CPU News

AMD is expected to launch "Beema" mobile APUs this year. "Beema" is is the third generation ULV core, that will succeed current "Kabini" core, however it will still make use of 28nm technology. "Beema" accelerated units will have up to 4 CPU cores, GCN graphics, and integrated Security Processor. According to AMD estimates, the APUs will have much better performance per Watt, up to 104%, than "Kabini" products. The actual TDP will depend on SKU, and it will range from 10 Watt to 25 Watt.   [Full story]
Earlier this week Intel added information on "Haswell Refresh" processors to their official pricelist. It is suffice to say that there were no unexpected surprises there, and the new SKUs had exact same prices as the first generation Haswell microprocessors. Although the list included the full line of "Haswell Refresh" desktop CPUs, they will be available only in May, due to their release date being tied to the launch of desktop 9-series chipset. The mobile refresh, on another hand, will work with currently available 8-series mobile chipsets, and for that reason we were not surprised to see new mobile processors in Intel's product database.   [Full story]
Czech website Tyden.cz have published rumored specs for the Maxwell based GTX 880 graphics card. While some of these specs look reasonable on paper, they don't all look like an upgrade compared to the GTX 780 and 780 Ti. There is another rumor that suggests this is not intended as a replacement for the GTX 780 or 780 Ti, suggesting it is more a replacement for the GTX 680.   [Full story]
Intel today added brief specifications and prices of all "Haswell Refresh" desktop and mobile microprocessors to the official pricelist. Built on "Haswell" architecture, upcoming products will not have any new microarchitecture and power optimizations compared to last year's products, and their only advantage will be slightly increased clock speeds. According to known specifications, many "Haswell Refresh" CPUs will be clocked 100 MHz higher than the Haswell 2013 lineup, which will result in up to 2% - 5% better performance, much smaller than we got used to in the past three years.   [Full story]
AMD this week launched low power desktop platform utilizing socket AM1, formerly known as a "socket FS1B". The socket AM1 processors are built on "Kabini" core, which combines "Jaguar" CPU architecture and "Graphics Core Next" GPU architecture, and integrates it all together with all necessary chipset logic on a single system-on-a-chip (SoC). The SoCs have 2 and 4 CPU cores, and the graphics unit comes with 128 shader cores. Feature-wise, new "Kabini" processors are similar to mobile Kabini products with the exception of somewhat higher clocks, and a new PGA type packaging, that simplifies processor upgrade or replacement.   [Full story]
Digitimes reported this week that Intel Haswell Refresh processors will be available next month. That information confirms older reports, like the one published by Hermitage Akihabara back in March. According to Digitimes, Intel originally planned to launch microprocessors in April, and follow with the 9-series desktop in a month. The latest plan is to launch the CPUs together with new chipset in early May. Hermitage Akihabara provided launch date of May 11.   [Full story]
Intel this week released three Quark processors, all featuring extended operating temperature range. New X1001, X1011 and X1021D are qualified to operate from -40°C to +85°C, while offering the same performance and feature-set as Quark X1000, X1010 and X1020D models, that were released in the past two quarters.   [Full story]
Intel is going to retire Atom S1200 microserver processors, and the first Z3700 systems-on-a-chip for tablets. The list of affected models includes Atom S1260, S1240 and S1220, as well as Atom Z3740, Z3740D, Z3770 and Z3770D. All of soon to be discontinued Atoms are based on the Silvermont architecture and manufactured on 22nm technology, however all of them use earlier core steppings B1, B2 or B3.   [Full story]
Intel recently completed transition of Xeon E5 and E7 families to one-generation old Ivy Bridge architecture, and the company is hard at work on the next generation processors, built on the latest Haswell microarchitecture. Similar to Ivy Bridge server lineup, "Haswell" processors will be branded as Xeon E5 and Xeon E7, but their suffix will change from "v2" to "v3". Intel is also going to release new Xeon E3 "Haswell" SKUs, and eventually introduce the next generation E3 microprocessors, utilizing Broadwell architecture.   [Full story]
AMD has announced the Radeon R9 295X2 Graphics card. Powered by the Vesuvius XT GPU, which comprises a pair of Hawaii XT units, this is the dual-GPU powerhouse for the Graphics Core Next based graphics cards. This is a pair of Radeon R9 290X in crossfire, on a single board. It is a power hungry beast, operating well outside the PCI-E power specs, but supplying premium processing power.   [Full story]
Apr 07, 2014   AMD A8-7600 APU delayed
On January 14 AMD launched the first Steamroller-based APUs for socket FM2+ platform, and announced an energy-efficient A8-7600 SKU. The quad-core A8-7600 was a bit slower than the A10-7700K, and did not have the unlocked multiplier. On a positive side, it had about 22% lower price sticker, and supported a "configurable TDP" feature. Rated at 65 Watt when running at stock 3.3 GHz frequency, the APU could be configured to fit into 45 Watt TDP, albeit in this case its base and turbo clocks would be limited to 3.1 GHz and 3.3 GHz respectively.   [Full story]
AMD has launched the FirePRO W9100 professional graphics card. Powered by the same Hawaii XT GPU seen in the Radeon R9 290X, the card runs at a slightly slower core clock speed than its desktop counterpart. It has an unprecedented 16 GB of memory, allowing users to work with much larger internal datasets than previously possible with no sacrifice in performance.   [Full story]
Just two weeks after the launch of the Xeon Phi 7120D co-processor in dense form factor packaging, Intel is ready to introduce another member of the 71xx family. The latest Xeon Phi product, which was added today to Intel's official pricelist, has the model number 7120A. The 7120A is packaged as an add-on card with PCI-Express interface. The co-processor has an active cooling solution on the board, which distinguishes it from the launched last year 7120P (it comes with a passive cooler).   [Full story]
AMD has added the entry level Radeon R5 230 to it's retail graphics card line-up. Previously only available for the OEM market, the R5 230 is a rebrand of the Radeon HD 6450 and HD 7450, albeit with the memory clocked to match the 7450, which is slower than the 6450. The card uses a Caicos Pro GPU, first launched in February 2011, which is powered by the TeraScale 2 Unified Processing Architecture, introduced a year earlier.   [Full story]
Intel last week introduced two entry-level Rangeley C2308 and C2508 processors. The newly launched "Rangeley" chips, together with the original line of C23xx, C25xx and C27xx systems on a chip are positioned as low end offerings for communications and networking embedded devices. The mid-end of the communications/networking market lineup so far was represented by dual- and quad-core Ivy Bridge Gladden processors, available under the Pentium, Core i3 and Xeon E3 brands. Earlier this week Intel released successors of the Ivy Bridge Gladden CPUs, namely Core i5-4402EC, Core i7-4700EC and i7-4702EC.   [Full story]
