[Oficial] Intel Core "all" Series

[Jorge-Vieira]

Durante o tempo dos K7 e K8 a Intel andava a comer pó e a fazer jogos sujos.
Esses foram os tempos aureos da AMD, a partir daí foi uma travessia no deserto que ainda continua e, até mesmos os mais defendores da marca apontam isso como principal critica por não se ver nada de novo nem sequer um refresh.
Não sei se no futuro vamos ver novamente a AMD a ter um longo periodo aureo de dominio nos CPUs como no passado, mas para bem de todos era bom que isso acontecesse já com a proxima arquitetura.

[Winjer]

A AMD com os Phenom II até conseguiu igualar os Core2. O problema é que passado pouco tempo a Intel lançou os Nehalem e a AMD voltou a ficar em desvantagem e teve de baixar preços.

[Jorge-Vieira]

Intel com 35 modelos Xeon

Intel Corp. is gradually increasing the number of custom microprocessors designed for specific customers in its lineup. The move helps the company to maintain its market share and address clients that have special requirements and which otherwise would have chosen proprietary solutions from other companies.
“The data center team is customizing our Xeon products for specific customers and workloads,” said said Brian Krzanich, chief executive officer of Intel, during the company’s quarterly conference call with investors and financial analysts. “Custom SKUs now number roughly 35 and over the last year volume from custom SKUs has grown at three times the rate of our off-the-shelf product.
Intel sells its custom Xeon microprocessors with special capabilities to companies like Amazon, Google and Facebook, which have unusual requirements in many cases. To run hyper-scale datacenters efficiently such companies need processors with reduced power consumption and sometimes special input/output capabilities. At present the aforementioned companies design and build their own servers. Some analysts believe that eventually they could also create their own server-class system-on-chips based on ARM architecture.

Earlier this year Intel said that it could develop custom Xeon chips with third-party intellectual property. For example, Intel could install FPGA [field-programmable gate array] chips into Xeon packages to provide certain functionality. Unfortunately, it is unclear how many of such models the chip giant has designed so far.
Advanced Micro Devices, Intel’s arch-rival on the market of central processing units, also designs semi-custom chips for various consumer electronics and other types of devices. So far the company has not announced semi-custom Opteron products, but it is known that it does certain levels of customization for select customers.

http://www.kitguru.net/components/cp...in-the-lineup/

[Jorge-Vieira]

A AMD com os Phenom II até conseguiu igualar os Core2. O problema é que passado pouco tempo a Intel lançou os Nehalem e a AMD voltou a ficar em desvantagem e teve de baixar preços.

Desde que a Intel passou para a tecnologia "Core" a AMD andou sempre atrás do prejuizo.
Depois os famosos bugs do TLB também não ajudaram a uma nova arquitura... e daí para cá é o que sabemos, a AMD combate a Intel nos preços para se tornar minimamente compettitiva e nesse aspecto têm produtos bons ou muito bons para o preço pedido.

[Winjer]

O bug do TLB apenas afectou os Phenom I. Mas realmente foi uma percalço enorme da AMD. Chegar tarde, com menos desempenho e ainda por cima com um bug destes.

[Jorge-Vieira]

E esse famoso bug nunca foi corrigido, lançaram um patch nas Bios... e a coisa ficou mais ou menos controlada, mas à custa desse bug ainda atrasou mais os produtos da AMD.

[Jorge-Vieira]

Ainda sobre o record de facturação da Intel, traduzido algum desse valor em vendas de todos os tipos de cpus traduz este numero.

Despite the economic recession, Intel seems to have scored big over the last quarter, shipping over 100 million processors within three months – a new company record. This includes chips for desktops, servers, laptops, tablets and mobile phones.

Excerto retirado desta noticia
http://www.tomshardware.com/news/intel-finance-record-revenue-3q13,27889.html

[Jorge-Vieira]

Primeiras amostras dos Intel Skylake

ntel Corp. has quietly started to sample its next-generation code-named “Skylake” microprocessors. Availability of samples of upcoming chips means that the company should be on-track with its second-generation offerings made using 14nm process technology and products on their base will arrive on time.
According to Zauba.com database, which monitors imports and exports to and from India, where high-tech companies employ a lot of software developers, Intel has been shipping Skylake processors and supporting hardware to India from different countries starting from late June or early July.
It is not completely clear whether Intel began to ship samples of its Skylake chips to its own software developers in India this summer, or provided the chips to third parties to make sure that programs will support its latest microprocessors when they become available in around mid-2015.

