar levels of gaming performance to the Nvidia RTX 4070 Founders Edition. ❌ You already own a decent RTX 30- or RX 6800-series GPU: At ~20% higher performance than the RTX 3070 Ti, and similar performance to the RX 6800 XT, spending another $600 so soon isn't worth it just for Frame Generation. Which makes it a potential performance uplift over the previous Ampere generation, rather than something that can be completely relied upon.īut when Frame Generation works, it really gives the RTX 4070 some serious 4K ray traced gaming chops. Still, with DLSS in play as well there's not as much traditional rendering per frame as there used to be.įrame Generation has to be added on a per-game basis, however, so not every game has it, or will ever be upgraded to include it. Well, not quite from scratch, but based on the frames around them. It's an unprecedented bit of AI tech magic that takes advantage of Ada's Optical Flow Accelerator hardware to create entire game frames from scratch. That last one being the key differentiator and the feature which promises to eventually push it past the RTX 3080 in gaming terms. It being an RTX 40-series GPU you also gain access to DLSS 3.0 and Frame Generation. Storage: 1TB WD Black SN850, 4TB Sabrent Rocket 4Q RAM: 32GB G.Skill Trident Z5 RGB DDR5-5600 Sometimes a little more, occasionally a little less. So, alongside nine times the amount of L2 cache attached to the GPU, those fewer CUDA cores are actually able to deliver pretty much the same level of performance. It's also running almost a full 1GHz quicker under gaming load. This is TSMC's 4nm process combined with Ada's efficiency, however, and that means there are billions more transistors in the package despite it being less than half the size of the RTX 3080's Ampere GPU. It also has a wider memory bus-320-bit against a 192-bit bus-and therefore a whole lot more memory bandwidth. ![]() That classic GeForce GPU comes with 8,704 CUDA cores, 22 more RT Cores, and more ROPs. Ampere is highlighted by looking at the core configuration of the RTX 4070 vs. That all means you're looking at a card whose gaming performance is maybe 20 - 25% off the Ti version.īut the generational improvement of Ada vs. You are, however, still getting the same 12GB GDDR6X memory configuration. That means you're getting 5,888 CUDA cores instead of the full chip's 7,680, you also get fewer RT and Tensor Cores. The AD104 GPU inside the RTX 4070 is the same as that used in the RTX 4070 Ti, just minus a few SMs. And hey, it's got a metal backplate and some visibly chonky copper heatpipes in action. So, while I was a little concerned the courier had wandered off with my card when he handed over the suspiciously lightweight box, I'm more than happy with the simplistic design choice PNY has taken with its reference card.
0 Comments
Leave a Reply. |