This week AMD revealed prices of soon to be available socket AM1 Sempron processors. Socket AM1 is a new low-power platform, that combines "Kabini" systems on a chip in an upgradeable PGA package, and an inexpensive motherboard with the minimum amount of logic. The Fusion Controller Hub (FCH), responsible for interfacing between the APU and peripheral devices, on this platform is located on the processor. In theory, this allows owners of such socket AM1 boards to upgrade not only the CPU/GPU, but also the chipset.   [Full story]
Intel prepares to refresh Haswell desktop processors in the second and the third quarters of 2014. The "K" CPUs with unlocked multiplier are expected in Q3, and the microprocessors with locked multiplier will be released a quarter earlier. It was reported that the non-"K" chips will run a bit faster than current generation of Haswell products, and they will be compatible with existing 8-series and future 9-series chipsets. In February we published complete list of Haswell Refresh processors, that will be launched in Q2.   [Full story]
Intel recently expanded Core i5 family with a new low power microprocessor, dubbed as Core i5-4460T. In our February's story (linked below) we listed it along with Haswell desktop refresh SKUs, anticipated to arrive in the second quarter 2014. However, the i5-4460T is not a speed bump of existing product, and that may explain why it was released earlier than the refresh processors.   [Full story]
Intel this week introduced two systems on a chip from Atom C2000 series, aimed at networking and communications markets. New products, Atom C2308 and C2508, are built on the latest 22nm microarchitecture, incorporate up to 4 CPU cores, and support QuickAssist technology. Although the chips come with reduced clock speeds compared to older family members, their main advantage is not a performance, but a lower TDP at the same price point.   [Full story]
Intel prepares to launch the second wave of Atom "Bay Trail-T" products for Android and Windows 8.1 tablets. New Atom Z37x5 lineup will boost graphics performance by as much as 16%, and offer greater variety of SKUs, from budget-oriented Z3735E and up to the new flagship Z3795. The lineup will be comprised of at least 9 different models, most of which were planned to launch this quarter or at the beginning of the second quarter. We now have detailed specifications of seven SKUs, including Atom Z3735D, Z3735E, Z3745, Z3745D, Z3775, Z3775D and Z3795.   [Full story]
Last week a user "Randylahey" in hardwarecanucks.com forum purchased a used AMD CPU, and, as it turned out, it was a never seen before FX-670K model. The most peculiar thing is that the processor uses socket FM2. As expected for the FX brand, the CPU has the integrated GPU disabled. The chip was removed from HP Pavilion 500-266ea desktop PC, but it is identified as the FX-670K even on non-HP retail motherboards.   [Full story]
NVidia unveiled the upcoming GeForce GTX Titan-Z at their annual GPU Technology Conference. Billed as a supercomputer that will fit under your desk, the card includes two fully enabled GK110 GPUs. NVidia say it is engineered for next-generation 5K and multi-monitor gaming. They also say that, unlike other dual-GPU cards, the two GPUs are tuned to run at the same clock speeds, which they claim is not the case for traditional dual-GPU cards.   [Full story]
Updated Mar 26
Mar 25, 2014   Intel Celeron N2807 processor released
In February Intel launched the first "Bay Trail-M" mobile systems on a chip (SoC) with a new core stepping C0, that had a few important USB controller enhancements. New models were Pentium N3530, Celeron N2830 and N2930. Last week Intel added another mobile Celeron with stepping C0 to their product database. The new Celeron, dubbed N2807, is supposed to replace the N2806. Some characteristics of the N2807 were slightly improved, which should result in faster performance and a bit lower power consumption.   [Full story]
At the beginning of this month AMD announced socket AM1 platform, that would bring "Kabini" Accelerated Processing units to desktop computers. New feature of socket AM1-compatible parts is a PGA packaging, that should make it easier to build a low power system with desired board features and performance, not to mention an ability to upgrade such system in the future. The processors will be branded as Athlon and Sempron, have 2 or 4 CPU cores, and operate at frequencies at up to 2.05 GHz.   [Full story]
Intel this week expanded Xeon Phi 7100 family with the 7120D coprocessor. The 7100 series consists of premium Xeon Phi products, aimed at the most demanding High Performance Computing tasks. These co-processors are clocked higher and have more on-board memory than more budget oriented 5100 and 3100 series. The first two Xeon Phi models, the 7120A and 7120X, were packaged as an addon cards with and without cooling solutions. The latest Xeon Phi 7120D is a "dense form factor" product, that looks like a 5" x 6" PCB board with a 230-pin edge finger connection.   [Full story]
Intel this week revealed desktop roadmap enhancements and new technologies at a Game Developers Conference in San Francisco. The enhancements to the roadmap consist of several unlocked products, including Haswell-based Pentium and Core microprocessors, "Haswell-E" Extreme Edition CPUs, and Broadwell Core CPUs.   [Full story]
AMD in January announced their latest family of desktop processors, based on "Kaveri" core, and the company is hard at work on the next generation of APUs, due in 2015. The future Accelerated Processing Units for desktop and mobile computers will be codenamed "Carrizo", and they will be built on "Excavator" microarchitecture. Some details of these products were revealed this week by "Bright Side of News" hardware news site. They got hands on a document, titled "Preliminary BIOS and Kernel Developerís Guide (BKDG) for AMD Family 15h Models 60h-6Fh Processors", that described internal architecture of Carrizo APUs.   [Full story]
As may already know, Intel "Haswell Refresh" desktop products are coming in the second and third quarters of this year. Due to a minor (in most cases) speed boost, the microprocessors are expected to have a few percent better overall performance than Haswell counterparts, launched in 2013. Xeon E3-1200 v3 family for single-socket workstations is also going to be refreshed, and it is likely that they will also have slightly higher clocks speeds, while the rest of the features will stay unchanged. Specifications of forthcoming Xeon E3s are not known yet, but we now have a list of their model numbers.   [Full story]
Intel's current mid-range communications/networking platform, codenamed Crystal Forest Gladden, is going to be replaced with Glen Forest platform in a matter of weeks. The new platform will combine Cave Creek chipset with Core i5-4402EC, Core i7-4700EC and i7-4702EC microprocessors, based on a "Haswell Mobile Comms" core. As the core name suggests, they are essentially mobile Core i5 and i7 "Haswell" chips with a couple of features disabled, and a couple of extra features added.   [Full story]
Mar 17, 2014   Site update: Changes to CPU benchmark database
I've made modifications to the CPU selection page for the CPU benchmark database, and placed them online last week. It is now possible to reset manufacturer, family or socket selection without resetting other fields. Selected families, sockets, part numbers and models have different color. The list of displayed part and model numbers now depends on the ES / special / CPU-World only options. There were other tweaks that should make the page easier to use.