Based on what is listed on Zauba.com, the Skylake chips that Intel has shipped so far belong to the Skylake-Y family of microprocessors with reduced power consumption; the CPUs feature two x86 cores, entry-level graphics engine as well as LPDDR3 memory controllers. Such chips are used for hybrid 2-in-1s, tablets and other ultra-thin devices.
Central processing units based on Intel’s brand-new Skylake micro-architecture are projected to feature considerably improved graphics core, a bunch of special-purpose accelerators (e.g., HEVC and VP9 hardware decoders) and support such technologies as AVX 3.2 (512-bit instructions), SHA extensions (SHA-1 and SHA-256, secure hash algorithms), MPX (memory protection extensions), ADX (multi-precision add-carry instruction extensions) and other innovations.
Intel did not comment on the news-story.

http://www.kitguru.net/laptops/anton-shilov/intel-begins-to-sample-skylake-microprocessors/

[Jorge-Vieira]

Intel vai liderar processo de 10nm

Intel Corp. said on Tuesday that it would be the first company to start volume production of chips using 10nm fabrication technology. The company has already demonstrated the first wafers processed using 10nm tech, now it revealed that it will gradually increase its investments in the forthcoming technology in the coming quarters.
“We have led the world to 14nm and plan to do the same with 10nm process technology,” said Stacy Smith, chief financial officer of Intel, during the company’s quarterly conference call with investors and financial analysts.
Back in September the company showcased its first wafers processed using 10nm fabrication technology, but it did not reveal any details about the process or its roadmap. This time the company said that it would gradually increase spendings associated with the 10nm ramp up starting the fourth quarter and over the next two years, which indicates that the first commercial 10nm chips could emerge on the market in late 2016.
“In the fourth quarter, you are seeing the front edge of the startup costs associated with the 10nm and that is kind of right in-line with the historical timing of what you would expect,” said Mr. Smith. “We [will] talk about how that [costs] might look over the next couple of years but we are seeing at least the front edge at the time I think you would expect.”

It is noteworthy that Intel did not reveal anything about possible setbacks for its 10nm because it is getting harder and harder to adopt new process technologies. Earlier the chip giant had to delay its code-named Broadwell microprocessors by about a year because of issues with 14nm manufacturing technology and yields.
The only thing currently known about Intel’s 10nm production tech is that it will continue to use FinFET transistors and will not rely on EUV [extreme ultraviolet] lithography.

http://www.kitguru.net/components/cpu/anton-shilov/intel-we-will-lead-the-world-to-10nm-microprocessors/

[Filipe]

A intel não pára. É impressionante a qualidade desta empresa.

[Winjer]

A luta nos processos de fabrico está mesmo acesa. Há anos que não víamos algo parecido.
Mesmo que a Intel seja a primeira a chegar aos 10nm, não me parece que vá ser a única durante muito tempo.

[Patuga]

Se há alguem a fazer isso será sem duvida a intel. não acredito que ninguem tenha capacidade tecnica para os ultrapassar nisso.

[Winjer]

A TSMC anda muito perto da Intel.
A GF está atrasada, mas vai receber ajuda da Samsung.
A Samsung está atrás da Intel, mas não por muito.

[Jorge-Vieira]

Mais detalhes sobre os Intel Skylake e chips graficos Iris integrados.

A few weeks ago, specifications of Intel’s Broadwell-U lineup were leaked which detailed the HD Graphics 6100 series chipset codenamed Iris. Being the flagship iGPU chip of Intel’s Broadwell lineup, the GT3e chip will be replaced by the more powerful GT4e when Skylake hits the market later next year.
Intel’s Next Gen Skylake CPUs To Integrate GT4e Graphics Chip With 72 EUs

