What is the best for 2019 and beyond?

With Computex and E3 2019 behind us, the time has come to examine the tug of war between AMD and Intel. Both have introduced new products, and AMD is expected to launch new processors and GPUs in July. Intel will not launch its new 10th generation processors until the end of 2019. Intel's new line of discrete graphics processors will not appear until 2020.

The state of AMD

AMD entered the x86 processor market as an Intel subcontractor. This contract allowed AMD to use Intel's 8086 design for manufacturing processor clones. These chips would help Intel to run the orders for new IBM PCs.

After Intel launched its 32-bit processor, AMD's contract stalled, forcing both companies to fight for years. AMD resumed making clones until 1996, when it launched its first internal x86 processor, the AMD K5 processor.

AMD launched the first 64-bit consumer-oriented processor, the Athlon 64, in 2003. It also launched the Athlon 64 FX for hobbyists and the Opteron chip for servers. The company's first dual-core chip, the Athlon 64 X2, arrived in 2005. Its first four-core chip, Phenom, arrived two years later. The Athlon and Phenom desktop components seemed to be the main focus of AMD.

That changed when Apple launched the mobile phone boom with its first iPhone and iPad.

The bulldozer years

After a brief presence between 2006 and 2008, AMD restarted its mobile efforts with Fusion. This initiative introduced AMD's first-ever accelerated processing unit (APU), which combines processor and GPU cores into a single chip. He also launched a chain reaction that would see AMD fall behind Intel in desktop computing until 2016. AMD's APU efforts dominated the bulldozer years.

For example, between 2011 and 2016, the FX chips were the only work of AMD. They are named Zambezi and Vishera and are based on AMD's Bulldozer architecture (Piledriver was a revised bulldozer). Meanwhile, Intel created chips for desktop and laptop computers every year. The company is also focused on the business sector given the uncertainty of desktop computers.

Opponents predicted that tablets and smartphones would kill the desktop and laptop markets. But the Ultrabooks, the 2 in 1 and the removable computers have apparently saved the PC industry.

To defend AMD, opponents predicted that tablets and smartphones would kill the desktop and laptop markets. But the Ultrabooks, the 2 in 1 and the removable devices have apparently saved the PC industry and have almost killed tablets. Nevertheless, OEMs typically keep Intel chips in their PCs while using ARM solutions in their mobile devices.

Despite its strong focus on UFAs, AMD had a master plan.

Consoles and graphics

Custom APUs based on its GPU Graphics Core Next architecture have landed in Xbox One, Xbox One X, PlayStation 4 and PlayStation 4 Pro. Developers working on x86 PCs can now create games running across the console. and PC without any portage involved. High Definition PC games have finally returned.

This is the reverse of the minimal presence of AMD's processor during the Bulldozer years: it is also a manufacturer of graphics cards. AMD acquired ATI Technologies in 2006 and began producing additional graphics cards for desktop computers. Intel will not participate in the GPU race until 2020.

AMD's first-generation Graphics Core Next (GCN) architecture has been integrated with the Radeon HD 7000 "Southern Islands" expansion card family. The Radeon RX Vega Series has completed the AMD GCN era with the 7nm Radeon VII graphics card, designed by Gigabyte, Sapphire, XFX, and more.

Similar to AMD's mobile focus, the two years between Radeon RX 300 and Radeon RX Vega have targeted consumer graphics. This allowed Nvidia to dominate the desktop and laptop space with its GeForce GTX 900 and GTX 10 series. At the same time, AMD released affordable RX 400 and 500 cards, which provide low-cost RV and Full HD graphics. on all desktops.

Meet Zen and Vega

In retrospect, AMD noticed a four-year gap between its Vishera FX processor family and its Ryzen 1000 chips. Two years passed between AMD's latest Radeon RX 300 high-end graphics processor and its Radeon RX family 400 affordable.

During this period, AMD secretly worked on a new processor architecture from scratch. Called Zen, this would put society back on a competitive trajectory.