We reported in January on new Intel Atom Z3700 series processors for tablets, spotted in an OEM manufacturer roadmap. Among new models, the roadmap mentioned Z3775D, Z3735D/E and Z3735G/F "Bay Trail-T" chips, aimed at $200+, $150+ and $99+ tablet market segments. VR-Zone in February posted overview of Intel roadmap for this year, that not only confirmed these model numbers, but also listed unreleased Atom Z3645, Z3745D, Z3775 and Z3795 SKUs. Yesterday Intel published Product Change Notification (PCN), that had more details on upcoming Bay Trail-T models.   [Full story]
NVidia has launched the GeForce 800M top to bottom line-up of new notebook graphics cards. The line-up comprises of the entry-level Fermi based GeForce 820M, Maxwell based 830M through GTX 860M and high-end Kepler based GTX 870M and GTX 880M. The GeForce GTX 850M and higher SKUs introduce a new ultra-high efficiency mode, Battery Boost. Additionally, the GeForce 840 and higher products introduce two new GeForce Experience features, ShadowPlay and GameStream.   [Full story]
Intel announced Quark family of ultra low power embedded systems on a chip in September of last year. The first launched SKU was Quark X1000, a single-core and single-threaded processor, that integrated DDR3 memory controller, a number of communication interfaces and legacy logic. The chip was clocked at 400 MHz and had very low TDP, just 2.2 Watt. Up to this month the X1000 was the only processor in Quark family, but recently Intel added two new models: X1010 and X1020D.   [Full story]
in January we reported on a new Xeon Phi 7120D co-processor, spotted in the "Xeon Phi Product Family Performance" document, published by Intel. The co-processor, that apparently has the same features and performance as production 7120P and 7120X models but lower TDP, will be available in dense form factor. The co-processors in this form factor require an external cooling solution, and use a 230-pin connector, compatible with PCI-Express x24. According to our sources, the 7120D will be released in the first quarter 2014.   [Full story]
Intel this week launched E5-4600 v2, the last series of server processors, powered by Ivy Bridge architecture. The E5-4600 v2 CPUs have the same built-in features as Xeon E5-2600 v2 family, however they can scale up to 4 sockets. Compared to the first generation E5-4600 products, the v2 chips offer more cores, more cache, and greater memory bandwidth, which ultimately results in much higher performance. The CPUs also incorporate new float 16 instructions, APIC virtualization, Digital Random Number Generator, and support for OS Guard feature.   [Full story]
AMD has announced the latest member of the Rx 200 Series graphics cards. The Radeon R9 280 is based on the Tahiti Pro GPU, and is basically a rebranded HD 7950 with a slight increase of clock speeds. It is expected in the retail market over the next few weeks, at an MSRP of $279, and looks like it is set to complete the R9 200 line-up.   [Full story]
Back in January, AMD launched the first "Kaveri" APUs, featuring the latest "Steamroller" architecture with improved per-clock performance. The chips also incorporated newer Radeon R7 GCN graphics unit, that added support for Heterogeneous System Architecture. The processors had 4 CPU cores, and were priced at $150 and higher. Less expensive dual-core "Kaveri" APUs are planned to be launched in the second quarter, and in the meantime AMD prepared several budget "Richland" models, including A4-4020, A4-6320 and A6-6420K.   [Full story]
Intel's line of "Sandy Bridge-E" extreme CPUs was launched in Q4 2011, and consisted of i7-3930K and i7-3960X models. The processors had such innovated features as 6 Hyper-Threaded CPU cores, up to 15 MB of level 3 cache, and quad-channel memory controller, but on another side they were priced at premium and required new and expensive socket 2011 motherboards. The family was expanded with Core i7-3820K and i7-3970X SKUs in 2012, and replaced with Core i7-4xxx "Ivy Bridge-E" products last year.   [Full story]
AMD today announced socket AM1 platform for entry-level and low-power desktop PCs. The platform, previously known as socket FS1b, works with dual- and quad-core "Kabini" Accelerated Processing Units, built on "Jaguar" CPU microarchitecture and Graphics Core Next GPU architecture. The APUs also integrate a DDR3 memory controller, that supports data rates up to 1600 MHz. Furthermore, the chips integrate a Fusion Controller Hub (FCH), providing such functionality as USB 2.0/3.0 and SATA 6 GB/s data I/O interfaces, as well as VGA, DisplayPort and HDMI video interfaces.   [Full story]
Last weekend Intel added Celeron N2830 "Bay Trail-M" system on a chip to the official price list. According to the latest revision of "Bay Trail" specification update, the N2830 uses a new core stepping, that fixes a number of USB defects. Also, CPU and GPU clock speeds of the SoC are slightly different from the Celeron N2820. More details on the N2830 were published by Intel this week, and they revealed that the processor has two new features. Additionally, Intel published specifications of Celeron N2930 and Pentium N3530, that were also launched this week.   [Full story]
Last year Intel launched first "Haswell" desktop processors, that boosted performance of the CPU side by up to 10%, and significantly raised performance of the integrated GPU. This year will be less exciting as the company does not plan to introduce new desktop architecture. Instead, Intel will offer products, based on a year-old Haswell core. Preliminary specs of "Haswell" refresh CPUs indicate that almost all of them will have 100 MHz higher clock rate compared to their predecessors.   [Full story]
Mar 1, 2014   Site update: CPU-World server upgraded
With the invaluable help of hosting company, the CPU-World site has been moved to a new and much faster server. The website is fully operational now, but due to different web server configuration and different versions of other software I still need to update our monitoring and administration scripts. Once that is out of the way, I will change my priorities, and concentrate more on CPU and families descriptions, as well as on processors, missing from the site.
Feb 28, 2014   CPU-World server upgrade: downtime expected
We're going to move CPU-World.com website and the forum to better server, that should improve website speed and reduce page load times. The move will start around 11PM EST (4AM GMT on March 1st), at which time the website will be temporarily shut down. The downtime will last up to 1 hour. We apologize for the inconvenience.
Last month we reported on a few new Atom Z3700 "Bay Trail" systems on a chip, coming in Q1 and Q3 of this year. Based on that report, in the first quarter Intel may release Atom Z3775D for Android and Windows tablets, along with Z3735D processor for Android-based tablets. One of upcoming chips, Atom Z3775, recently appeared in specifications of new Hewlett Packard ProPad 600 tablet, announced at the Mobile World Congress 2014. The tablet can will also shipped with previously unheard of Atom Z3795.   [Full story]
Earlier this quarter Intel released two families of Xeon server microprocessors. Xeon E5-2400 v2 family, which debuted in January, is aimed at two-socket servers and workstations, while recently launched E7 v2 family targets mission critical and big data processing servers with up to 8 sockets. The last server family to launch in this quarter will be Xeon E5-4600 v2 for 4-way systems, utilizing socket 2011. According to recent Intel roadmap, the family consists of E5-4603 v2 and E5-4607 v2 entry-level chips, Xeon E5-4610 v2 and E5-4620 v2 standard parts, E5-4640 v2 and E5-4650 v2 "advanced" parts, and E5-4627 v2 frequency optimized model.   [Full story]
Two weeks ago Intel discontinued large number of desktop and mobile systems on a chip (SoC), based on "Bay Trail" core. Bay Trail chips were launched last September, and they featured such improvements as out of order microarchitecture, increased number of cores, higher clock speeds and new graphics unit. The second wave of Bay Trail products, released last November, also had Burst technology enabled. Most of these processors have been retired in a matter of months.   [Full story]
