CPU's Article (28)
AMD Vishera FX-8350 Review
» October 21, 2012 05:00PM
AMD 2nd Generation A10 5800 & A8 5600 Desktop APU Review
» October 1, 2012 05:00PM
Sapphire HD 7970 Vapor-X GHZ Edition Review
» August 22, 2012 05:00PM
Intel Third Generation Core i7 3770K Review
» April 22, 2012 05:00PM
Intel Second Generation Core i7 3820 Review
» March 14, 2012 05:00PM
CPU's News (442)
Posted: May 7, 2013 11:29AM
Author: Tobias Thydal
Intel has announced its new Silvermont architecture, which is going to improve the Atom line significantly. The new architecture will be based on Intel's 22 nm 3D Tri-Gate SoC, which is faster while using less power than the current generation. Intel claims that Silvermont will deliver up to three times more peak performance or about the same performance as the current generation using only 20% of the power. The new architecture allows for a new out-of-order execution engine, a multi-core and system fabric architecture that is scalable up to eight cores, new IA instructions, and improved power management capabilities, such as low power C states.
Intel foresees that these chips will be used in various applications, such as smartphones, microservers, tablets, entry level laptops, and network infrastructure. Release dates of some of the CPUs based on Silvermont have been announced:
- Baytrail: Quad-core processor for tablets - holiday 2013
- Merrifield: Smartphone processor - by the end of 2013
- Avoton: Microserver chip - second half of 2013
- Rangeley: Routers and other networking gear - second half of 2013
The success of these chips are crucial for how well Intel will do in the growing mobile market. ARM has been dominating the market for quite a while, and other competitors are trying to get their market share as well. Intel needs to get some results soon, if it wants to have a say in the future of mobile computing.
Posted: May 5, 2013 01:53PM
AMD has launched a new division which aims allow companies to design their own processors for specific applications using the basis of AMD's existing APU technology, while incorporating the company's own intellectual property. This has already been applied in the upcoming Playstation 4, which utilises a heavily modified APU based on the Jaguar core, but until now, there has not been a dedicated division for custom chips. It is hoped that the advent of this new division will help generate the profit that AMD so greatly needs.
One example of the application of this new initiative would be use in Smart TVs, where the APU could contain hardware-accelerated decoders for specific video file types. Interestingly, the new division could enable AMD to incorporate ARM cores into its existing APUs, which may eventually open doors for AMD to enter the ARM SoC market. This could then lead to more powerful graphical rendering capabilities on smaller ARM-based devices such as smartphones and tablets.
Posted: May 1, 2013 08:04PM
Intel's Haswell processors are on the way, but it appears some power supplies may have some issues getting things started. A new report shows that Haswell's C6/C7 states require 0.05 amps on the 12V2 rail, which could be a problem as some PSUs can't supply that low level power. If that's the case then it opens up a lot of stability problems or the simple fact of having a PSU shut off entirely. Intel's Reseller Center website includes a handy list of power supplies, and when you sort by a minimum 12V2 load of zero amps, there's a grand total of 23 units that meet the requirement: 19 Corsair, three InWin, and one Seasonic. There could be more that support it that aren't included on Intel's list, but chances are it won't be too many more.
However, despite all of this there is still some good news. Corsair's Robert Pearce believes motherboard manufacturers could disable the C6/C7 states in the BIOS to ensure compatibility with more PSUs. Users can always enable those states later on once their power supply supports them. Corsair is working to make sure all of its PSUs support the C6/C7 states, and hopefully other companies do too.
Source: Tech Report
Posted: April 30, 2013 01:40PM
AMD has today announced two new desktop processors based on its current Piledriver processor architecture. The FX-4350 and FX-6350 follow the current naming convention for AMD processors, the former being a 4-core variant, and the latter featuring 6 cores, but offer improved clock speeds and L3 cache.
The FX-4350 boasts a 4.2GHz clock speed, coupled with 8MB of L3 cache. This makes the FX-4350 400MHz faster than the current FX-4300, and also doubles up on the L3 cache. The FX-6350 on the other hand boasts a 3.9GHz clock speed, and also offers 8MB of L3 cache, similar to the FX-4350. The FX-4350 has a 95W TDP, and ships for $122, whereas the FX-6350 pushes the TDP up to 125W, 30W higher than its FX-6300 counterpart, and is priced at $132.
