Processors - Page 120

ARM has unveiled a new chip that is set to enable more efficient and affordable mobile processors. Cortex A7 is based on ARM's 28nm fabrication and reportedly consumes five times less power and measured just one-fifth the size of the 45nm Cortex A8, which is found inside Apple's A4, Samsung's Hummingbird and Texas Instruments' OMAP3.

ARM will compliment the efficiency by pairing the Cortex A7 with quicker, more power hungry processing cores. The A7 is set to pave the way for sub-$100 smartphones, which will boost adoption rates in developing regions. ARM CEO, Warren East told the BBC:

Given the ill received launch of AMD's FX (Bulldozer) CPU family no thanks to the lack of what mattered most, performance, it's hard to imagine how AMD can do much to rise above anytime soon, or even maintain the foothold they had in the market previously in terms of offering a more aggressive price/performance alternative.

However, it's looking like AMD still may have another playing card up its sleeve yet; this in the form of a new (B3) revision of the chips being worked on. A mention of the new stepping was spotted in a public AMD document entitled "BIOS and Kernel's Developers Guide (BKDG) for AMD Family 15h Models 00h-0Fh Processors"

This is potentially a good sign if what we saw from the Phenom launch is anything to go by. Many of you would recall When the first B2 stepping processors in the Phenom lineup were found to be flawed by the TLB (translation lookaside buffer) issue, which incurred quite the performance hindrance of at least 10%. It wasn't until AMD rectified the issue with the launch of the B3 stepping Phenoms four months later that Phenom became a lot more attractive.

Thanks to a chunky roadmap leak via a Chinese forum earlier today, we can gather up some more spicey details on Intel's upcoming Ivy Bridge CPUs and associating platforms; also removing some of the haze thats still been surrounding it in terms of TDP ratings and backward/forward compatibility characteristics.

If said roadmaps are the real mcCoy, we can look forward to a maximum TDP of just 77W for the high-end models which is a good drop lower than the 95W peak for Sandy Bridge thanks to the brand new 22nm fab process. The diagram above seems like a bit of a dogs breakfast at first, but if you focus primarily on the upper region that's where the more interesting tid bits lie.

We can ascertain from this chart that just like Sandy Bridge, there will be unlocked 'K' series SKUs for Ivy Bridge as well; these of which top out at the aforementioned TDP of just 77W. Looking further down the chart, there are entries for quad core 65W 'S' and 45W 'T' SKUs not unlike Sandy Bridge, and also an SKU entry for a 35W dual core.

Even amongst all the lawsuits, bitching and injunctions, Samsung are still producing the A6 quad-core mobile chip for future Apple iOS-based devices. It was previously reported that Apple had signed a deal with Taiwan Semiconductor Manufacturing Company (TSMC) last month to supply A6 chips, but that deal seems to be on shaky grounds now. The Korean Times is now reporting that Apple and Samsung have maintained their buyer and supplier relationship and will do so going forward for the foreseeable future.

The Korean Times cites an executive from an Apple parts supplier in Korea, where it says Apple have concluded that Samsung is a critical business partner and can't sever times with them just yet. Samsung is already producing the A6 processor for future iOS devices at a manufacturing plant based in Austin, Texas which opened back in 2007. TSMC will still manufacturer chips for Apple, but the volume is expected to be very low.

The reason behind this is that TSMC haven't stabilized their manufacturing process and Apple isn't willing to take the risk of switching suppliers under such circumstances. What of the legal issues between Apple and Samsung? The article does point out that during an interview with Samsung CEO Choi Gee-sung, in which he told reporters that Samsung is handling legal issues with Apple independently of their business relationship with Apple.

While many will take these numbers with a grain of salt since they came direct from Intel, it's not hard to believe how strong performing Intel's Sandy E is up against its old aged X58/LGA1366 predecessor, and with the (quite disappointing) benchmarks out on Bulldozer, it only puts AMD in an even dimmer light.

Intel has shared some internal test results of its soon to be launched Core i7 3960X Extreme Edition processor; a six cored / 12 threaded beast clocked in at 3.3GHz with 15MB of cache. The results were directly compared to the top dog in Intel's current flagship platform, the Core i7 990X EE, which runs a stock clock of 3.46GHz.

