AMD launched the Radeon HD 8970M notebook graphics card on May 15. Built around the GCN architecture, it is hailed as the fastest available notebook graphics solution at the time of launch. It has 1280 shaders, or processing cores, clocked at 850 MHz, with a boosted clock speed of 900 MHz. This gives it a single precision compute capability of 2304 GFLOPS, almost as fast as the mainstream Radeon HD 7870 on the desktop.   [Full story]

Several unreleased processors and co-processors were added to Intel's Material Declaration Data Sheets (MDDS) database in the last day or two. Listed Pentium microprocessors are G2030 and G2140, and Xeon Phi cards are 3120A, 3120P, 7120P and 7120X. The listing doesn't reveal much, besides tray and product part numbers of Pentium CPUs and Xeon Phi co-processors, but it can be considered as a confirmation of their impending release.   [Full story]

The official launch of Haswell microarchitecture in June will include desktop, mobile and server products. Mobile Haswell lineup will be comprised of a few mainstream quad-core processors with TDP 47 and 57 Watt, and a handful of ultra-low voltage parts: Core i3-4100U, i3-4010U, i5-4250U, i5-4200U, i7-4650U, i5-4350U, i7-4550U and i7-4500U. Their specifications are pretty much unknown at this point, and the only thing known about them is the type of integrated GPU.   [Full story]

In March, we reported about AMD plans to release A4-4000 processor in May, that is right before the official launch of "Richland" 6000-series APUs. The A4-4000 was supposed to have new core revision, and was expected to be priced at $40. The processor recently appeared in stock in several online US shops.   [Full story]

AMD introduced first ultra-low power "Kabini" APUs for embedded market last month, and is set to launch consumer "Kabini" products later this quarter. Future ULV processors will employ "Jaguar" 28nm microarchitecture, and have from 2 to 4 CPU cores, paired with HD 8000 series graphics. Details of specific SKUs were sketchy so far, as we only knew model numbers, the number of cores and the size of L2 cache.   [Full story]

We reported in March that Intel plans to launch new Xeon Phi co-processors in Q2 2013. Current Xeon Phi lineup includes 3 models from 5100 series, aimed at high density servers. These cards integrate up to 61 cores, based on Pentium design, and coupled with up to 30.5 MB of L2 cache. Additionally, the cards have 8 GB on-board GDDR5 memory, operating at 5.5 GT/s. In May, Xeon Phi family will be expanded with 5 new products.   [Full story]

An "admin" user at RedQuasar forum recently posted some benchmarks of Core i5-4570 and Core i7-4770K 4th generation processors, built on Haswell microarchitecture. What makes these benchmarks different from previous authorized and unauthorized Haswell tests is that they involved production versions of CPUs, and not engineering samples. The "admin" user also posted pictures of production i5-4570 and i7-4770K chips with SR14E and SR147 S-spec numbers.   [Full story]

In a month, Intel is going to launch Haswell microarchitecture, that will offer better performance and lower power consumption than current generation of Ivy Bridge CPUs. One of the drawbacks of new architecture is incompatibility with existing socket 1155 motherboards. Fortunately, Intel still have plans to make LGA1155 processors, and a handful of them will be released in June. We reported their specifications in February and March of this year.   [Full story]

The latest reports from Fudzilla indicate that the first 700 series GeForce graphics cards from NVidia will be launched later this month, and that the series will have a top down launch schedule. The first to market will be the GK110 based GTX 780, launching on May 23. It will be slower than the Titan, but will be known internally as the Titan LE. Next up will be the GTX 770 on May 30, based on the GK104-425 core, an improved GK104 GPU.   [Full story]

First two generations of Accelerating Processing Units were positioned by AMD as inexpensive processors with good enough all-around performance, whether it be CPU calculations, graphics or general purpose computing. AMD also released cheaper versions of APUs with disabled integrated graphics. These parts were marketed as Athlon X2 or Athlon X4, depending on the number of cores. Next generation of desktop APUs, codenamed "Richland", will also have at least one SKU with disabled graphics.   [Full story]

In January AMD announced ultra-low power Temash APUs for tablets and ultra-thin notebooks. The chips will make use of recently launched Jaguar architecture, and they are expected in Q2 2013. Some specifications of future Temash APUs were revealed by AMD in the announcement, and in subsequent roadmap updates. Dual-core microprocessor received A4-1200 model number. This part is clocked at 1 GHz, and has HD 8180 graphics. The second processor is a quad-core A6-1450.   [Full story]

We had a quite a few stories dedicated to Atom "Silvermont" architecture and products, that are going to replace current generation of Atom "Salwell" CPUs. Manufactured on the latest 22nm technology, the microarchitecture will bring such features as out-of-order execution, newer instructions, up to 4 cores, and Ivy Bridge graphics to Atom family. Depending on targeted market, there will be a few different variations of "Silvermont" processors.   [Full story]

Details of a new GCN based entry level graphics card from AMD have surfaced around the web. Named the Radeon HD 7730, it is the slowest and cheapest GCN based card to date, and is believed to be based on the Cape Verde GPU. The reference card is a single slot design, and sports one DVI port, one HDMI and one DisplayPort outputs. It is 168 mm long, and has a TDP of 55 W.   [Full story]