Posted: April 24, 2013 10:31PM
Author: Tobias Thydal
Prices for Intel’s next generation of processors, Haswell, have been released. The information comes from a Chinese source that was able to get access to information about the MSRP of Haswell. Please keep in mind that the prices below are lower than the actual prices consumers will see; however, they can serve to give an idea of the prices for the new CPUs. For reference, the Core i7-3770k has a MSRP of $332.
It's in the lower part of the table that it gets interesting, because that is where the new generation is slightly cheaper than Ivy Bridge. The i5-4570 is only $189 compared to its predecessor, i5-3570, which costs $213. An interesting detail is that Intel is introducing more chips in the low- to mid-range segment. This could be a sign that Intel intends to win the market for laptops and tablets, since manufacturers of those will be able to choose the best suited chip for the job. The old x6xx chip is back as well, which was left out in Ivy Bridge. So it should now be possible to choose a CPU that suits ones needs and budget exactly.
A reminder as to why you should be excited for Haswell. It is expected to be up to 10-15% faster than Ivy Bridge, the iGPU is said to be twice as fast as HD 4000 iGPU and to top it off it is going to require less power to operate.
Posted: April 17, 2013 09:28AM
According to a report on PC fan site Guru3D, Intel's next generation Haswell chips will be easier to overclock than any previous generation. Haswell chips will reportedly allow overclocking using both the multiplier and modifying the base clock of the chip, which should allow even those using chips with a locked clock multiplier to gain a modest overclock.
As well as this, Haswell chips are said to feature an integrated VRM controller which should allow enthusiasts and overclockers to squeeze the last bit of performance out of their chips. With these features, next generation Intel chips should be able to hit 8GHz using LN2 cooling, which should be enough to rival AMD's FX series which currently dominate the record frequencies.
Posted: March 28, 2013 10:20AM
Intel has announced that its upcoming Haswell series of processors will feature two new DirectX extensions, PixelSync and InstantAccess. The former, PixelSync, enables Order Independent Transparency (OIT), which speeds up the sorting of transparent graphical elements, for example windows, by ensuring they are rendered in the correct order. OIT is a common feature found in DirectX 11 games, however it is not a feature of the API.
Current methods of OIT involve using large amounts of VRAM and bandwidth, neither of which most integrated graphics solutions, including Haswell, excel at. For this reason, Intel has developed its own implementation of the technology which relies less on the bandwidth and VRAM of the processor.
The other extension, InstantAccess, enables the CPU part of the processor to access a location in the GPU memory (even though the CPU and GPU parts of the chip share the same memory, they each have their own allocated sections). This alleviates the requirement to 'copy' the data to the CPU memory, which is very costly in terms of performance.
Posted: March 12, 2013 11:03AM
AMD has today announced the release of its long awaited 'Richland' Accelerated Processing Units (APUs), now called the AMD Elite A-Series APUs. These chips boast up to four traditional CPU cores as well as an integrated GPU based on AMD's upcoming HD8000 series of graphics chips. The chips are designed for use in mid to high-end notebooks, and will begin appearing in stores soon. Low-voltage versions of these APUs are expected to become available later this year.
The improved Richland APU is claimed to outperform mobile Intel Core i7 chips in both straight-out performance and graphical rendering. If this claim transfers into real-world applications, Intel may need to rapidly rethink their mobile CPU line, which is often much more expensive than its AMD counterparts. The AMD A4-5150M and A6-6350M are both dual-core chips, while the A8-5550M and A10-5750M are quad-core chips with improved graphics processors. The chips were first shipped to vendors in December, however no products boasting the chips have yet been released.
Posted: January 28, 2013 10:33AM
AMD has released some additional details on its upcoming Richland desktop APU lineup, which is geared to replace the current Trinity architecture. Trinity was introduced late last year, but lacked the power many users were hoping for. Although Richland remains on the same 32nm Piledriver core as Trinity, it adds support for DDR3-2133MHz RAM, increases CPU clock speeds, and integrates a HD8XXX GPU into the chip. The flagship APU in the Richland series will be the AMD A10-6800K boasting four cores, an unlocked clock multiplier, and an integrated Radeon HD8670 GPU.
The unlocked quad-core Richland APUs all have a 100W TDP, while the other locked chips and dual-core variants will feature a more modest 65W TDP. The Richland series will remain compatible with the current A55/75/85X chipsets, but will also be compatible with new A68/78/88X series chipsets.
Posted: January 22, 2013 01:17PM
According to a DigiTimes report, Intel is planning to host a conference ahead of Computex 2013, held in June where it will announce its upcoming Haswell series of processors. The new processors will be made available from June 2nd, with vendors showcasing their Haswell-based products at Computex. These processors are expected to account for 14-16% of Intel's processor shipments during Q3.