AMD has also developed the Vega graphics platform based on its fifth-generation GCN architecture. This design served as AMD's top-of-the-line successor to the RX 300 family. AMD also created Radeon DNA (RDNA): this is the first new GPU architecture from scratch of the company since the introduction of GCN in 2012.

According to AMD, Ryzen processors can match the performance of Intel processors at half the cost.

According to AMD, Ryzen processors can match the performance of Intel processors at half the cost. Although it sounds good, there is a huge setback: Ryzen desktop chips do not include built-in graphics. If you are a PC player, it will probably not matter if you want a specific graphics card. If you do not need a separate GPU, most Intel desktop and mobile processors include built-in graphics. AMD's Ryzen APUs also include embedded graphics.

The original Ryzen 1000 series is based on AMD's first Zen design using 14-nm processing technology. The Ryzen 2000 desktop series is based on an improved zen architecture (also called Zen +) and 12-nm processing technology.

On the moving front, AMD's 2000 Ryzen APUs for laptops and desktops use the original 14nm zen architecture. The new Ryzen 3000 APUs arrived in July are based on the 12 nm Zen Plus (or Zen +) design. Compared to desktop chips, AMD's Ryzen brand APUs are lagging behind Zen's Gen1-Refresh-Gen2 update model.

Ryzen 3000

The new Ryzen 3000 desktop processors are based on AMD Zen's second generation architecture (Zen 2) and TSMC's 7 nm + processing technology. This is remarkable, given that Intel's 10-nm Ice Lake chips will not ship until the end of 2019. Even more, the new AMD package includes the upcoming Ryzen 9 3950X, a 16-core chip with 4.7 GHz for $ 749. The equivalent of Intel is the Core i9-9960X which costs at least $ 1,725. Ouch.

But wait! There is more! AMD's new Ryzen 3000 series supports PCI Express 4.0, unlike current Intel products. Abbreviation for Peripheral Component Interconnect Express, PCI Express is a standard for high-speed connections between the processor, graphics card, storage, and so on. PCI-SIG approved the PCIe 4.0 specification in October 2017, allowing data transfers of up to 64 GB per second (16GT / s).

Twenty months later, PCI Express 5.0 is now ready for hardware manufacturers. Like PCIe 4.0, devices that support this standard may not appear for 20 months. It promises up to 128 GB per second using an x16 configuration (32GT / s). AMD, Intel, Nvidia and others have already committed to this new standard.

The performance formula per watt

Thanks to Ryzen, AMD targets performance per watt, which allows for more cores and frequencies at half the cost. For enthusiasts, AMD offers its Ryzen Threadripper processors like the 32-core 2990WX priced at $ 1,799. At present, the 1999 Core i9-9980XE dollars has 18 mainframes.

AMD is now a leader in the desktop market. The company does not overload its Ryzen portfolio with an insane number of products. With the Ryzen 2000 Series, AMD provides eight desktop processors, four HEDT processors, ten mobile APUs, and twelve desktop APUs. In the only 9th generation Coffee Lake refreshment family, the company sells thirty-four desktop processors and nine notebook processors. We expect the HEDT tokens to arrive later this summer.

But keep that in mind: even if the price of AMD chips is lower, they consume more energy. Look at this comparison between AMD and Intel:

	Basic speed	Maximum speed	Power	Price
Ryzen 7 3800X	3.9GHz	4.5 GHz	105 watts	$ 399
Core i9-9900K	3.6 GHz	5.0 GHz	95 watts	$ 488

With Intel Turbo Boost technology, the Intel chip can reach the 5.0 GHz ceiling using two cores. The number of amps drops to 4.8 GHz with four cores and at 4.7 GHz with eight cores. At the same time, AMD's Precision Boost 2 technology increases the speed of any number of cores. The increase is based on an analysis of the current environment involving the use of heat, electricity and available height. This is what AMD calls the "reliability triangle".

That said, the AMD chip has a basic speed advantage, while the Intel chip has a higher turbo speed cap. And while the Ryzen 7 chip costs $ 89 cheaper, it consumes an extra 10 watts. Unfortunately, we do not have reference numbers for comparison purposes as the Ryzen 3000 table components will only be delivered in July.

