It seems like Nvidia announces the fastest GPU in history multiple times a year, and that’s exactly what’s happened again today; the Titan V is “the most powerful PC GPU ever created,” in Nvidia’s words. It represents a more significant leap than most products that have made that claim, however, as it’s the first consumer-grade GPU based around Nvidia’s new Volta architecture.
Check out @Smith7044’s Tweet: https://twitter.com/Smith7044/status/893224528264015874?s=09
AMD today launched Ryzen Pro (styled “PRO” in AMD’s branding, but we’re not going to do that here), a series of processors designed for the corporate desktop. Close counterparts to the existing line of consumer-oriented Ryzen chips, the Pro parts are aimed at Intel’s vPro-compatible processors, which enable a number of additional administrative, security, and management capabilities.
Most of the regular Ryzen models have corresponding Pro versions, albeit topping out at a 1700X rather than the 1800 and 1800X of the consumer parts. This means that at the high end, there’s a couple of eight core, 16 thread parts, which is twice the number of cores and threads of comparable Intel chips.
In general, each Ryzen numbered series approximately corresponds to Intel’s numbering—Ryzen 7 against Core i7, and so on—but AMD is claiming some advantages over Intel. In particular, Intel doesn’t make any vPro-enabled i3-series processors. Even if your computational needs are modest, and an i3 is otherwise sufficient, you’ll have to step up to at least an i5 if you want vPro capabilities. Not so with Ryzen Pro; at the low end, AMD has a pair of Ryzen 3 Pro chips with just four cores and four threads each, operating at the same kind of price point as the two core, four thread i3, but with the same management and security capabilities as other Pro parts. Ryzen 3 Pros also have some amount of turbo boosting, which i3s don’t.
When pushed on what exactly those Pro features are, however, AMD is unclear. In the slides it prepared for the media, AMD listed features such as memory encryption, Secure Boot, and a firmware-integrated TPM as Ryzen Pro features. This is peculiar, though, because the regular Ryzen chips support Secure Boot and a firmware TPM. They also purport to support memory encryption, and at least some motherboard firmwares for Ryzen chips have settings for turning on this memory encryption. As such, these features don’t appear to be unique to Ryzen Pro.
The bigger difference seems to be in non-silicon features. The Pro CPUs have 3 year warranties, instead of the 12 months of the consumer parts; AMD also promises to manufacture the Pro parts for at least 24 months. This makes it much easier for enterprises to standardize on particular system configurations, safe in the knowledge that they’ll continue to be able to buy identical systems even over multi-year deployments.
AMD’s relative lack of product segmentation—extending the full range of features all the way down the line to the lowliest Ryzen 3 Pro—is certainly a welcome contrast to Intel’s segmented line-up and forced upsells. But there’s a big hurdle for the company to overcome in the corporate desktop space, especially when it comes to low-end parts: none of these chips have integrated GPUs. While the high-end Ryzen 7 Pro parts are likely to be used for high-end workloads with discrete GPUs, the mid-range Ryzen 5 Pros and low-end Ryzen 3 Pros are much less likely to justify that additional expense. Intel’s competing chips, in contrast, all come with an integrated GPU, enabling a range of small form factors, reduced power consumption, and lower bills of materials.
Ryzen chips with integrated GPUs will materialize eventually, but until they do, we feel that Ryzen Pro has something of an uphill battle ahead of it. Ryzen Pro systems should ship in the second half of this year; Ryzen Pro mobile parts will be released in the first half of 2018, with a full reveal due in late August.
After the leaks, the announcements, and the controversy, the Core i9 reviews we’ve been waiting for are here. The embargo has dropped on the gaming hardware sites that have been putting Intel’s newest processors through the ringer, and we can finally see that it’s not just all one elaborate Intel PR move.
Microsoft Xbox One X has been announced at E3 2017. We already know the internal specifications of the Xbox One X, which puts the console in direct competition with Sony’s PlayStation 4 Pro. So which console are you planning to buy?
First off, this is going to become quite confusing for people less familiar with computer hardware. Intel has Z270, AMD has X370, and now Intel is going to have X299, and AMD is going to have X399/X390. But I digress.