Haswell is the successor to the Ivy Bridge architecture currently applied in Intel Core 3XXX processors, and will make up all Core 4XXX processors, except high-end Ivy Bridge i7 Extreme processors. The launch of Haswell is anticipated by much of the PC market, and will feature a new microarchitecture which allows new chips to be introduced, such as the low power Haswell Y series.
Posted: January 9, 2013 12:53PM
Earlier, during this year's CES, Intel revealed its new Ivy Bridge Y series of processors rated at 7W designed for use in low power laptops and tablets. One of the most surprising details of the new chips were the TDP (Thermal Design Power). This is particularly surprising considering the current-generation 17W Ivy Bridge chips, with the new chips representing a decrease of 10W. It has since been established, however, that the 7W quoted value is not as it may seem on the surface.
Intel has since, however, released a statement that reveals the rated TDP of its new Ivy Bridge chips is, in fact, 13W. This is because Intel's 7W rating is based on the chips' SDP (Scenario Design Point) rating, which is a designed to represent the processor under the stress of mainstream use. SDP is not comparable to the industry standard TDP, and is unlikely to be adopted in the near future.
Posted: January 8, 2013 12:01PM
Intel has recently announced its plans to add a quad-core variant to its Atom series of mobile processors during a press conference at CES. The project, code-named Bay Trail, remains on the existing 22nm process used in most current Intel processors, and will be clocked at 1.2GHz. These Bay Trail processors are to be be ready in time for the 2013 holiday shopping season. Intel also demonstrated a number of tablets with the new Atom chip running the Windows 8 Operating System.
The quad-core chip is targeted at the smartphone and tablet market currently dominated by the likes of Qualcomm, Samsung, and Apple. The reference design phone Intel produced features a 3.5 inch screen and is relatively thin. Intel must make significant improvements over its current series of Atom chips before it can take on much of the smartphone and tablet market.
Posted: December 27, 2012 07:01AM
AMD's eight-core FX-8300 is now available for enthusiasts to play with, after weeks of informed speculation on whether the company would be releasing another cost-effective chip. The big news here is FX-8300's 95W TDP, which is considerably less when compared to the rest of AMD's eight-core Vishera lineup. Running at a default clock speed of 3.30 GHz with 3.60 GHz of TurboCore speed, the AMD FX-8300's four modules packs in eight cores, along with 8MB total L2 cache and 8 MB shared L3 cache. Priced somewhere between the FX-8350 and FX-8320, the FX-8300 retails for an estimated $190. Despite the chip being slower than its siblings (and priced so closely), the drop in TDP is thought of as its main selling point.
Posted: December 18, 2012 11:16AM
According to a recent report, Intel is to release its high-end Haswell i7 mobile processors in April 2013. These high-end processors will all be quad-core, and will have Hyper-Threading technology enabled, allowing two threads to run on each physical core. The chips will have a base clock of between 2.7 and 3.0GHz, and will have turbo frequencies between 3.5 and 3.7GHz. The chips will have 47-57W TDPs and are likely to be used in high-end gaming laptops as well as all-in-one desktops where low power consumption is of paramount importance.
The Haswell series of mobile processors will also come in a variety of variants, including a new H series that will all be quad-core chips with TDPs between 37 and 57W. The report also states the H series chips will use the BGA (ball-grid-array) package in which the processor is soldered directly onto the motherboard itself. The information has not yet been confirmed by Intel, however it would appear that the firm will be introducing BGA processors as an alternative to the mainstream socketed processors. It will be interesting to see the price difference, if any, between the two types of processor.
Posted: December 12, 2012 01:10PM
The Chinese version of tech website VR-Zone has published a leaked spec sheet that allegedly contains information on the first wave of Intel's upcoming Haswell processors releasing in Q2 2013. The processors detailed range from the midrange Core i5 4430S to the high end Core i7 4770K.
The Haswell chips will be produced on a 22nm die, with frequencies ranging from 2GHz to 3.5GHz. The maximum TDP of Haswell-based processors will be 84W, 7W higher than the current top Ivy Bridge processors. Low voltage versions of the chips will also be available with a 35W TDP. Haswell-based Core i3 processors will arrive along with the ultra high end Ivy Bridge i7 Extreme processors during Q3 2013, after the first wave of i5 and i7 processors.