A new deal with Samsung

AMD re-enters the portable consoles market with a new contract with Samsung. The company had a brief presence in the non-gaming console market after its acquisition of ATI Technologies in 2006. It is now back in the game, giving up its GPU technology to Samsung.

Prior to the acquisition, ATI provided two SoCs (all-in-one processors, or system-on-chips). Xilleon has accelerated video decompression for broadcast networks. Imageon has integrated graphics embedded in portable mobile devices that support 2D and 3D rendering.

After the acquisition, AMD renamed the chips in AMD Imageon and AMD Xilleon. Two years later, AMD decided to focus primarily on x86 processors and graphics chips. This involved ceasing its manufacturing operations as GlobalFoundries and selling its ATI-related SoC divisions in 2008. Qualcomm purchased Imageon technology and renamed Adurino, while Broadcom purchased Xilleon technology.

This new agreement with Samsung integrates AMD's Radeon core graphics technology with Samsung Exynos chips used in smartphones and tablets. Samsung generally uses its Exynos chips in devices sold in the international market, while relying on Qualcomm Snapdragon chips in North America.

A new deal with Google

Samsung is not the only company looking for AMD's GPU technology. The company announced in March that Google's next game streaming service, Stadia, would use Radeon branded custom GPUs designed for data centers. Based on the AMD multi-user GPU technology introduced in 2015, these GPUs include 56 computing units (3,584 stream processors) and dedicated HBM2 memory to produce 10.7 teraflops of graphics processing power.

The great idea received at the 2019 GDC was that AMD would provide a custom APU, as is the case with the consoles. This is not the case. AMD makes it clear that Google will use its GPUs designed for data centers. There is no mention of APU or processor cores manufactured by AMD. Presumably, these will be custom built by Intel and clocked at 2.7 GHz.

What might be the case is that Google's data centers already have Intel-based processors. The company has probably signed a contract with AMD for the installation of GPU Radeon Datacenter (if they are not already). The purchase of AMD Opteron APU-based systems may not be ideal because of the power needed to run and broadcast multiple virtual machines. In addition, AMD server APUs target small businesses looking for high performance, low energy costs.

However, Google Stadia is a big win for both AMD and Intel. In addition, since console games already target custom APUs, there is no compromise in functionality as the games run on AMD's GCN architecture. The only big loser in this scenario is Nvidia.

The state of Intel

Intel does not really need backstory. The doors opened under the name of N M Electronics in 1968 and then changed to Intel – abbreviation of Integrated Electronics – a month later. The era of x86 processors began with the Intel 8086 chip used in the new IBM family of personal computers launched in 1981. Intel's 80286, 80386 and 80486 microprocessors followed suit.

Intel started using a tick-tock production model in 2007. The "tock" represented a change in the microarchitecture of the processor, while the "tick" piled up the revision in a smaller chip configuration. For example, Intel used its fourth-generation 22nm "Haswell" microarchitecture in processors launched in 2013. "Fifth-generation" fifth-generation Intel processors arrived the following year on the basis of Intel's latest development. a 14 nm version of "Haswell".

Tick ​​Tock's death

The move to 14-nm process technology effectively killed Intel's Tick-Tock model and introduced a new model called Intel. Process Architecture-Optimization. With its 14nm process node already operational, Intel has designed a new microarchitecture code name Skylake. This design served as a foundation for its fifth to ninth generation processor families. Intel has officially ended its tic-tac model with the launch of its seventh generation of "Kaby Lake" processors.

Kaby Lake builds on the first optimization 14nm Intel process technology (dubbed 14nm +) in 2016. Intel has refreshed Kaby Lake for 2017 in the first wave of eighth-generation mobile processors using the same process node. This improved design has increased energy efficiency and added four cores to the Intel Core i5 family of processors. The eighth generation really only started when the second Intel Skylake * 14nm ++ optimization, nicknamed "Coffee Lake".

From there, we witnessed a third optimization in 2018 (14nm +++) with "Whiskey Lake", a success only mobile for Kaby Lake Refresh. We also attended the launch of "Amber Lake", Kaby Lake's successor for mobile phones.