Alongside this weeks launch of the GeForce GTX 750 and GTX 750 Ti, NVidia also announced the imminent release of the GeForce GTX Titan Black. The Titan Black, NVidia's new flagship desktop product, will take the best features of the GTX Titan and the GTX 780 Ti, and improve on them. With the Maxwell architecture having recently been introduced in the GTX 750 and GTX 750 Ti, it is likely this will be the last new model to use the Kepler architecture.   [Full story]
Intel this week announced Xeon E7 v2 processors, boasting new architecture and manufacturing process, as well as quite a few new features, that greatly enhance their performance. The best performing chip in the E7 v2 lineup is a Xeon E7-8890 v2, which has 15 cores, 37.5 MB L3 cache, and 2.8 GHz base clock. The CPU supports Turbo Boost technology, that can "overclock" it up to 3.4 GHz. However, there was faster unannounced E7-8895 v2 SKU launched this week, that has all the features of the E7-8890 v2, with the exception of 200 MHz higher Turbo Boost frequency.   [Full story]
NVidia has launched their first Maxwell GPU, the GM107, in the form of two graphics cards, the GeForce GTX 750 and GTX 750 Ti. The GM107 GPU is built using the tried and trusted 28 nm process. The cards are intended to fit between the GTX 650 and GTX 660, replacing the GTX 650 Ti.   [Full story]
Intel this week launched the second generation of Xeon E7 microprocessors for mission critical applications and big data computing. Branded as Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2, the CPUs provide up to 50% more cores, up to 25% more cache, up to 4x times of I/O bandwidth, and up to 3x times of supported RAM than the first generation of Xeon E7 products. These improvements, combined with microarchitecture enhancements, allow Xeon E7 v2s to perform up to 2 times faster than their predecessors.   [Full story]
Next quarter Intel is going to launch a bit faster models of Haswell-based processors for LGA1150 platform. It is expected that the company will refresh complete desktop lineup, starting from low cost Celeron microprocessors and up to premium Core i7 CPUs. All in all, more than 20 different SKUs will be released. We already reported on some of these models, but now we have a complete, or almost complete list of these processors, including their clock speeds.   [Full story]
DigiTimes last week reported that the next generation of Intel processors, built on Broadwell 14nm architecture, was delayed. The Broadwell processors were scheduled to launch in Q3 2014. However, due to slow sales of Haswell-based products, the Broadwell parts were pushed back by one quarter, and the current plan is to launch priority SKUs in the fourth quarter of the year, while the remaining models will be available in Q1 2015. Today VR-Zone revealed more detailed information on a launch schedule of mobile CPUs, that confirmed DigiTimes' report.   [Full story]
AMD is about to release ultra low power "Kabini" Accelerated Processing Units for a PGA socket, originally called FS1b. Based on earlier reports, the processors with model numbers 5150 and 5350 (or A4-5150 and A6-5350) will be marketed as Athlons, while slower 2650 and 3850 models will be sold as Semprons. The socket FS1b will be also re-branded, and it is going to be called Socket AM1. Although the new Athlon and Sempron APUs are expected to perform slower than their socket FM2 counterparts, the "Kabini" chips will have much better power consumption and should will be very inexpensive.   [Full story]
Intel launched a number of embedded products this year, including Celeron G1820TE for desktop platform, Celeron 2000E and 2002E for mobile platform, and a few low-power Xeon E5-2400 v2 CPUs for server and communications platforms. We also reported that the company will release embedded Xeon E5-4624L v2 CPU in Q1 2014. Furthermore, Intel plans to expand embedded server and communications platforms with three new models, Core i5-4402EC, Core i7-4700EC and i7-4702EC.   [Full story]
Last September Intel released budget Bay Trail systems on a chip for mobile and desktop markets, branded as Celeron and Pentium. The processors were manufactured on 22nm technology, and featured such enhancements as greater number of CPU cores, higher clock speeds, beefed up graphics unit, not to mention an out-of-order microarchitecture, that improved per-clock CPU performance by up to 30% compared to their predecessors. The first wave of Bay Trail SKUs was followed in November by Burst-enabled models, that offered even higher operating frequencies.   [Full story]
In January AMD added to their product database three "Kabini" systems on a chip (SoC), including dual-core E1-2200, together with quad-core E2-3800 and A4-5100. One of these processors, the E2-3800, is now shipped in HP Pavilion 23-g010 All-in-One PC, which is offered through the HP online store. The E2-3800 is also referenced in specifications of HP 245 G2 and HP 255 G2 HP notebooks. Perhaps, more interesting is the fact that that the specs also mention (PDF file) two unreleased models, an E1-2150 and an A4-5050.   [Full story]
In October of last year we reported on new Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2 CPUs, expected to launch this quarter. The processors will utilize Ivy Bridge architecture, and will have 50% more cores and 25% more cache than the first generation of Xeon E7 products, that were built on Nehalem architecture. In total, there will be 21 different models. Using multiple sources, like this one (PD file), we were able to come up with specifications for most of these processors.   [Full story]
Advanced Micro Devices this week revealed more information on ARM-based "Seattle" server products. New Opteron processors will be manufactured on 28nm process, and will be branded as A1100 series. According to AMD, the A1100 server chips will have from 4 to 8 CPU cores, integrate the latest SATA, Ethernet and memory interfaces, and offer 2x - 4x times of performance of Opteron X-Series APUs. AMD is also going to release "Seattle" development platform, which includes an evaluation board and Linux-based software tools.   [Full story]
Last September Intel launched the first "Ivy Bridge" Xeons for dual-processor servers, based on socket 2011. Branded as E5-2600 v2 series, these microprocessors had up to 50% more cores and up to 25% more cache than previous generation of Xeon E5s. More socket 2011 products are coming this quarter, and this time the manufacturer will release Xeon E5-4600 v2 series, designed to work in quad-processor systems. In addition to 8 server SKUs, that we already reported on, Intel is going to introduce embedded Xeon E5-4624L v2.   [Full story]
Intel offers three variations of Xeon Phi coprocessors: 31xx for compute-bound workloads, 51xx for memory/capacity bound tasks, and high-performance 71xx series for the most demanding applications. The high-end 71xx lineup currently consists of 7120P and 7120X models, implemented as PCI-Express cards with and without a passive thermal solution. In addition to these products, Intel prepares 7120D coprocessor in a dense form factor.   [Full story]
VideoCardz.com have published specs for an NVidia GeForce GTX Titan replacement, the GTX Titan "Black Edition". They also showed possible specs for a dual-GPU GTX 790. The Titan Black Edition is said to use a fully enabled GK110 GPU, as used in the GTX 780 Ti, but with more memory. The GTX 790 is expected to have two GK110 GPUs, with more CUDA cores per GPU than a GTX 780.   [Full story]
As was reported by Digitimes in December, Intel is going to release budget versions of Bay Trail-T systems on a chip (SoC) later this year. Model numbers and possible launch window of these processors were revealed by an Emdoor OEM manufacturer roadmap, published by padnews.cn. Based on the roadmap, in the course of the first half of 2014 Intel will launch Atom Z3775D and Z3775D series models, covering different tablet price points.   [Full story]