Posted: December 11, 2012 05:24AM
Author: Nick Harezga
Representatives from Intel and IBM presented papers detailing new 22nm technology at the International Electron Devices Meeting in San Francisco yesterday. IBM is using a 3D process to create server processors with performance that is 25-35% greater than current 32nm technology. IBM plans to further discuss the 3D technology on Wednesday. Intel has created several variations to its 22nm process for system on a chip units. This offers performance improvement up to 65% over Intel's 32nm planar process. The process will also offer increased performance in fast interface devices such as Ethernet and HDMI.
Posted: December 10, 2012 04:18PM
A supposed Intel powerpoint slide got leaked with some details on the company's upcoming low-powered Ivy Bridge processors, designated as "Y" variants scheduled for a Q1 2013 release. Similar to Intel's next-generation Haswell processors, some of the new Ivy Bridge chips consume lesser power than current models, boasting a TDP of only 10W. The Core i7-3689Y will be the most powerful CPU of the lineup with only two cores at 1.5GHz, but takes only 13W of power. Hyper Threading is enabled with up to four threads, along with 4MB L3 cache and a maximum of 2.6GHz at Turbo Boost. The 1GHz P-2129Y Pentium is among the least power-hungry of the bunch at 10W, but as expected is hindered with no other extras apart from being a dual core with embedded graphics support. The Ivy Bridge "Y" chips are speculated to be targeted primarily for the ultrabook market, and are considered to be the stopgap until the 22nm Haswell architecture makes its way to consumers.
Posted: December 6, 2012 12:53PM
A report was recently published that suggested processor manufacturer Intel may move away from LGA-based processors and move towards BGA-based processors for the consumer market. LGA processors have a socket, so the CPU can be removed from the motherboard and replaced or upgraded, whereas BGA processors are soldered directly to the motherboard itself, which would make processors very difficult to upgrade. This sparked an outcry among PC enthusiasts who often choose to upgrade processors instead of changing an entire motherboard. BGA processors do offer many advantages, including reduced power consumption and heat output, as well as a smaller die size.
Initially, Intel did not respond to the report, but rival AMD stated that it had no plans to abandon socketed processors in the near future. More recently, however, Intel has responded by saying that it has no plans to replace socketed processors with BGA packages, which will come as a relief to PC enthusiasts everywhere. The official statement from Intel spokesman Daniel Snyder does state that Intel "cannot comment on specific long-term product roadmap plans".
Posted: December 6, 2012 07:12AM
Intel has caught a bit of flak amidst rumors that it will ditch offering socketed CPUs after its Haswell lineup to focus on BGA packages which permanently embeds processors on motherboards, and AMD has responded by proclaiming that it will not abandon production of socketed client processors. While Intel has diffused the gossip somewhat, AMD maintains that the DIY PC enthusiast market remains a critical part of the company's offerings.
Here's what AMD's Chris Hook had to say: "AMD has a long history of supporting the DIY and enthusiast desktop market with socketed CPUs & APUs that are compatible with a wide range of motherboard products from our partners." The next two years will bring in the "Kaveri" APU and FX CPU lineup, which will remain as socketed products. He adds that while AMD has a diverse range of BGA-packaged processors that caters to ultrathin notebooks, all-in-one desktops, along with embedded applications and tablets, the company has no current plans to move to BGA only packaging and will continue its support of this critical market segment. "For the desktop market, and the enthusiasts with whom AMD has built its brand, we understand what matters to them and how we can continue to bring better value and a better experience."
Posted: December 6, 2012 06:27AM
It will take Intel around two years to manufacture processors in 18-inch diameter wafers using the 14 nanometer silicon fabrication process, said its chief technology officer Justin Rattner. The CTO elaborates that this will be made possible as development between the technology and machinery that will make it happen steadily inches closer to completion. With that said, Rattner adds that the rate at which Intel pursues these advances will extend Moore's Law by 10 more years. The latter part of 2013 will see Intel starting its 14nm CPU production with batches of the P1272 and P1273 SoCs. Once 14nm becomes more mainstream, Intel will head further on into 10nm, 7nm, and 5nm manufacturing processes, which should begin by 2015.