Meanwhile, Intel has launched a new processor based on 10nm processing technology called Cannon Lake. Still based on Skylake, the eighth generation chip has appeared but has not been generalized. Which Cannon Lake made For that, it was necessary to restart the optimization engine of the Intel process architecture.

Caffeine and ice water

This brings us to the latest Intel processors. Initially launched in October 2018, the ninth generation family is updating Coffee Lake on the 14 nm ++ process node. Three desktops arrived in October, followed by six in January at CES 2019 and another 24 in April. This deployment number does not even include mobile, server, and HEDT products.

Not stopping there, Intel introduced its tenth generation "Ice Lake" family at Computex 2019, based on the new "Sunny Cove" architecture. This is the architecture part of Intel's process architecture optimization model. The first eleven chips target the suffixes "U" (ultra-low consumption) and "Y" (very low consumption). You'll see up to four cores and eight threads, speeds up to 4.1 GHz and processor speeds up to 1.1 GHz.

Unfortunately, we do not know anything about these chips, apart from the little things provided by Intel. They feature a revised integrated GPU architecture (Gen11) promising smooth frame rates in Battlefield V in 1080p. They also support 3200 MHz DDR4 memory. Intel 300 Series chipsets add Wi-Fi connectivity 6 and support for Intel Optane memory.

Ice Lake processors and chipsets are supposedly now being shipped to OEMs for laptops arriving during the 2019 holiday season.

AMD vs. Intel clash

That said, the battle between AMD and Intel will pit the third-generation Ryzen "Zen 2" chips against Intel's "Coffee Lake" ninth generation products. But as we said earlier, Ryzen 3000 is only shipped in July. We therefore have no benchmarks for comparison.

What we can is comparing a second generation AMD Ryzen chip with a similar ninth generation Intel processor. We dug into Geekbench to find their partitions to one and more kernels:

	Ryzen 7 2700X	Core i9-9900K
nuclei:	8	8
Son:	16	16
Base Speed ​​(GHz):	3.7	3.6
Max speed (GHz)	4.3	5.0
Power:	105 watts	95 watts
Single-hearted score:	~ 4994	~ 6519
Multi-core score:	~ 29453	~ 37440
Price:	$ 279	$ 484

As the results show, even though the Ryzen 7 2700X has a slightly higher base speed at a lower cost, it still does not outperform the Intel Core i9-9900K processor. This is a big argument in the debate between AMD and Intel: the heart of the Intel processor is simply better to execute instructions per cycle. In addition, the AMD chip consumes more power and does not even come with built-in graphics. In the end, it may be better to spend an additional $ 205 on the Intel chip.

Let's do another comparison for laptops:

	Ryzen 7 2700U	Core i7-8559U
nuclei:	4	4
Son:	8	8
Base Speed ​​(GHz):	3.3	2.7
Maximum speed (GHz):	3.8	4.5
Power:	25 watts	28 watts
Single-hearted score:	~ 4011	~ 5689
Multi-core score:	~ 10929	~ 19133

Here we see that AMD's second-generation APUs consume three watts less energy. Despite its higher base speed, the chip is lagging behind Intel's eighth-generation single-core Geekbench processor. It is also late in multicore testing partly because of its lower maximum speed.

according to Intel Business Unit"All hearts are not created equal, and more hearts do not always mean better overall performance."

According to Intel, performance also depends on the optimization of memory and architecture. The company clearly explained this after comparing its new Epyc "Rome" processor to 64 second-generation cores to the Intel® 28-core Intel Server Core Xeon Platinum 8280 "Cascade Lake" upgradeable processor. AMD has demonstrated that its chip is running twice as fast as the Xeon in a benchmark. Intel said that AMD had not correctly configured the test system, producing substandard results with the Xeon chip.