AMD today launched two server microprocessors for two- and quad-socket servers. New Opteron 6338P and 6370P chips integrate 12 and 16 CPU cores respectively, utilize lower power envelope, and offer better performance per watt per dollar than existing products from the same family. The 6338P model is priced at $377, and the 6370P is priced at $598.   [Full story]
AMD launched new Steamroller architecture and new "Kaveri" core this month, and the first two "Kaveri" models are already available in stores. Another part, the A10-7600 APU, should be available later in the first quarter. As reported by VR-Zone, the company also plans to release one quad-core and 2 dual-core SKUs in the second quarter, expanding A10-Series family with a mid-power A10-7800, and launching A4-7300 and A6-7400K budget products.   [Full story]
Last Sunday VR-Zone published details of Accelerated Processing units for socket FS1b. According to VR-Zone, the processors with E1-2650, A4-3850 and A4-5350 model numbers will be based on "Kabini" core, have up to 4 "Jaguar" CPU cores, and come with GCN graphics. This not the first report on these chips, and the previous one was posted in December. The earlier report already revealed most important features of Sempron 2650, 3850 and Athlon 5350 APUs, that have similar specifications to the E1-2650, A4-3850 and A4-5350.   [Full story]
Intel today added 9 Haswell-based mobile microprocessors to the official price list. The list of new models includes Core i5-4310M, i5-4310U, i5-4340M and i5-4360U, along with Core i7-4610M, i7-4810MQ, i7-4860HQ, i7-4910MQ, and Core i7-4940MX Extreme Edition. Based on brief specifications in the pricelist, the CPUs provide only a minor speed boost over their predecessors, launched in the summer and the fall of last year.   [Full story]
Current generation of high-end AMD Opteron 6300 series microprocessors for 2 and 4 way servers was launched in November 2012. Codenamed "Abu Dhabi", these chips packed up to 16 "Piledriver" CPU cores, 16 MB of cache, and supported the latest (at the time) x86 extensions. According to AMD official roadmap, the "Abu Dhabi" processors will be followed by "Warsaw" products, also produced on 32nm technology. There are not many details of Warsaw processors right now. It is known that they will have 12- and 16 CPU cores, and offer better performance per watt ratio than "Abu Dhabi" parts.   [Full story]
New AMD "Kaveri" processors, that were announced last week at CES show in Las Vegas, have been launched. The first two retail models are A10-7700K and A10-7850K, both boasting 4 "Steamroller" CPU cores and HSA-enabled Radeon R7 graphics. Furthermore, AMD introduced A8-7600 quad-core model with 65 Watt and 45 Watt TDP. The A10-7700K and A10-7850K are available in stores now. The boxes come with a free coupon for Battle Field 4 PC game. The A8-7600 has an official price of $119, and it will be available later in Q1 2014.   [Full story]
New CPUs: AMD A8-7600
Rather than transitioning to new architecture this year, Intel plans to refresh current desktop lineup with slightly faster versions of Haswell-based CPUs. We already reported on Core i5-4460, i5-4590, i5-4690 and Core i7-4790 parts, coming in Q2 2014. According to Digitimes, the refresh will not be limited to Core i5 and i7 products, and we will see new Core i3, Pentium and Celerons as well.   [Full story]
NVidia recently announced the upcoming release of the Tegra K1 mobile processors. Available in both 32-bit and 64-bit, the processors will use an ARM CPU core and a Kepler GPU. The 32-bit version will have an NVidia 4-Plus-1 Quad-Core ARM Cortex-A15 "r3" CPU with 4 cores, clocked at 2.3 GHz. The 64-bit "Denver" variant will have an NVidia-designed 64-bit dual Super Core CPU. It will be based on ARMv8 microarchitecture, and will be pin-compatible with the 32-bit version of the Tegra K1.   [Full story]
Earlier this week we reported on a launch of Xeon E5-2400 v2 family for Romley-EN (socket B) platform. Utilizing all features and enhancements of Ivy Bridge microarchitecture, new "Ivy Bridge-EN" CPUs have more cores, larger L3 cache size, and higher clock speeds than previous generation of Sandy Bridge-EN based microprocessors. The E5-2400 v2 products work in single- and dual-processor servers, but this week Intel also introduced several Ivy Bridge-EN SKUs, targeting only single-processor embedded systems.   [Full story]
AMD introduced three new mobile graphics cards at CES 2014. The initial launch of second generation GCN based mobile GPUs included 3 models - Radeon R9 M290X, R7 M265 and R5 M230. The R9 M290X and R7 M265 are rebrands of existing GPUs, while the R5 M230 uses a new entry level GPU, with specs similar to those of the Radeon HD 8500 series. At launch time, these models could be found in systems from Alienware, Clevo, Lenovo and MSI.   [Full story]
AMD this week formally introduced Kaveri products, and published details of their performance and features. Forthcoming "Kaveri" A10-7700K and A10-7850K processors, that will be available in less than a week, will combine two new pieces of technology: "Steamroller" CPUs cores with enhanced per-clock performance, and R7 series GCN graphics unit, which comes with such innovative technologies as Heterogeneous System Architecture (HSA) and Mantle programming interface.   [Full story]
Intel today launched second generation of Xeon E5-2400 products for Romley platform, featuring more cores, larger caches and higher clock speeds, while operating in the same power envelope as previous generation of server chips. New microprocessors are built on Ivy Bridge-EP core, that provides a number of performance and power enhancements, such as internal speed optimizations, new Float 16 instructions, APIC virtualization, and lower idle power consumption. The chips also have such security technologies as Secure Key and OS Guard. The E5-2400 v2 CPUs are targeted at dual-processor systems, and use socket B (LGA1356).   [Full story]
In December we reported about upcoming refresh of Haswell desktop processors. In the second quarter 2014, Intel will launch new mainstream desktop Core i5 and i7 SKUs, featuring slightly better performance than their predecessors. Later this year, the manufacturer will also introduce faster mid-power Haswell products for socket 1150. One of them, Core i5-4460S, was recently spotted in CPU support lists of ASRock motherboards, such as B85M-ITX, H81M-GL and several others.   [Full story]
This week AMD added specifications of several Kabini systems on a chip to their product database, which usually means that the processors were either launched, or are about to be launched in the near future. New chips are A4-5100, E1-2200 and E2-3800 low-power "Kabini" APUs for tablets, hybrids and entry-level notebooks. The E1-2200 and A4-5100 SoCs are 50 MHz faster versions of existing E1-2100 and A4-5000.   [Full story]
In the next two weeks, AMD is expected to launch new Steamroller architecture, that should boost the CPU performance of AMD Accelerated Processing Units (APUs). The first products, that will make use of the microarchitecture, will be desktop "Kaveri" processors, which will combine up to 4 "SteamrollerB" CPU cores and R7 series graphics on a chip. The specs of two known desktop Kaveri APUs were published about a month ago. According to them, A10-7700K and A10-7850K will have 4 CPU cores, 4 MB L2 cache, no L3 cache, and 10% lower operating frequencies than their "Richland" predecessors.   [Full story]
Intel today lowered prices on a handful of Core i5 and Core i7 mobile processors, as well as on a few desktop CPUs for All In One systems. The company also dropped prices on budget mobile Celeron CPUs, cutting them by at least 13%, and as much as 22% - 26% on selected models. Overall, the prices were cut by an average of 13.4%.   [Full story]
Dec 22, 2013   Happy Holidays! / Leaving for vacation
I will be on vacation the week of Christmas, and I will not be able to do any site updates. During that week I will provide basic site support, such as site monitoring, approving comments, and forum moderation. I wish all site visitors Merry Christmas and Happy Holidays!