The information comes from Zauba (via Computerbase) who had listed several of Intel Skylake series chips in a shipping manifesto along with some key details of the processors. Previously, Zauba has given details of a few NVIDIA graphics cards related to GK110 and GM204 which turned out to be accurate so let’s get on with the details. As we previously detailed, Broadwell will feature four iGPU configurations which include GT1, GT2, GT3 and GT3e.
The GT3e “Iris Pro” level graphics chipsets which include the HD 6100 “Iris” GT3e with 48 Execution Units and 128 MB eDRAM cache, The GT3 HD 6000 comes with 48 Execution units and no L4 cache, HD 5300/5500 GT2 graphics chip with 24 Execution units and GT1 HD Graphics chip with just 12 Execution units. Broadwell has increased the EU count by 50% and there are seven H/W threads per EU with 128 GRF registers per thread along with simultaneous multi-threading support.
While Broadwell has already launched with the Core M “Broadwell-Y” series processors, the real chips will start shipping at CES 2015 where Broadwell-U series will enter the market. Broadwell K-Series is being planned for a Q2 2015 launch but at the same time, Intel has announced Skylake-S series which will be available to consumer market featuring lower TDPs. The Skylake-S series chips will be part of a diverse desktop and notebook based platform but being based on a 14 nm process node, Skylake processors would feature a new architecture along with several CPU and GPU side enhancements whereas Broadwell is meant to improve upon the architecture which was first introduced on Haswell processors. The following chart shows the main differences in the next generation Skylake processor configurations:

WCCFTech.com	Intel 14nm Skylake Configuration Table
Variant	SKL-Y (BGA)	SKL-U (BGA)	SKL-H (BGA)	SKL-S /K (LGA)
Cores Configurations	2	2 / 2	4 / 4	4 / 2 / 4
Graphic Configurations	GT1 12 EUs GT 1.5 18 EUs GT2 24 EUs	GT2 24 EUs / GT3e 48 EUs	GT2 24 EUs / GT4e 72 EUs	GT2 24 EUs / GT2 24 EUs/ GT4e 72 EUs
eDRAM	-	64MB – GT3e	128MB – GT4e	64MB – GT4e
Memory	LPDDR3 1600Mhz	LPDDR3 1600Mhz	DDR4 2133Mhz	DDR3L/DDR3L-RS 1600Mhz, DDR4 2133Mhz
TDP	4W	15-28W	35-45W	35-95W

Compared to Haswell and Broadwell, we are looking at some decent extension in the number of Execution Units till the GT3 iGPU but the GT4e is way beyond the GT3e chips with 72 Execution Units. The chip will be featured across the Syklake S/H/K series processors with 64 MB eDRAM cache on LGA variants and 128 MB eDRAM cache on BGA variants.
Aside from all these details, the SKUs for Skylake have also been confirmed along with their TDPs. The desktop variants will have six different variants which will include the flagship 95W Quad Cores with GT2 and GT4e GPU while the lowest 35W TDP variant will feature Dual Cores and GT2 graphics. The H-Series lineup will also include the 95W Quad Core with GT4e graphics chip but that’s the only flagship chip followed by 65W, 55W and 45W Quad Core variants with a mix of GT4e and GT2 graphics. This H-Series doesn’t seem to include a dual core model at all. The U-Series lineup will have TDPs of 25W and 15W with the flagship part featuring Dual Cores and GT3e graphics while the other models consist of GT2 graphics. Last up is the Y-Series Skylake lineup featuring two SKUs, both featuring 4.5W TDP and 2 Cores plus GT2 graphics.
Broadwell will stick to the LGA 1150 socket and remain compatible with current 9-Series chipset motherboards while Skylake will shift focus to the next mainstream chipset known as the Z170. The chipset will feature up to 20 PCIe Gen 3 lanes, 6 SATA Gen 3 ports, 10 USB 3.0 ports and a total of 14 total USB ports (USB 3.0 / USB 2.0), up to 3 SATA Express capable ports, up to 3 Intel RST capable PCI-e storage ports which may include x2 SATA Express or M.2 SSD port with Enhanced SPI and x4/x8/x16 capable Gen 3 PCI-Express support from the processor. Aside from that, we know that the Skylake processors would be compatible with the latest LGA 1151 socketed boards and Z170 chipset which is part of the new 100-Series chipset replacing the 9-Series “Wild Cat” PCH. The second most interesting thing covered is that Skylake processors would have both DDR3 and DDR4 memory controllers so the different SKUs can be configured to use either of the memory type.
The difference between them is not clearly mentioned however we think its related to clock speeds. You can get a better representation of the variants in the following chart:

SKUs	x86 Cores	Graphics	TDP
Skylake DT	4	GT4e	95W
Skylake DT	4	GT2	95W
Skylake DT	4	GT2	65W
Skylake DT	2	GT2	65W
Skylake DT	2	GT2	35W
Skylake DT	4	GT2	35W
H-Series	4	GT43	95W
H-Series	4	GT4e	65W
H-Series	4	GT4e	55W
H-Series	4	GT4e	45W
H-Series	4	GT2	65W
H-Series	4	GT2	55W
H-Series	4	GT2	45W
U-Series	2	GT3e	25W
U-Series	2	GT3e	15W
U-Series	2	GT2	25W
U-Series	2	GT2	15W
Y-Series A	2	GT2	4.5W
Y-Series B	2	GT2	4.5W

Intel Mainstream Platforms Comparison Chart:

	Intel Sandy Bridge Platform	Intel Ivy Bridge Platform	Intel Haswell Platform	Intel Broadwell Platform	Intel Skylake Platform	Intel Cannonlake Platform
Processor Architecture	Sandy Bridge	Ivy Bridge	Haswell	Broadwell	Skylake	Cannonlake
Processor Process	32nm	22nm	22nm	14nm	14nm	10nm
Processors Cores (Max)	4	4	4	4	4	TBA
Platform Chipset	6-Series “Cougar Point”	7-Series “Panther Point”	8-Series “Lynx Point”	9-Series “Wild Cat Point”	100-Series “Sunrise Point”	200-Series “Union Point”
Platform Socket	LGA 1155	LGA 1155	LGA 1150	LGA 1150	LGA 1151	TBA
Memory Support	DDR3	DDR3	DDR3	DDR3	DDR4 / DDR3	DDR4
Thunderbolt	Yes	Yes	Yes	Yes	Yes “Alpine Ridge”	Yes
Platform	Desktop LGA	Desktop LGA	Desktop LGA	Desktop LGA	Desktop LGA	Desktop LGA
Launch	2011	2012	2013-2014	2015	2015	2016

Read more: http://wccftech.com/intel-skylake-processors-integrate-iris-gt4e-level-graphics-chip-features-72-execution-units-128-mb-edram-llc/

[Jorge-Vieira]

Intel Skylake fotografado, mas aparentemente as frequencias não são o que era esperado.

There hasn’t been alot of news on Intel’s desktop front after Skylake was officially revealed. However, now the good folks over at VR-Zone.com have posted pictures of the ES-1 Engineering Sample of a Skylake Processor along with some very very important information. Get ready for a shocker, because this one ain’t pretty.
Intel’s Skylake-S ES-1 processors pictured- disappointing stock clocks leaked

As you probably know by now, Skylake is the new architecture succeeding the die shrink that is Broadwell (which still hasn’t arrived yet). It will come only in two models, namely an i5 and an i7 variant. Both SKUs will have their multiplier locked, which means you can’t overclock these using the conventional methods. Now the source mentions that you can expect the processors sometime in Q2 2015, and recall that we have already told you that the month(s) to look forward to, are June and July. Interestingly Broadwell-K with the unlocked multiplier has not arrived yet, implying that it may arrive alongside or just before the Skylake processors. The socket is LGA 1151 and the mainstream chipset used will be Z170/H170. These will be coupled with the B150, Q150 and Q170 chipsets.
However, the thing that literally (well, not really) shocked the socks off me, was the utterly disappointing stock clocks. The processors will feature a 2.3Ghz Core Clock and a 2.9Ghz Turbo Clock. Another configuration will have 2.2 Ghz Core Clock with an absolutely depressing 2.4Ghz of Turbo. TDP will be 95W and 80W respectively. Heed my words folks, If this leak is true, and I really don’t know what to think with this one, then Intel is engaging in a paradigm shift. I previously expressed my concerns in my analysis about Intel’s delays and this just adds to the pie. Intel is forcing any and all enthusiasts one generation behind the current architecture, and the HEDT enthusiasts one more generation behind. I assume the latest generation will be reserved for OEMs. Once again, if this is true then Intel is well on its way to becoming a monopolistic tyrant. Intel will release the Greenlow platform sometime in 2015 so we should definitely hear more about this.

Read more: http://wccftech.com/intel-14nm-skyla...#ixzz3Gif3F24w