It’s not yet time to pen a full, in-depth comparison between Intel’s forthcoming Kaby Lake chipsets, including Z270, H270, and whatever may become of the lower-end H- and B- lines. There’s still data we’re waiting on, and won’t have access to for a little while yet. Still, some preliminary Z270 & H270 chipset specs have been reported by Benchlife, including information on PCI-e lane count and HSIO lanes. This coverage follows the same format as our Z170 vs. H170, H110, B150, & Q150 differences article.
If the early information is to be believed, the Kaby Lake-ready platform primarily focuses its efforts on largely minor improvements, like additional HSIO lanes to support a burgeoning PCI-e-enabled SSD market. Z270 will move from Z170’s 26 HSIO (High-Speed I/O) lanes to 30 HSIO lanes, providing an additional 4 lanes for M.2 and PCI-e AICs (add-in cards). H270, meanwhile, will move from H170’s 22 lanes up to parity with the Z-series platform, also hosting 30 HSIO lanes. The additional lanes fall into the category of “general purpose” PCI-e lanes, resulting in the following configuration:
Ask GN is a little bit more personable than our usual reporting, given the nature of the Q&A style of content. Taking that as an opportunity to update everyone on our server migration, we’ve finally moved; it was a success. The new server has been operational for a few days now, and it’s faster, more stable (seriously – 0 downtime, whereas the other began crashing every few hours), and will soon™ be home to an updated front-end for the website. We will be rolling-out small changes over the next few weeks, and those will eventually lead to a full refresh.
Anyway, Ask GN was the original topic. We’re on episode 36, and this week addresses questions of DirectX 12 scalability, GameWorks & GPUOpen efforts, electrostatic discharge and management of ESD, and more.
Leading into Black Friday and Cyber Monday, arguably America's most worshiped holidays, we're launching our next installment of Ask GN: Episode 35, which talks CPU socket technology, GPU silicon manufacturing (and what happens to failed chips), synthetic benchmarking and its variance, and super-sampled versus native resolution 4K gaming.
As we've expanded the Ask GN series, we've started advancing its collection and distribution of information. A few of the questions from this week's episode were passed through our contacts at AMD and nVidia, and next week is shaping up to hold similar Q&A as passed through manufacturing contacts. We're also working with some game developers to better answer a couple of CPU threading questions, as it seems a good opportunity for everyone to learn -- GN staff included.
Here's the latest episode:
This week's news recap segment features updates from the Super Computing conference 2016, including updates to AMD's GPUOpen and Boltzmann initiatives (ROCm, HIP), an Intel Xeon refresh, and Intel's investment in self-driving cars. Outside of SC16, we also have news pertaining to Lian-Li's (finally) shipping PC-O10 case and Thermalright's AXP-100H Muscle cooler.
The only rumor in this news segment is that of AMD's Summit Ridge naming scheme, which Chip Hell suggested will be branded with numerical 3-5-7 suffixes, similar to Intel's CPU branding. Beyond an allegedly leaked slide, there's no way to validate this rumor -- so take it for what it's worth. It's likely that we'll find out more about Zen at the time of CES, or shortly thereafter. That tends to be when Intel and AMD make some CPU / architecture announcements.
The PC-O10 features a binal design composed of two chambers aimed at isolating heat generating components, comparable in form to other models like the PC-O8 and PC-O9. The PC-O10 exhibits a striking resemblance to the PC-Q37, so much so that the *-O10 appears to be a larger, more rectangular version aimed at accommodating the ATX form factor motherboard. Motherboard support is broad, with Micro-ATX, ATX, and E-ATX all being able to fit in the chassis.
This episode of Ask GN addresses reader and viewer questions relating to boost technologies for GPUs (DPM states and GPU Boost), "game mode" for monitors, and a couple questions related to CPU benchmarking. We talk loose plans for Zen tests and scalability of the 2500/2600K Sandy Bridge CPUs in the modern era. Even Nehalem got a few mentions.
Monitor "game modes" presented a topic with which we're not intimately familiar, but some research did grant us enough information to hopefully answer the question in a helpful fashion. The rest, like the boosting functionality on GPUs, is stuff that we've discussed on-and-off in review articles for several months -- it's just now laid-out in a quick Ask GN video.
Corsair recently announced two new additions to their peripherals lineup: the HARPOON RGB mouse and the K55 RGB keyboard, priced to appeal to gamers on a budget. This follows competitor Logitech's recent release of the Prodigy series, also targeted at entry-level gamers.
Corsair's Harpoon is purchasable right now, while the K55 will be available starting November 22.
High-end monitors are really starting to get pumped-out now, it seems. This generation of ~$250+ GPUs supports resolutions of 1440p with relative ease, and UltraWide displays are proliferating on the market to popularize the 21:9 aspect ratio.
Episode 33 of Ask GN answers a few more questions than we normally take, resulting in a 20-minute run-time for the episode. We talk about hyperthreading and its impact in some specific games, DirectX 12 performance, scaling of Dx11 and Dx12 performance on AMD and nVidia GPUs, Furmark, and a few more topics. One of those other topics was on recording software's impact on framerate, like Shadowplay, OBS, FRAPS, or GVR.
You can find the embedded Q&A video below. We've been running these for a bit more than a year now, making Ask GN our longest-running series by a longshot. As for content this week, because we often take Ask GN articles as an opportunity to discuss what's coming, keep an eye out for more EVGA VRM updates, BF1's performance with various RAM speeds, and a couple smaller items.
The GPU news is probably the most interesting. AMD looks to be positioning an "RX 470D," also called "RX 465" and "RX 470 SE," to compete more directly with the GTX 1050 Ti. Our GTX 1050 & 1050 Ti review noted that the 1050 gives the RX 460 a tough fight, but that the RX 470 handily outpaces the 1050 Ti in all tested scenarios. The only problem, as always, is the price gap -- it's a $30 jump from entry-level GTX 1050 Ti cards to the entry-level RX 470 cards. That's where the 470D is supposed to land, and should fight the 1050 Ti directly.
Video below for the news discussion, or find the script below that:
We moderate comments on a ~24~48 hour cycle. There will be some delay after submitting a comment.