Three new budget Intel processors are about to be released in the next few months. Celeron G1820, G1820T and G1830 will be the first Celeron-branded SKUs, that make use of Haswell architecture. Haswell Celerons were originally planned to be launched in the first quarter 2014, possibly in January, however two of them are already available for sale in US from various online retailers, including Rakuten.com (former Buy.com), and from third party sellers at Amazon.   [Full story]
Last month we reported on preliminary lineup of Broadwell mobile microprocessors. In that report we mentioned several variations of Broadwell chips, such as "Y", "U" and "H" versions, having different power requirements. The "H" and "U" chips will be available with more than one type of GPU. Furthermore, there will be 1- and 2-chip versions. All "Y" and "U", and a handful of "H" parts will be produced as systems on a chip (SoC), that will integrate the CPU and GPU units, as well as the PCH-LP chipset. Later in the article we will refer to them as SoCs, 1-chip or single-chip processors.   [Full story]
Recently leaked AMD desktop roadmap stated that next year the company will introduce low power "Kabini" systems on a chip (SoC), and release new mainstream APUs, based on "Kaveri" core. Only quad-core "Kaveri" products will be available at launch, and dual-core mainstream processors will be built using older "Richland" core. Prices and brief specs of all these microprocessors were published last week by SweClockers.com, they also revealed new A4-4020 "Richland" SKU.   [Full story]
Last year Intel launched "Romley" platform, that has being used with the first two generations of Xeon E5-2600 processors. The next generation of Xeon E5 products, based on "Haswell-EP" core, will utilize "Grantley" platform, that will employ slightly different variation of socket 2011. Haswell-EP microprocessors will incorporate up to 14 CPU cores and up to 35 MB of last level cache. They are expected to support all Haswell technologies, along with new RAS features. In the summer/fall of 2015 Intel is going to release 4th generation of E5-2600 CPUs, built on 14 nm "Broadwell" microarchitecture.   [Full story]
AMD has announced the Radeon R7 260 GPU, a cut-down version of the Bonaire GPU used in the Radeon R7 260X graphics cards. The card will be released in mid January 2014, although the exact date is yet to be revealed. For price and performance, the new card will replace the Cape Verde based Radeon HD 7770, and it is likely to compete with NVidia's GTX 650 and 650 Ti, both of which are similarly priced.   [Full story]
Digitimes this week revealed Intel roadmap for tablet- and smartphone oriented systems on a chip, based on recently introduced Bay Trail-T core, and on future 14nm technology. Several Bay Trail-T processors from Z3700 series already launched this fall, and by the end of the year Intel is expected to release new models, that will offer support for Android operating system.   [Full story]
We recently reported on several third-generation business class APUs, being shipped in Hewlett Packard's desktop PCs. Based on "Richland" core, the A4-6300B, A6-6400B, A8-6500B and A10-6800B processors have all of the features of consumer-class A4-6300, A6-5400K, A8-6500 and A10-6800K products with the exception of unlocked multiplier. In addition to these "B" APUs, AMD also plans A10-6790B model, that should fill the gap between the A8-6500B and A10-6800B quad-core SKUs.   [Full story]
In the past few years we got accustomed to Intel launching new microarchitecture each year, using new technologies and core enhancements, as well as clock speed increases to boost performance of each new generation of chips. The things will be slightly different next year, because, with the exception of some unlocked Broadwell-K models, desktop Intel microprocessors will still be built on Haswell architecture, and the only new features will be introduced in 9-series chipsets.   [Full story]
The next Intel microarchitecture, codenamed "Broadwell", will be mostly aimed at mobile computers, and will offer increased CPU and GPU performance, along with significant power savings. The first mobile processors to switch to new architecture will be "H", "U" and "Y" types. Because Broadwell products will be released in the second half of 2014, it is too early to talk about specific model numbers and frequencies, however we already have preliminary information on some Broadwell lineup features.   [Full story]
Dec 10, 2013   New layout of CPU pages
I have a new CPU page design ready, that is a bit wider, uses wider side bar, and no longer uses tabs. While the layout has changed, due to all tabs been flattened into one long page, it still should look familiar enough to those who used CPU pages before. The new design is now enabled for many modern families. The rest of the families will be converted within the next two days.
AMD is getting ready to launch in January two Kaveri Accelerated Processing Units, that, by some reports, should boost both CPU and GPU performance by up to 20% - 30% compared to A10-6000 series APUs. Model numbers and specifications of new Kaveri processors emerged last week, and the A10-7700K and A10-7850K chips are now available for pre-order in US. Based on the pre-order prices, the processors are $12 - $34 more expensive than boxed A10-6800K products. What's more, unheard of before A4-6320 and A6-6420K dual-core APUs are also available for pre-order.   [Full story]
In October we reported on several Haswell-based mobile microprocessors, planned to be released in Q1 2014. The list included Celeron 2957U and 2981U, along with Pentium 3558U and 3561Y ULV chips. The main new feature of these microprocessors was support for Wireless Display technology, while the rest of the specifications was identical to Celeron 2955U, 2980U, and Pentium 3556U and 3560Y. Two of upcoming models, the 2957U and 2981U, were launched in November, and the other two were added to Intel's product database this week.   [Full story]
AMD "Richland" processors were launched in June 2013, and since then the company released about a dozen of consumer models, ranging from a dual-core 370K CPU and up to a quad-core A10-6800K APU with unlocked clock multiplier. In addition to consumer processors, AMD also planned 4 business class products with model numbers A4-6300B, A6-6400B, A8-6500B and A10-6800B. Recently these processors appeared in HP Compaq Pro 6305 Desktop PCs.   [Full story]
Dec 5, 2013   Site Update: Website design tweaks
I changed CPU family pages to use wider design without tabs, and I plan to do similar changes to CPU pages in the near future. I also converted some CPU news pages, and all S-spec pages to somewhat different design, that has wider right sidebar. Both designs (with a narrow and wide sidebar) will coexist on the site, and will be used on different pages depending on their content type. Once I'm done with CPU pages, I will convert all comparison pages to newer design without tabs, and then convert the rest of the site's pages.