Navi vs Xe in 2020

Another problem faced by AMD is Intel's upcoming entry into the complementary graphics card market. The former chief architect of AMD Radeon, Raja Koduri, joined Intel in late 2017 to assume the duties of chief architect and senior vice president of a new group of directors. basic computer and visual. His first task was to create a discrete graphics card by 2020. The agreement also included the integration of Radeon cores into Intel-based modules containing Kaby Lake processor cores and HBM2 video memory.

Intel's new discrete GPUs will be based on its scalable Xe architecture. You'll see solutions for the data center, desktop enthusiasts and laptops. You'll see parallel computing with real-time hardware-based ray tracing, competing with the Nvidia GPU RTX 20 Series family. The Nvidia GTX 10 series only supports ray tracing via acceleration or GPGPU software.

This is great news, especially for a processor that is coming back to the GPU market. The tracing of rays on a consumer workstation is a big step forward, promising a photo-realistic rendering with no horrible waiting times. This is the novelty in the field of games developed by the RTX 20 family of Nvidia for desktops and laptops.

AMD CEO Lisa Su spoke about hardware and software based tracing in January, but made no mention of ray tracing at her E3 2019 conference in June. Instead, she revealed the new "Navi" cards arriving on July 7, 2019:

	Radeon RX 5700 XT Edition 50th Anniversary	Radeon RX 5700 XT	Radeon RX 5700
Units of calculation:	40	40	36
Basic speed:	1,680 MHz	1.605 MHz	1465 MHz
Speed ​​of play:	1 830 MHz	1.755 MHz	1.625 MHz
Maximum speed:	1,980 MHz	1 905 MHz	1,725 ​​MHz
Memory (GDDR6)	8GB	8GB	8GB
Performance:	10.14 TFLOPS	9.75 TFLOPS	7.95 TFLOPS
Competitive product:	GeForce RTX 2070	GeForce RTX 2070	GeForce RTX 2060
Price:	$ 499	$ 449	$ 379

With Intel entering the discrete GPU space, AMD and Nvidia will not be the only competitors to fight for your dollars. For Intel, it can be difficult to penetrate this market given the huge customer base dedicated to AMD and Nvidia. Hardware ray tracing seems to be an ace in the hole and a great alternative to the Nvidia RTX 20 series. Unfortunately, AMD's new Radeon RX 5700 series does not include hardware support.

So who wins this war?

AMD vs Intel in the desktop

In the battle between AMD and Intel for the workstation, Intel is expected to continue to dominate in the near future. Nevertheless, AMD represents a significant threat.

The high number of processors and the low prices of AMD are an interesting selling point. In turn, they require more energy and do not include embedded graphics. Customers can get an AMD 12-core processor priced at $ 499, but there is currently no 12-core chip marketed by Intel. The second half of 2019 should see the release of AMD's third-generation Ryzen 3 chips. W can even see new Threadripper HEDT parts.

Meanwhile, Intel may introduce new XED Series HEDT processors to compete with newer Threadrippers. Since the Ice Lake processors will not arrive until the end of 2019, Computex is perhaps Intel's latest on this topic in consumer processors. Until then, the Athena project should generate a lot of buzz before the start of Ice Lake: the successor of Ultrabook based on Intel 10nm processors.

AMD vs Intel in laptops

In the quarrel between AMD and Intel laptops, Intel is expected to continue to dominate because of the trend.

AMD's APUs typically resided in low-cost notebooks until the arrival of Ryzen in 2017. Intel still exceeds AMD in laptops, but you can find some excellent solutions such as Acer Predator Helios 500 and Aspire 3. APU brand too.

Unfortunately, AMD still does not offer an eight-core smart chip, despite all its core heaping in the office space. Instead, Intel is currently taking the lead with its i9-9980HK and i9-9880H eight-core notebook processors. You will probably find them in game laptops associated with a separate Nvidia GeForce graphics chip. Heck, Intel also sells six-core notebook processors.

AMD vs Intel – Beyond the PC

Despite its fight against Intel in three main markets, AMD will continue to dominate in the field of consoles. AMD's close connection to the console and computer partners gives it a central place in game play and development. With the PlayStation 5 and the Scarlet project using AMD components, this domination probably will not change before five years. Nvidia, meanwhile, has Nintendo.