X299 is a bit interesting as it’s going to have two separate ranges of CPU’s: Skylake-X and Kaby Lake-X.
Honestly the less interesting of the two, Kaby Lake-X will be limited to 4 cores and 8 threads, so basically you’re not getting an improvement over the standard Kaby Lake platform. In fact, it’s limited to 16 PCIe lanes and dual-channel memory, so besides maybe some extra USB and SATA ports versus the Z270 platform, I really don’t see what purpose Kaby Lake-X has. Maybe it’ll just have more power available to it for more stable overclocks? I’m not really sure.
Skylake-X, on the other hand, will be the next HEDT platform, and will now be expanded to include the first consumer 12 core, 24 thread CPU. It’ll also feature up to 44 PCIe lanes and quad-channel memory.
Of course the big question is price. I’d expect that the Kaby Lake-X chips would start around $350, because otherwise if they’re more expensive you might as well just get a 7700K. As for Skylake-X, who knows?
Taking past specs and pricing into account, I’d maybe expect the 7800K to be a 6-core for around $400, the 7850K to be an 8-core for around $600-$650, the 7900K to be a 10-core for around $1,000 and the 7950X to be a 12-core for around $1,700, following the 6000-series pricing.
Of course, Ryzen may convince Intel to drop their prices, considering that the 1700X and 1800X can obtain comparable performance to a 6900K for less than half the price. In that case, the prospective 7850K may drop to only around $500 to compete with the 1800X, but I honestly expect the 7900K and 7950X to stay at high pricing just because that’s what Intel has always done with their ultra-high end products. Although if AMD comes out with 10 or 12-core options on their X390/X399 platform that may influence Intel to drop the prices of those as well.
Intel is planning to get rid of the royalties it charges third-party chipmakers that use its Thunderbolt 3 specification sometime next year, according to a report from Wired, which could make it easier for hardware manufacturers to use Intel’s data transfer specification.
USB-C is a complicated specification — just because a port is physically a USB-C port, doesn’t mean it’s got the same technology driving it on the inside. One of those standards is Intel’s Thunderbolt 3, the current generation of Intel’s ultra-fast data transfer system, which switched from using the Mini DisplayPort connection to USB-C.
Thunderbolt 3 is one of the most versatile USB-C solutions out there, with fast speeds and charging, but Intel’s licensing fees also meant that its more expensive to use. By opening up the protocol, Intel hopes to encourage “broader adoption in the ecosystem, with a lot of different peripherals and other devices,” Jason Ziller, who leads Intel’s Thunderbolt development, tells Wired. Intel is also reportedly working to integrate Thunderbolt 3 into future CPUs, which should also help adoption.
The high-end desktop market looks set for a massive showdown this summer between Intel and AMD. The former is set to release its new X299 platform and the latest rumors point at up to six Core i9 and Core i7 CPUs sporting up to 12 cores as it looks set to replace its current X99 platform
AMD on the other hand had already shown part of its hand, by announcing Threadripper – an entirely new range of CPUs based on its Zen microarchitecture that will sit above the Ryzen 7 series. These will have up to 16 cores and if current rumors are anything to go by, AMD is planning a top to bottom range with nine CPUs with 10, 12, 14 and 16 cores.
Traditionally, these high-end desktop platforms are used as workstations and for multi-threaded performance but do have the benefit of supporting multiple GPUs with extra bandwidth too for high-resolution gaming. This could mean that Intel will have a real fight on its hands with AMD already offering excellent value in multi-threaded applications with its Ryzen CPUs, which have offered similar performance to Intel but for much lower costs.
One area AMD has struggled is with low-resolution gaming, but that’s not likely to be a target for owners of high-end desktop systems so Threadripper could be even more disruptive here than Ryzen was in the mid-range.
Several websites have leaked data regarding the new CPU ranges and you can see all the supposed new CPUs listed below. Hexus pointed at a post on Anandtech that claims to list the entire Skylake-X or Core i9 CPU line-up. Meanwhile Wccftech cites https://en.wikichip.org as the source for its list of Threadripper 10, 12, 14 and 16-core CPUs from AMD.
All information is unconfirmed at this point so take it with a generous pinch of salt.