Last month AMD announced some features and a launch schedule of the next generation of Accelerated Units, codenamed Kaveri. More details on these processors surfaced last week, when Prohardver.hu published several slides with specifications of A10-7700K and A10-7850K models. Future APUs are not clocked as high as their predecessors, however they are better optimized for CPU, GPU and compute intensive tasks, owing it to newer microarchitecture, the latest R7 series graphics unit, and hUMA (heterogeneous Uniform Memory Architecture) support.   [Full story]
AMD updated their official roadmap last month, and announced "Kaveri", "Beema" and "Mullins" desktop and mobile processors coming next year. A bit more information about AMD desktop plans for the next two years emerged in a leaked roadmap slide, posted in Hardforum.com. According to the slide, AMD prepares to launch Kaveri and Beema APUs in 2014. Also in 2014, AMD will introduce "Kabini" processors for socket FS1b. A successor to Kaveri, codenamed "Carrizo", will be available in 2015.   [Full story]
Next quarter Intel is going to introduce "Ivy Bridge-EP" microprocessors, branded as Xeon E5-4600 v2, for quad-socket servers. In October we published specifications of two Xeon E5-4600 v2 models, E5-4610 v2 and E5-4657L v2, that were found in a CPU support list on the ASRock website. Since then three new Xeon E5 microprocessors were added to the support list: E5-4607 v2, E5-4627 v2 and E5-4650 v2.   [Full story]
Intel's embedded low power lineup consists mostly of Haswell- and Bay Trail-based microprocessors, and only Celeron 1020E is yet to be transitioned to newer architecture. Two possible successors of the 1020E are Celeron 2000E and 2002E, and both recently appeared in published motherboard or system specifications, such as specs of Fanless Intel Medical Station (PDF file), and MB 8301 industrial mini-ITX motherboard (PDF file).   [Full story]
Next year Intel is going to refresh existing Haswell lineup, and introduce two 9-series chipsets, Z97 and H97. Closer to the end of the year the company will also launch first desktop "Broadwell" microprocessors, codenamed Broadwell-K, for socket 1150 platform. As the name suggests, the processors will have unlocked clock multiplier. Perhaps, the most interesting feature of new products will be Iris Pro (GT3) graphics with integrated 128 MB eDRAM memory.   [Full story]
Earlier this month Intel quietly launched ultra low voltage Celeron and Pentium systems on a chip (SoC), based on "Bay Trail-M" core and aimed at mobile devices. Several key improvements in new mobile processors were higher CPU clocks, added support for Burst technology, and increased maximum GPU speed. Mobile Celerons and Pentiums were not the only "Bay Trail" products released this month. Last week Intel added specifications of desktop "Bay Trail-D" chips to their product database.   [Full story]
Two months ago Intel launched Xeon E5-2600 v2 microprocessors, boasting such new features as more efficient Ivy Bridge architecture, increased number of cores and larger L3 cache. While the E5-2600 v2 lineup included more than 20 different server and embedded SKUs, only Xeon E5-2695 v2 and E5-2697 v2 had the maximum of cores (12). Priced at $2336 and $2614, these were also the most expensive from the lineup. Recently some US online listed another 12-core model, E5-2651 v2.   [Full story]
Next generation of microprocessors for tablets will be codenamed "Cherry Trail-T", and will be manufactured on 14nm technology. Compared to currently produced "Bay Trail-T" products, future chips will feature higher clock speeds, better graphics performance and more memory bandwidth. These and other details on "Cherry Trail-T" processors were published this week by VR-Zone China.   [Full story]
AMD yesterday provided details on upcoming mobile microprocessor cores, including low power "Beema" and ultra-low power "Mullins" cores. The company also revealed performance estimates of future APUs, and released updated desktop and mobile CPU roadmaps for 2014.   [Full story]
AMD "Kaveri" core is one of the most anticipated AMD products next year, partly because it is based on a new "Steamroller" microarchitecture, which is expected to fix shortcomings of Bulldozer/Piledriver processors and improve CPU performance, and partly because it will support Heterogeneous System Architecture (HSA), expected to boost GPGPU performance. "Kaveri" chips were previously reported to become available in February 2014. Yesterday AMD revealed exact launch date, as well as a model number of the flagship desktop Kaveri APU.   [Full story]
There is not much known about the next generation of Intel processors, codenamed "Broadwell", and much less known about their successors, based on "Skylake" microarchitecture. It is known that "Skylake" products will be released in 2015, and will use the same manufacturing process as "Broadwell" CPUs. Additional details on "Skylake" Xeon E3-1200 v5 server chips and associated "Greenlow" platform were published this week by VR-Zone China.   [Full story]
In September Intel launched the latest generation of "Bay Trail-M" ultra low voltage processors for netbooks and mobile devices. New mobile CPUs, branded this time as Celeron and Pentium, offered up to 2x times better CPU performance, and up to 3x time faster graphics, while operating in the same power envelope as their "Cedar Trail" predecessors. As if those improvements were not enough, Intel this week refreshed Bay Trail-M lineup, releasing faster versions of Celeron and Pentium products.   [Full story]
Last week AMD announced A10-6790K APU, which was positioned as a slightly slower, but also less expensive version of the flagship A10-6800K chip. Today the 6790K was added to AMD's official price list. Furthermore, the company added prices of FX-9370 and FX-9590 microprocessors for AM3+ platform. Finally, AMD reduced prices of A6-6400K, A8-6500 and A8-6600K models by 12% - 13%. The A6-6400K now costs $62, and the price of A8 SKUs was slashed from $112 to $97.   [Full story]
Yesterday Intel published several Product Change Notifications, advising their customers about forthcoming discontinuation of a number of Atom and Itanium microprocessors. Retired Atom models are D2500, D2550, D2560, N2600 and N2800. These dual-core processors were introduced in a time frame from September 2011 to November 2012, and all of them are based on one generation old "Cedar View" core. End-of-lifed Itanium models are 9310, 9320, 9330, 9340 and 9350.   [Full story]
Last Sunday VR-Zone China posted a picture of what seems to be an engineering sample of AMD "Kaveri" APU. The processor has part number ZD356195I4468, and the back of the chip resembles pinout of "Richland" based Accelerated Units with a couple of pins added here and there. VR-Zone have not disclosed any specs for the sample, but some information can be deduced from the part number. Assuming that Kaveri engineering samples use the same naming conventions as older APUs, the "D35" can be decoded as "a desktop processor with 3.5 GHz clock speed".   [Full story]
Intel today updated the official price list, adding several ultra-low voltage microprocessors to the mobile lineup. New Celeron products are 2957U, 2981U, N2806, N2815, N2820 and N2920, and they are priced from $132 to $137. All new processors are implemented as a system-on-a-chip (SoC), and they integrate on a die up to 4 CPU cores, a graphics unit, along with chipset logic.   [Full story]
In June, AMD leaked model numbers of presumably upcoming mobile processors in a list of supported products for the Catalyst Windows 8.1 Preview driver. The list mentioned E1-2200, E1-2600, E1-2650, E2-3100, A4-5100, A4-5150, A6-5250 and A6-5350 Accelerated Processing Units, based on "Kabini" low-power core. Their features were not known at the time, but now AMD revealed details of three APUs from the list. Based on published information, AMD E1-2200 and A4-5100 are incremental upgrades to existing E1-2100 and A4-5000 models.   [Full story]
The first Intel "Haswell" quad-core processors, along with the microarchitecture, were launched in June of this year, and they were followed by Pentium and Core i3 "Haswell" CPUs in September. The last processors to transition to Haswell architecture are desktop Celerons, which are expected to launch to Q1 2014. The initial Celeron lineup will consist of 3 SKUs, including standard-power G1820 and G1830, and low-power G1820T. Two of these processors are now available for pre-order in US from several online shops.   [Full story]