A breakdown of all the results can be read via the source link below, but to summarize here, the i7 3960X looks to be about 15% quicker on average. Its best improvement was actually seen in 3DMark 11's Physics test where it bested the Core i7 990X by 36%. And throughout the tests there was one particular standout attribute; its memory benchmarks where it left the 990X for dead over and over and scored up to 111% better in some instances.

Intel is currently staying quiet on whether there will be a new Core i7 introduced to combat the impending launch of AMD's Bulldozer, but it seems there is going to be a bigger, badder brother released in the LGA1155 lineup. It is rumored that Intel is going to drop the 2700K onto the market, which will be a minor clock speed increase over the already available Intel Core i7 2600K. The 2600K is a 3.4GHz part with the 2700K set to be a 3.5GHz part, with Turbo enabled those speeds will be 3.8 and 3.9GHz, respectively.

The 2600K known is for its super high overclocking ability, with 4.5GHz+ achievable with decent air cooling, so the 2700K should achieve the same type of overclock as its just a 2600K with an extra 100MHz in the trunk. The conservative speed bump will keep the gates open for a future 2800K, as Intel need to keep a space between regular Sandy Bridge and the Extreme chips. We should soon what happens, but there are already people who have cranked up an engineering sample of the 2700K to an impressive 5GHz on air cooling.

Impressive? It's impressive, with a silent K.

Samsung's latest Exynos 4210 SoC is a great piece of tech, it powers their super smartphone, the Galaxy S II and is a very, very fast phone. The 4210 sports two ARM Cortex A9 cores running at 1.2GHz along with an ARM Mali-400 MP4 GPU. The A9 chip is equal to that of TI's outing, and a bit better than NVIDIA's Tegra 2. In various tests, the Mali-400 MP4 beats out all other smartphone GPUs that have been tested.

Samsung have today announced a 32nm addition to the family, the Exynos 4212. CPU clock speeds get a nice bump up to 1.5GHz, but GPU clocks are unknown. Samsung is however claiming that GPU performance improves by up to 50% "over the previous processor generation". The move to 32nm could have allowed a 50% increase in clock speed. Samsung's 32nm process uses high-K + metal gate transistors, similar to Intel's transition to 45nm.

According to Samsung, these new transistors allow for either a 40% (or more) increase in performance at the same leakage or a 10x reduction in leakage current at the same speed. Density also improves by 2x enabling even more complex chip designs (more cores, bigger GPUs, more integration) or smaller die with the same feature set (which will eventually improve supply and profit). The Exynos 4212, Samsung is claiming a 30% reduction in power compared to its predecessor.

AMD's Bulldozer is set to enter the construction site shortly, and these new benchmarks are leaned heavily in their favor. It looks as though a set of leaked press slides of the AMD FX Processor found their way into the grubby hands of DonanimHaber ahead of the official launch with AMD claiming huge performance leads over rival, Intel. AMD claims that their FX-8150 processor is aiming to take down Intel's Core i7-980X in game tests and even passing it in some DirectX 11 titles.

It's performing on the same level as the Core i7 2600K in several popular CPU benchmarks such as WinRAR 4, X.264 pass 2, Handbrake, 7Zip, POV Ray 3.7, ABBYY OCR, wPrime 32M, and Bibble 5.0. The AMD FX-8150 benefits greatly from the FMA4 instruction set that is lacking in Intel's Sandy Bridge. In the OCL Performance Mandelbrot test, the FX-8150 thrashes the 2600K by as much as 70%.

But, performance isn't everything to a lot of people. Considering that the FX-8150 is confirmed to hover at around $250, we should see massive love for AMD's new Bulldozer processor.

After recently hearing of Intel's plans to release a new, top end entry in its LGA1155 Core i7 line, the i7-2700K, new information has come out not only confirming its soon to be existence, but also covering pricing and availability.

Where Intel usually has a tendancy to roll in new top end parts at the same price point as what the predecessor sat at, this time things are expected to be different and apparently the i7-2700K will carry a higher price tag than the 2600K. While the exact price isn't allowed to be revealed at this time, a source has made it clear the i7-2700K will in fact be significantly higher than the 2600K's current $317 price tag - most likely somewhere around the ball park of US$340 to $350.

The i7-2700K will run at 3.5GHz out of the box, a typical 100MHz bump over the previous top model (2600K), while the chip is otherwise pretty much identical, including the multiplier with the highest available setting being 57x. On the upside, it's been rumoured that Intel is cherry picking its best Sandy Bridge cores for 2700K allocation, and this is backed up by other reports that it is one hell of an impressive overclocker.

IDF 2011 - Intel is committed to low-power computing and to show their commitment to it, Intel CEO Paul Otellini challenged the Intel Labs to create a demo of just how far Intel can push technology at this current moment in time. Paul added "the ultimate goal is the most power-efficient devices known to man". Intel Labs were able to create a CPU that was capable of running Windows, powered by a tiny solar cell.

An Intel engineer stepped onto the IDF 2011 stage to show off a demo which involved a nodding cat wearing headphones. The engineer said:

AMD held a tech day for its upcoming Bulldozer architecture and while not much is allowed to be said, Anandtech have come out swinging with the overclocking potential of the Bulldozer chip. Most people wouldn't think it could hit these speeds, but just wait until you see what AMD can do. AMD used different overclocking experiments and had three different stations setup for people to look at.

The first one used a sub-$100 closed-loop water cooling solution from Antec in the form of the Kühler series, on this solution the Bulldozer was able to reach 4.8GHz as seen below:

The next test used phase change cooling. With the phase change cooling system, AMD was able to push the 8-core FX CPU up to a decent 5.894GHz at 1.632V, with a video of this achievement below:

This news certainly won't gel well for many of you at a time when we thought launch was imminent - AMD is once again choosing to delay the launch of its FX-Series desktop processors, now planned for release in October.

AMD's rescheduling for October is supposedly due to AMD wanting to make sure they hit all the right performance price-points with the first wave of models; thus they are said to be introducing seven models straight out the gate comprising four 8-core FX-8000 series chips, one 6-core FX-6000 series chip and two 4-core FX-4000 series chips.

Looking at a chart compiled by the folks at Xbit labs after gathering data from unknown sources, the initial line-up of models will include the FX-8150 (a 3.6GHz part), FX-8120 (3.1GHz - avail. in both 95 and 125W variants) and FX-8100 (2.8GHz) as far as octo-core models go. The six-core model to accompany them is the FX-6100 running at 3.3GHz, while the two quad-core models are the FX-4170 and FX-4100, running at 4.2GHz and 3.6GHz respectively.

AMD has released some nifty documents which outline the architecture behind their upcoming Core i7 competitor, Bulldozer. The Bulldozer CPU architecture is one of the biggest things to come from AMD in the form of CPU technology in a long time and the slides below outline Zambezi for the Socket AM3+ platform. The AM3+ socket infrastructure adds quite a few new things and a few surprises along the way. Check out the below images for a bit of a tease of things to come from team RED.

One of the surprising things is older AM3 processors are plug-in compatible with AM3+ motherboards, which would allow people to pick up a board in preparation for the new CPUs and then upgrade as they see fit. As well as the impressive 72-bit DDR3 memory channels which will see support for DDR3-1866. Bulldozer is quite power-efficient too, as it will increase clock speeds if there is spare TDP headroom and when there is not much workload on the CPU, it will disable un-used cores.

Very impressive, AMD. If it can keep up with Sandy Bridge-based Core i5/i7's, then we'll finally see some true competition in the CPU fight.

While inevitably not bringing anywhere near the amount of excitement the upcoming FX-Series will once launched, AMD has just kicked out three new APUs catering to its Brazos platform - the E-450, E-300 and C-60.

Using AMD's Bobcat x86 architecture, all three models are dual-core, built on the 40nm process. The new top dog "Zacate" E-450 APU clocks in at 1.65GHz (where the previous top model, the E-350 ran at 1.60GHz). The integrated Radeon HD 6320 GPU has its core clocked at 508MHz and with its TurboCore tech it can be driven at up to 600MHz.

THe E-300 lacks TurboCore on both its CPU and GPU components, with the CPU clock at a fixed 1.30GHz, and the Radeon HD 6310 GPU at 488MHz. As for the C-60, this is AMD's latest ultra low power APU with a TDP of only 9W (the other two are double that). The C-60 does have TurboCore technology embedded for both its CPU and GPU components. The CPU's stock clock rate is 1.00GHz, but with the help of TurboCore can head north to 1.33GHz, while the Radeon HD 6290 GPU has its core clocked at 276MHz, but TurboCore can increase it to 400MHz.

It's been discovered today that Intel is planning to shave pricing on several of its current Sandy Bridge desktop CPUs next month.

The drop looks to be consistently around about $10 on all nine models, which are the dual-core Intel Pentium G630 and G850, dual-core Core i3-2120 and i7-2390T, and quad-core Core i5-2400S, i5-2405S, i5-2500S, i5-2500T, and i7-2600S.

Unfortunately, where most of us would've been especially wanting to see cuts on the more popular enthusiast models such as the i5-2500, i5-2500K, i7-2600 and i7-2600K, they are not mentioned in the list.

Detailed info on AMD's upcoming Bulldozer (Zambezi-FX) processors has hit the Internet with Chinese website Corescn.com claiming that the B2 Revisions of the CPU's have hit production and would be on shelves by late Q3 2011. The Bulldozer CPU's have been delayed because of a performance-related bug, with the B2 revision going into mass production sometime in August 2011 and hit markets by September 2011.

The first two CPU's based on the B2 revisions will be the FX-8150 (FD8150FRW8KGU) Featuring 8Cores/8Threads, 8MB L3 Cache, Stock Frequency of 3.6Ghz (Turbo - 4.2Ghz) while the other FX-8120 (FD8120FRW8KGU) is also an 8 Core CPU with same specs and just a little low clock frequency of 3.1Ghz (T.C - 4.1Ghz). Both CPU's feature TDP of 125W.

We should also expect the FX-6100 (Hexa Core), FX-4100 (Quad Core) and FX-8100 (Octa Core) CPU's featuring 95W TDP. This will be followed by another lineup which would release in Q1 2012 and is set to include 4 new CPU's detailed below:

It has just come to light that Intel does not intend to ship its upcoming Sandy Bridge-E (LGA2011) Core i7 processors with a heatsink/fan (HSF). Instead, Intel plans to offer the certified HSF for these processors as a separate, standalone purchase through the same retail channels.

While there's no doubt some people are satisfied with, or even prefer an Intel certified cooler for these high TDP processors, Intel would be banking on the majority of the target users wanting to use third-party cooling options, as it is expected that many of the big name CPU cooler makers will have lineups of LGA2011-ready coolers readily available upon the Sandy Bridge E series launch.

Sandy Bridge E processors have a rated TDP of 130W, but early testing on engineering samples has proven that it can reach as high as 180W under load (and that's at stock speeds).

From time to time we'll see one of those crazy world renowned overclockers kick out another incredible world record result from a recent piece of hardware thanks to the most extreme cooling methods used. But it's not so frequent that a new WR really jumps out at you as being exceptionally above and beyond anything we've seen before.

Well, Mad222 of Team Centralfield has managed to do just that with an incredible overclocking result with a Phenom II X4 955 processor. Using GIGABYTE's 890FXA-UD7, the CPU was pushed all the way up to a jaw dropping 7378MHz.

Dropping it back only ever so slightly to 7155MHz, a 1M SuperPI result was also completed, as you can see in the below screenshot :-

This can't be confirmed as 100% truth yet, but it seems AMD has "accidentally" leaked out the pricing of their upcoming eight-core Bulldozer chip. AMD announced a give away of 5 x AMD FX-8150P processors and if you scroll through the terms and conditions of the competition, the following line can be found:

Top tier prizes: Five (5) AMD FX series eight-core processors. Approximate Retail Value: $300 USD each.

Oh really... well, it seems the processor will compete against the Intel Core i7 2600 Sandy Bridge CPU which is currently retailing at $300. Can AMD bulldoze Intel over with this chip and the competitive pricing, we'll have to wait and see!

While we're yet to get our hands on some octo-core processors in the form of the upcoming FX series from AMD to battle it out against their thriving competitor, a freshly leaked roadmap today reveals that AMD has even bigger plans already in the works, with some beefy 10-core desktop CPUs scheduled for release next year.

These 10-core chips (codenamed Komodo) will be the successor of the upcoming (Zambezi) FX-Series models. Word of Komodo was out as far back as last year, but at the time was said to have up to eight cores.

A bit about the CPU. It will use 10 'Piledriver' cores, have Turbo Core 3.0 technology, built-in DDR3 memory controller, housed in an FM2 package and manufactured on 32nm tech. It's being designed for the 'Corona' platform along wit hthe Hudson D4 chipset which supports four native USB 3.0 ports and eight SATA 6.0 Gbps ports.