Posted: December 4, 2012 05:29AM
Author: Nick Harezga
The AMD Opteron line of processors helped me to get my start in overclocking with a pair of chips that clocked exceptionally well. It has been a long time since I broke 3GHz with my Opteron 148 and 170, and AMD has just announced an update to the Opteron line of processors today. The Opteron 4300 and 3300 are low-power units that are targeted at "cloud servers that process Web transactions." The new Opterons will be featured in the SeaMicro servers from AMD, upgrading from the current 4200 and 3200 series. The 4300 series has six different options ranging in price from $191 to $501, delivering speed from 2.2GHz to 3.1GHz, and drawing just 35 to 65 watts. The 3300 series has three CPUs ranging in price from $174 to $229, clock speeds between 1.9GHz and 2.6GHz, and drawing power between 25 and 65 watts.
Posted: November 23, 2012 02:13PM
Overclocking has become much more popular and accessible in recent years, not only for professional overclockers, but also for PC enthusiasts, with the world record being pushed higher and higher. The current record is held by an AMD Bulldozer FX-8150 running at 8,709.06MHz using liquid nitrogen cooling. Very few of even the most aspiring PC enthusiasts could hope to get even close to this value, even with the best watercooling gear and especially for sustained periods of time.
Having said that, AMD’s new Piledriver core is getting very close to its Bulldozer counterpart, with the record set by an FX-8350 currently at 8,470.74MHz. The new Piledriver core, although still on a 32nm die, is no longer drawn by hand, which AMD claim provides the boost that generally accompanies a process shrink. This should in theory provide more overclocking headroom for professional overclockers and PC enthusiasts alike, so we can expect the FX-8150’s crown to be contested soon enough.
Posted: November 22, 2012 10:56AM
Want an affordable CPU upgrade that's easily overclockable to 5GHz? According to PC Gamer, that's what a budget AMD FX-6300 Piledriver CPU can offer with ease. Sure, Intel's offerings are probably better choices if your wallet could spare a little bit more, but the embattled AMD just might surprise a few folks looking for more bang for their hard-earned bucks. The recently released 3.5GHz Socket AM3+ Vishera chip runs at a rated 95W, and retails for around $140. At stock, its six cores make it a respectable competitor against Intel's i3-3225, but AMD's overclocking advantage takes its edge even further. Unfortunately, while the 1.5GHz boost for the FX-6300 is certainly impressive, the article makes no mention of the cooling options used for successfully overclocking the chip, not to mention the inevitable increase in power consumption which adds to more internal heat. Still, it's a little encouraging to see AMD wiggle some overclocking headroom in its Piledriver CPU lineup, something that enthusiasts will no doubt happily experiment with.
Posted: November 21, 2012 05:07PM
It wasn't too long ago when a leak appeared about the Intel Haswell Lynx Point chipset, and now there's a leak on the Intel Bay Trail-T processor. These chips are designed with tablets in mind and feature some improvements over the older Clover Trail parts. Bay Trail-T is built on a 22nm process (code name Valleyview-T SoC) instead of 32nm like Clover Trail and 28nm of ARM Cortex-A15, which translates to improved performance and power consumption. The Valleyview-T SoC is a quad-core part with an HD 4000 GPU based on Ivy Bridge, and both together are capable of powering a 2560x1600 resolution display. Oh, and it can record stereoscopic, 1080p 3D video at 60 fps in case 3D tablets take off.
There aren't any concrete numbers of battery life, but Intel does promise Bay Trail-T parts use 50% less power than its predecessor. That would mean at least ten hours of battery life in a tablet, which is nothing to sneeze at. The only downside to all of this information is tablets running Bay Trail-T aren't expected to arrive until 2014. Click here for the pictures.
Posted: November 21, 2012 12:04AM
Author: Marko Jurac
There are rumors that Samsung is going to be building a new SoC for its high-end smartphone and tablet devices based on a new ARM chipset design. Named the "big.LITTLE", it will feature eight cores, four of which run at 1.8GHz geared for performance with the other four running only at 1.2 GHz for power efficiency. Based on ARM's Cortex-A15 core architecture, the design is taking queues from NVIDIA's Tegra 3 chipset with its '"4-PLUS-1 Quad-Core with a 5th Battery Saver Core" design. The basic idea of the design is it allows for variable switching between performance and power saving cores. The Samsung variant is rumored to follow the same design and be built upon a 28nm manufacturing process. With rumors looming of the new Samsung Galaxy S4 coming next year, it's very likely we'll hear more about Samsung and this supposed chipset.