AMD announced new Accelerated Processing unit this week. The AMD A10-6790K is a quad-core processor with integrated Radeon HD 8000 series graphics, that should perform a bit slower than the flagship A10-6800K, while on the other hand it is expected to cost less. Specifications of this model already appeared on AMD's website, but the official pricing is not known yet.   [Full story]
One of the highly anticipated microprocessor products, coming in the next few months, is AMD "Kaveri", primarily because it is going to introduce much improved "Steamroller" processor core and new GCN graphics. AMD's current desktop roadmap places the launch of "Kavery" APUs in the second half of 2013, however some reports indicated that the processors will be available only next year. Chinese VR-Zone news site revealed this week additional details on a launch schedule of Kaveri Accelerated units.   [Full story]
In September Intel introduced first Xeon E5 server products, built on Ivy Bridge-EP core. These processors utilize socket 2011, integrate up to 12 cores, up to 30 MB L3 cache, and boast several new embedded technologies and interfaces. New Xeons are aimed at single- and dual-socket servers and workstations, however Intel also prepares microprocessors for quad-socket systems. Coming in Q1 2014, the CPUs will be branded as E5-4600 v2, and will have a feature-set similar to already launched Ivy Bridge-EP parts.   [Full story]
AMD have launched the new high-end single-GPU graphics card, the Radeon R9 290X. Priced to compete with NVidia's GeForce GTX 780, the new card makes no pretense at trying to compete with the ultra high end GTX Titan, although in some benchmarks it does get close. While many of the R7 and R9 cards released two weeks ago are rebrands of existing Radeon HD 7000 series cards, the R9 290X uses the entirely new Hawaii XT GPU, and supports TrueAudio technology, available in the R7 260X and in the soon to be seen R9 290.   [Full story]
Next quarter Intel is going to launch new budget ultra-low power microprocessors, based on Haswell architecture. Celeron 2957U, Celeron 2981U, Pentium 3558U and Pentium 3561Y are planned as a replacement for 2955U, 2980U, 3556U and 3560Y parts, that were introduced in September. Model numbers of new products are almost identical to their predecessors, which suggests that their major specifications will be similar. The new CPUs will support Wireless Display feature, and, as far as we know, that is the only difference between new and old products.   [Full story]
Earlier this month we published some details on Intel Xeon E7 microprocessors, coming in Q1 2014. The processors for 2-way, 4-way and 8-way systems will be branded as Xeon E7-2800 v2, E7-4800 v2 and E7-8800 v2, and will be built on Ivy Bridge microarchitecture. Besides new architecture, the CPU will feature up to 15 cores, up to 37.5 MB of last level cache, and support for larger amounts of physical memory. They will also add DMI 2.0 and PCI-Express 3.0 interfaces.   [Full story]
Oct 18, 2013   Related S-Specs section added
Many S-Spec pages now show "Related S-Spec numbers" section. That section contains a list of all known S-Specs, including pre-production ones, that were produced for specific part. If the section is not displayed it means that we know only one S-Spec number, associated with the part.
Yesterday Intel released new Product Change Notification (PCN), detailing changes to packaging of socket 2011 and socket 1356 server microprocessors. Besides the description of the change, the PCN also had a list of affected CPUs, comprised of released Xeon E5-2600 v2 models, along with unreleased Xeon E5-1400 v2, E5-2400 v2 and E5-4600 v2 SKUs. The document also listed three not yet released Pentium-branded processors for workstations.   [Full story]
Several new socket FM2 processors, based on a "Richland" core, were released not long ago, or about to be released in the near future. As we previously reported, boxed AMD A4-6300 dual-core APU emerged in US stores last week, while A8-6500T and A10-6700T low-power quad-core processors are still on pre-order in many online US shops. AMD also plans A10-6790K Accelerated Unit, which recently surfaced on the ShopBLT.com website.   [Full story]
New CPUs: AMD A10-6790K
Alongside the previously reported release of the R9 series GPUs, AMD have also launched the full range of R7 series mainstream GPUs. The Radeon R7 260X, R7 250 and R7 240 bring with them another new feature, AMD TrueAudio, a technology not available on the high end R9 cards. The cards also support the new Mantle API, mentioned in our earlier article on the R9 cards.   [Full story]
Intel this week released Atom E3800 family of embedded processors, based on recently launched Silvermont micro-architecture. New Atom products are implemented as systems on-a-chip (SoC), and feature up to 4 CPU cores, 7th generation (Ivy Bridge) graphics, and support for DDR3 memory. They also come with a variety of on-chip interfaces, including PCI-Express 2.0, USB 3.0 and SATA.   [Full story]
AMD this week launched the first of their new high-end to mainstream R9 graphics cards. The R9 280X and R9 270X are basically rebadges of earlier 7000 series cards, albeit with a new naming convention and modified specs. What makes them stick out from the previous generation of GPUs is that they are the first cards to ship with drivers supporting the new Mantle API, a low level API enabling programmers to access the graphics hardware at a lower level than previous APIs.   [Full story]
In September we reported that newer boxed versions of FX-9000 extreme microprocessors will soon be available bundled with a liquid cooling kit. Launched in June, the FX-9370 and FX-9590 packed 8 cores, 8 MB of L2 and 8 MB of L3 cache on a chip. The processors run at very high 4.4 and 4.7 GHz clock rates, and can reach even higher 4.7 and 5 GHz frequencies when a Turbo Core feature is activated. Since the FX-9370 and FX-9590 are effectively overclocked siblings of the FX-8350 model, they dissipate significantly more power and require a liquid cooling kit to operate.   [Full story]
Intel Xeon Phi line of coprocessors was announced last November, and it initially consisted of a few SKUs from mid-range 5100 and high-end 7100 families. One of the first Xeon Phi models was 5110P, that packed 60 Pentium-like x86 compatible cores, and 8 GB of GDDR5 RAM on a PCI-Express card. The co-processor was clocked at 1.053 GHz, and delivered double-precision performance of slightly more than 1 TFLOPS.   [Full story]
Oct 6, 2013   Tweaks to S-Spec pages
I tweaked design of S-spec pages to use slightly wider layout, and removed the tab interface from the pages. The tab interface uses javascript code, that build tabs on the fly, and this causes the pages to re-render once they are loaded. Without tabs the pages may feel like loading faster, and they are less dependent now on javascript support in the browser. Besides these visible modifications to the S-Spec web pages, there was also one very important internal change to the S-Spec database: the data is now stored differently, and I have a new data entry interface working, which allowed me to add about 70 new S-Spec codes in the last two days. More S-spec codes will be added next week.
New retail AMD processor recently surfaced in US stores. AMD A4-6300 Accelerated Processing Unit was rumored to launch in June, but so far it was shipped only in OEM desktop systems, such as HP Pavilion 500-003a. Last month the APU was listed for pre-order in smaller US online stores, and now it is available for sale from NewEgg.com.   [Full story]
A report on Techpowerup hints at the final clock speeds of the new Radeon R9-290X, based on an unpublished GPU-Z screenshot. They say the new cards will have a base clock speed of 1050 MHz, with no turbo clock speed. It will however use a technology that reduces the clock speed, by an undisclosed amount, depending on loads and temps, so is not that different from recent cards that have base and boost clock speeds.   [Full story]
In September Intel launched first "Ivy Bridge" server processors, and more of them will be released in the coming months. The first "Ivy Bridge" Xeons were E5-1600 v2 and E5-2600 v2 series, designed for single- and dual-socket LGA2011 systems. Manufactured on 22nm technology, these products offered higher or considerably higher overall performance while operating in the same power envelope as the current generation of "Sandy Bridge" server CPUs. More microprocessors, such as Xeon E5-4600 v2 for quad-socket configurations and Xeon E5-2400 v2 for socket 1356, will be introduced next year.   [Full story]
Terms and Conditions · Privacy Policy · Contact Us (c) Copyright 2003 - 2014 Gennadiy Shvets

Search CPU-World

Search site contents:

Identify part

Identify CPU, FPU or MCU: