TweakTown NewsRefine News by Category:
This morning AMD announced the next big advancement concerning their APU technology. AMD heterogeneous Uniform Memory Access (hUMA) is an intelligent computing architecture that enables the CPU, GPU, and other processors to work in harmony from a single piece of silicon in a single pool of memory and seamlessly move task to the best suited processing unit.
This means that in a single application, some calculations will run on the CPU while others run on the GPU accessing the same memory though the same addresses without worrying about which software touched the data last. AMD has been able to achieve this by moving the GPU and CPU onto a single die and then AMD enabled the GPU to have direct access to the CPU memory from the same address space. Finally AMD was able to simplify the data sharing by updating the GPU memory set so that it can follow pointers and complex data structures in the same way that the CPU does. These advancements allowed for better efficiency and lower power consumption.
AMD is touting hUMA as restoring the GPU to the world of Uniform Memory Access. As it sits now the GPU utilizes non-uniform memory access and creates a mass of coding headaches for developers. With hUMA application coding can be simplified, and made more efficient throughout the code base.
VR-Zone has posted up a picture of what appears to be box art for the upcoming Intel line of processors known as Haswell. Intel is really pushing a conceptual computing and creativity aspect this year, so the box design definitely fits within that strategy. But, is it the actual box art?
The box art shown above departs from the typical box art used by Intel. For instance, the top part of the front of the box features a wider case-badge instead of the typical small badge. We likely won't know for sure until Haswell launches at Computex in early June, which we reported earlier today.
After a very successful first quarter for the Snapdragon 600, Qualcomm is moving forward with production on its next-generation system on a chip processor for mobile devices. A Qualcomm exec has confirmed with Engadget that the Snapdragon 800 is expected to move into mass production beginning late May.
Based around the Krait 400 architecture, the Snapdragon 800 is a quad-core beast that could likely see clock speeds as high as 2.3GHz with very low power consumption. An internal Adreno 330 GPU will be capable of 4K playback at 30fps. 802.11 B/G/N Wi-Fi, USB 3.0 and 4G LTE are also built into the chip, alongside a quick charging circuit and voice command capabilities.
Qualcomm is keeping quiet on which handset manufacturers have placed orders and as such, there is no official word, or speculation on which upcoming smartphones will feature the new SoC. We will tune an ear to the market as always and report back once we get word on any upcoming devices featuring the Snapdragon 800.
This morning Intel announced via its Facebook page and Twitter account that it would be unveiling its next-generation Core processor family at Computex on June 4th. Codenamed Haswell, the new processor family is no secret at this point, though.
Haswell architecture is designed to optimize the power savings and performance benefits from the move to FinFET transistors on the improved 22nm process, and feature a fully integrated voltage regulator, moving the power scaling duties from the motherboard and onto the processor itself.
The new king of speed is said to be the Core i7-4770K, which is a quad-core, eight thread 3.9GHz beast. While the middle of the road low power consumption chip is said to be the Core i5-4670T that runs four cores at 3.8GHz.
Our own Anthony Garreffa posted a complete lineup of the upcoming Haswell processors back in December.
Toppc, a member of Chinese overclocking forum Coolaler.com, has managed to get his hands on a pre-release Ivy Bridge-E CPU. This new CPU is the successor to the Sandy Bridge-E line of CPUs on Intel's LGA 2011 platform. It's widely expected to feature up to 12 cores and 30MB of L3 cache.
The results from the comprehensive benchmark suite show that the Ivy Bridge-E CPU is somewhere between 5 and 10 percent faster than the current generation Sandy Bridge-E processors. The benchmark suite included SuperPi mod 1.6, CPU Mark '99, WPrime 1.63, Cinebench 11.5, 3DMark Vantage (CPU score), and 3DMark 06 (CPU score).
The new line of Ivy Bridge-E CPUs are expected to launch towards the end of the second half of this year. They should be mostly drop-in compatible with current LGA 2011 systems, though you'll likely need a BIOS upgrade from your manufacturer.
AMD are definitely doing their best to jump back into the market, as they're really struggling right now. They've won contracts to have various products in all three next-gen consoles, and their Never Settle promotion for Radeon GPUs is doing very well.
In the CPU market however, things aren't so well, as Intel are just dominating too much. Haswell is about to launch, Intel's fourth-generation Core processor, and AMD have, well, nothing - or do they? There are now rumors going around that AMD are to release a super-FX CPU based on the Vishera architecture that makes the FX-8350 come to life, which runs at 4.2GHz.
This new CPU, codenamed 'Centurion', will be released in very limited quantities and would ramp the clock speed up to 5GHz - all on air. There's no details on what voltage or supporting hardware needed to run at this speed - but if the rumors are true, we're looking at around $795 for this limited edition FX CPU.
I think this CPU would have to run rings around Intel's Haswell CPU - especially at $795, which would most likely buy you a Core i7 4770K, a good motherboard and 16GB of RAM in just a few months time.
Intel has detailed overclocking the new Haswell architecture at IDF Beijing and not much has changed from overclocking Ivy Bridge. One major difference between Ivy Bridge and the upcoming Haswell architecture is that Haswell features just one voltage input that can be as high as 3.04v. Ivy Bridge had numerous different voltage inputs that were controlled externally.
The one new rule of thumb for overclocking Haswell is that the input voltage (VCCIN) needs to be at least 400mV higher than the core voltage (VCORE). This yields VCCIN >= VCORE + 400mV.
Other changes include an increased max multiplier--80x vs 63x--though this doesn't mean Haswell will be able to achieve 8Ghz. Intel has also included BCLK ratios that are present on the X79 platform.
There are plenty of other nuances for overclocking the new Haswell CPUs and HWBot has done a great job of detailing all the ins and outs, so head over to source #1 below to check it out!
Intel's 2013 update to their Xeon family isn't being rushed, but it will pack quite a punch when it arrives
With no real competition in the server/workstation space, Intel aren't just rushing out their 2013 update to their Xeon platform of processors. Toward the end of this year, we should expect the Ivy Bridge-based Xeon E5 and E7 processors, while the Xeon Phi MIC/co-processor/accelerator gets segmented into more SKUs and thanks to a new stepping, we're seeing clock speeds pushed from 1.1GHz to 1.238GHz with nearly the same TDP and form factor.
What we will see when the new Xeon processors arrive, is a 12-core part with 30MB of cache and quad-channel DDR3-1866 server memory paths per socket with clock speeds pushing 3GHz and Turbo Boost to boot. Power consumption on these insane 12-core processors should see a helping hand from its 22nm FinFet '3D transistor' process.
Toward the end of the year, we should expect the E5-4600 v2 series, which is the same as the above solution, but 4-socket-enabled where the dual QPI links make a square interconnection layout between the CPU. At around the same time, we should see the 15-core IVB-EX enterprise Xeon processors - the E7-4800 v2 and 8800 v2 - which will replace the Westmere-EX processors we have had for the past couple of years now.
The time is nearly upon us, when we get to see what the fourth-generation Core processor from Intel can do. Haswell, as it's otherwise known, has begun "shipping to customers now and will launch later this quarter" according to a source of CNET's.
Intel is expected to make a statement at the IDF Beijing conference next week, and should arrive in June for Computex in Taipei. Haswell is a special chip, as it is a next-generation processor that will scale incredibly well, from Ultrabooks to smart devices, right up to the high-end desktop and more. Intel CEO, Paul Otellini, has said that Haswell's new micro architecture will deliver "the single largest generation-to-generation battery life improvement in Intel history."
The chipmaker has confirmed the USB bug attached to chipsets, known as errata, will be "in production" during the initial ramp which happens later this month:
4th gen Core is on track for a mid-year launch. Intel issued a PCN documenting a chipset USB errata and stating that chipsets with the errata will be in production during the initial ramp. But Intel has confirmed that there is no chance of data loss or corruption. This issue has only been observed with a small subset of USB SuperSpeed thumb drives and does not affect other USB peripherals. We take all customer issues seriously and should any customer have a question or concern they can always contact Intel customer support.
In a recently released presentation, AMD has hinted that its new Steamroller micro-architecture is on track to be released this year. This comes at a surprise as most industry insiders thought that the new architecture would not arrive until mid to late next year.
In a presentation slide titled "AMD Opteron Technology: Delivering multiple generations of greater functionality and improved performance," it was hinted that AMD will be releasing server-class Opteron chips based on the new technology later this year.
X-Bit Labs is reporting that AMD has pushed forward its plans for Steamroller and the new chips will be compatible with current sockets and that the new server microprocessors will support PCI-E 3.0. Desktop chip's featuring Steamroller technology will most likely not be seen until next year though.