Posted: November 14, 2012 10:59AM
Author: Marko Jurac
Advanced Micro Devices (AMD) has enlisted the aid of JPMorgan Chase & Co. for financial advice as it struggles with chip sales in an ever changing martket more geared towards mobile computing. In the last year, AMD's shares have fallen over 60 percent with a current market cap of $1.4 billon. Long term debt and captial lease contracts have AMD owing some $2 billon to banks. Since the recent rush in mobile, AMD has had trouble in the last few years trying to keep up with names like Intel and ARM, which produce cheaper and more power efficent chips in a PC market now surging with an emphasis on mobile. AMD reported last month it would slash the workforce by 15 percent in an effort to refocus resources towards areas like communications, industrial and gaming hardware. Goldman Sachs analyst Jame Covello recented reported to clients that chances on AMD liqudating the PC processor divison were between 15 and 30 percent. AMD's current target market cap is $1.1 billion due to the company's decreasing size.
Posted: November 13, 2012 03:19PM
Author: Marko Jurac
New information on Intel's Haswell micro-architecture has recently been leaked after slides showed up on the Chinese site EXPreview. According to the leaked slides, desktop chipsets are expected to support up to 14 USB 2.0 ports, six USB 3.0 ports, six SATA 6Gbps ports, and eight PCI-e 2.0 lanes. Two more USB 3.0 and SATA 6Gbps SATA ports are listed to be shared along the other eight lanes of PCI-e 2.0. CPU overclocking on Lynx Point-LP is to be dropped as well. On Haswell ULT processors, a new CPU throttling feature drops the base clock from 100MHz to 24MHz. Also, new C-state power levels (C8, C9, C10) are included on the ULT chips to further reduce power usage. Haswell is being aimed towards the ultra-book range of laptops with an emphasis on power efficiency and platform integration.
Posted: November 12, 2012 08:07PM
Author: Marko Jurac
Intel announced today its new Xeon Phi co-processor lineup, which is to incorporate Intel's Many Integrated Core (MIC) architecture. The new Xeon Phi co-processors are being packaged in discrete graphics card-like form to work in tandem with Intel's Xeon E5 family of server and workstation processors. According to Intel, the Xeon Phi is the accumulation of eight years of work to produce a chip to tackle both high performance computing (HPC) and super-computing markets. This puts it in competition with products like the NVIDIA Tesla K20 series and AMD's recently announced FirePro S10000 series.
Some key specifications of the Xeon Phi are a 22nm manufacturing process with Intel 3-D Tri-Gate transistors, up to 60 cores/1.053 GHz/240 threads, 1 teraflops of double-precision performance, 8 GB GDDR5 memory and 320 GB/s of memory bandwidth, 512-bit memory interface, 32 KB L1 I/D cache, 512 KB L2 cache per core (30MB total), 225W TDP, and a PCIe x16 form factor. Intel reports that the Xeon Phi co-processors are to become generally available on January 28, 2013, with the Xeon Phi 5110P priced at $2,649 and the Xeon Phi 3000 family to be priced under $2,000 per SKU.
Posted: November 9, 2012 08:06PM
NVIDIA released its third quarter earnings earlier, but there was something else in there too that should be of great interest to many. NVIDIA's CEO, Jen-Hsun Huang, announced Project Denver will be a 64-bit desktop processor based on the ARMv8 architecture. However, not much else can be said about the chip other than Huang saying it's "going great." ARM recently showed off its ARMv8 cores with the Cortex-A50 series, which is what AMD will use as its SoC Opteron chips. Those Cortex-A50 chips aren't expected to arrive until 2014, but NVIDIA may have Project Denver out before then. Last March, Huang said the chip was around three and a half years into its five year design cycle, which could mean Project Denver is pretty much completed. We'll just need to wait and see what NVIDIA decides.
Posted: November 8, 2012 07:03PM
Author: Marko Jurac
Early this morning, Intel announced the launching of its new Itanium 9500 processors, which are to date the most powerful enterprise-based CPUs from Intel for data center computing. With the new chip containing 3.1 billion transistors, it is undoubtedly the most advanced and complex server-based CPU by Intel yet. With twice as many cores (eight instead of four), up to 54 MB of on-die memory cache, and up to 2TBs of low-voltage RAM to couple with this beast, it is sure to be a killer. Intel reports the Itanium 9500 being as much as 2.4 times faster than the older Tukwila-era design. The Itanium 9500 family is aimed to handle more demanding data heavy workloads such as business analytics, database, and enterprise resource planning applications. The new frequencies will range from 1.73GHz to 2.53GHz and TDP ranging from 130 to 170 watts. The Intel Itanium 9500 series is available now and is priced from $1,350 to $4,650 in quantities of 1,000 units.