Earlier this week AMD announced new processors, expanding the line of high-end FX CPUs with quad-core FX-4350 and six-core FX-6350. It's not uncommon for AMD to lower prices when launching new products, and this time it was not an exception. Based on the latest official AMD price list, the company cut prices of ten A-Series and FX-Series chips by as much as 19%.   [Full story]

Next generation of extreme Intel microprocessors is coming later this year. The CPUs will be based on Ivy Bridge-E core, that brings all performance enhancement features of server products to desktop market. Branded as Core i7 Extreme or Core i7, these processors will offer up to 6 Hyper-Threaded CPU cores, up to 15 MB L3 cache, PCI Express 3.0 interface, and quad-channel memory controller, supporting memory with 1866 MHz data rates.   [Full story]

AMD FX microprocessors, based on "Vishera" core, were available since last October. "Vishera" CPUs use the latest "Piledriver" architecture, and come with 2, 3 or 4 modules, which corresponds to 4, 6 or 8 cores. The chips also have up to 16 MB total cache, and support the latest instructions, including AVX and FMA3. So far AMD had 4 different retail models, and 2 other SKUs were shipped in computers only. Yesterday AMD refreshed FX family with two new products, FX-4350 and FX-6350.   [Full story]

We previously reported that Intel plans to launch a few Ivy Bridge-based Celeron and Pentium mobile processors before the families are transitioned to Haswell microarchitecture. New CPUs, Celeron 1005M, 1017U, and Pentium 2127U, will offer slightly better performance than Celeron 1000M, 1007U and Pentium 2117U products. The details of future microprocessors were confirmed last week by a recently updated specification update for Mobile 3rd Generation Intel Core Family.   [Full story]

Intel this week revealed a launch date of their next microarchitecture, codenamed Haswell. In a blog post, titled "4th Generation Intel Core Processors Coming Soon", they stated that first Haswell CPUs will be released in about 3,337,200,000,000,000 nanoseconds. This huge number is another way to say 3,337,200 seconds, or 38 days and 15 hours.   [Full story]

AMD this week officially launched the Radeon HD 7990 graphics card, using a pair of Tahiti XT GPUs. As the first cards with HD 7990 part number were introduced around 8 months ago, this seems like a marketing ploy, although it also opens the door for more AIB manufacturers to release dual-GPU GCN based products. Of note is the fact that with the launch, AMD are bundling cards with 8 games - many of them are popular, but none will fully tax the card.   [Full story]

As we're getting closer to the launch of desktop "Richland" APUs, they started to appear in OEM desktop PCs from major manufacturers. Lenovo recently added new product pages for H535 and H535s computers to their UK website. Both computers list current generation of "Trinity" processors, along with upcoming "Richland" APUs.   [Full story]

Haswell processors are not the only forth-generation desktop products to be launched this year. In the second half 2013, Intel will introduce high-end Core i7-4xxx CPUs for LGA2011 platform. These i7 microprocessors will use Ivy Bridge-E core, while keeping most of the Sandy Brdge-E features unchanged. That includes 4 or 6 CPU cores, up to 15 MB L3 cache, quad-channel memory controller, and support all SIMD instructions up to AVX.   [Full story]

Intel's Haswell microarchitecture, coming in about 6 weeks, is going to bring new performance improvements to desktop Core i5/i7 lineup. Unlike Ivy Bridge CPUs, that achieved this by tweaking the microarchitecture and raising clock speeds, Haswell microprocessors will get up to 10% better CPU performance owing just to architecture redesign. The specs of Haswell microprocessors will be identical to their Ivy Bridge counterparts, and it seems that the pricing will not change much.   [Full story]

In June, AMD is going to launch the third generation of desktop APUs, featuring "Piledriver" microarchitecture and HD 8000 series graphics. Specifications of upcoming A-Series consumer desktop processors were known for a while, but it's not well known that AMD will also release several business class SKUs, based on "Richland" core. Some details of these models appeared in CPU support lists on the Biostar website.   [Full story]

Last week Hewlett Packard launched Moonshot system and the first Moonshot server cartridges, built with Atom S1200 series processors and aimed at web hosting market. A single system with 45 of such cartridges can run up to 180 threads simultaneously. Later this year HP will introduce Avoton-based cartridges, that will double the number of threads, and improve server's performance. The Moonshot system is not going to be limited to Intel CPUs, though.   [Full story]

Last week we reported on two mobile Celeron microprocessors, 1005M and 1017U, currently scheduled to launch in the third quarter. What is interesting is that both processors are currently listed under H2 2012 in the latest Intel's public roadmap. The document is obviously incorrect because both processors are not launched yet. In addition to two Celerons, the roadmap also mentions Pentium 2127U.   [Full story]

I added benchmark pages for all tested processors from our CPU database. The pages currently show processor's multi-threaded performance in comparison with other Intel and AMD CPUs. For instance, see how AMD FX-6350 benchmarks page looks like. The benchmarks pages will be expanded in the future with more information.

I've made changes to CPU pages, that should make easier for first time visitors to understand various tools and links, that we have on those pages. I added brief explanations to several buttons in the "Specifications" tab, moved them into new "Tools" section, and styled links and tools in the "Benchmarks" and "CPUID" tabs in the same way. I also tweaked "News" tab design. This is how it looks now on the Pentium G640 page.

Many CPU comparison page now compare not only performance, TDP and price of microprocessors, but also some of their features, like EM64T, virtualization and Execute disable bit. Please see how it looks on the Core 2 Duo T5600 vs Core Duo T2050 page.