When Intel removed the Ice Lake cover, we found that the CPU performance data was complex. When it comes to the graphics processor, Ice Lake is a huge step forward, with performance far better than anything we've seen before with Intel integrated graphics. The processor, however, was a rather mixed bag. Limited to a TDP of 15 W, Ice Lake processors were not necessarily faster than the Coffee Lake chips that they had to replace and were often a little slower. If you give the processor extra room, the problem is solved – but of course, giving the chip more play power has a negative impact on the heat and battery life.
When Intel invited reviewers to ice lake test, the proposed test systems included a rocker switch to switch envelopes from 15W to 25W. This is how PCMag and other publications have been able to test the laptop in both modes, as shown below:
Users do not usually have this type of option. The TDP ranges are usually predefined by the OEM and can not be changed by the end user for obvious reasons – starting the TDP for laptop is a good way to overheat the system if you do not know not what you do and if the laptop is not specifically designed to operate at a higher power level. To the best of our knowledge (up to today), no consumer notebook could change its TDP values on the fly. In the Ice Lake testing session, Intel told critics that retail Ice Lake laptops would not have this option either.
However, there seems to be at least one exception to this rule. The Razer Blade 13 will have a Adjustable TDP can be configured via the Razer Synapse software. Supposedly, this ability has always existed, dating back to the original Razer blade. If this is true, the company does not seem to have put it forward before. Google does not provide results that reference an adjustable TDP on previous versions of the Razer blade. unless you count the fact that the laptop would slow down the load under certain circumstances. To be clear, the ability to run the processor in an energy-efficient envelope under load is not the same as being able to voluntarily put it in a higher TDP mode and make it work with a additional power reserve.
Given that Intel had already told critics not to expect adjustable TDP ranges to be a major feature for laptops, this begs the question: is it specific to Razer or will we see do more laptop manufacturers take advantage of these new features? Will Intel make adjustable TDPs a feature that high-end customers will be able to buy as they wish?
Razer's website for the new blade indicates that the system will use a 25W Ice Lake processor, but does not mention anything about the fact that the system can be adjusted in a power envelope of 15W instead of 25W.
We've heard for a while that Intel could react to the 7 nm AMD attack with a higher processor count on desktop processors. A new leak suggests that this is exactly what the company will do, with a new chipset supporting up to 10 processors based on its mature 14 nm process. This will supposedly require a new processor connector because Intel is increasing the power supply and capacity of its desktop motherboards to offset the higher power requirements of a 10-core chip.
The new take is supposed LGA 1200 and the high-end chips offer 10C / 20T configurations if we believe the rumors. The TDP is finally rising, up to 125W. The latter is something of an interesting point. The power consumption of the Intel processor currently has little relationship with TDP if you let the CPU grow. The TDP is measured at the base clock and not at the amplification. Intel may need to extend TDP to add more processor cores, but in the past, Intel still has processors in the same TDP media by cutting the base clock.
Our hypothesis is that Intel is raising TDP because it does not want to do it anymore. It is probably possible to further reduce the base clock to stay in the old 95 W TDP support with 10 cores instead of eight, but risk creating negative comparisons with previous generation components or AMD hardware. Intel reduced the base clock speed when it went from Core i7-8700K to Core i9-9900K – the 9900K has a base clock of 3.6 GHz, while the 8700K is 3 , 7 GHz. The old 7700K had a base clock of 4.2 GHz, although the overall performance is significantly lower.
The relatively low base clock may not have been a cause for concern when AMD's Ryzen 7 base clocks were also in the 3.6-3.7 GHz range, but AMD slightly adjusted its own clock ranges on 7 nm. The 3700X has a base clock of 3.7 GHz, while the Ryzen 3800X has a base of 3.9 GHz and the 3900X is a chip at 3.8 GHz. Intel may want to bring clocks up slightly to make sure that it fits on the base, and the only way to do it is to push the TDP higher.
Supposedly, the new 400 series adds 49 extra pins to the LGA1200, with the extra pins used for power supply. There would be some new features, such as integrated 802.11ax support and probably an easier method of integrating Thunderbolt 3, similar to what we saw in mobile. The 65W and 35W processors would still be supported (and released) during this latest 14nm revision. This is simply enthusiastic TDP support that could stretch up to 125W. Intel will probably try to keep the tempo boost as high as possible, but I do not want to speculate on what it will be.
Anyone who has focused his attention on the relative rankings between AMD and Intel has already realized that a 10-core Lake Comet will not live up to AMD in most performance areas. The 16-core Ryzen 9 3950X is on its way and we've already seen what happens when a 10-core Intel HEDT processor adopts a 16-core AMD Threadripper: The 10-core processor loses. Above all, he loses a lot.
But if that sounds absurd, defeating AMD's goal in terms of absolute multicore performance is probably not the goal. The two companies work together on their respective strengths: for AMD, this means favoring the multi-core while seeking to improve the mono-core, where Intel still has a limited advantage in some games in 1080p. For Intel, it means trying to improve a single core while competing more effectively with multi-core. Composing up to 10 cores and raising the basic clock via the increase in TDP probably help the company achieve this goal. It will take more than +2 hearts to put Intel back into the multi-threading game seriously, and the company knows it.
The rumors of a 10-core Comet Lake are strong enough and last long enough for me to think they are strong enough. We believe that this generation will also witness the return of hyper-threading to strengthen Intel's competitiveness against AMD at lower price ranges. Without price information, we obviously can not think of how the two companies will come together, but Intel has always introduced better price / performance ratios when launching major products. This suggests that we will see the company adjust its basic account / dollar strategy at the next major launch.
Intel has announced a new slice of 10th generation mobile chips, this time based on 14 nm. This is the third recent announcement of the Intel 10th generation company and the first to show us how 10nm and 14nm products will live side by side in the same product families. The main news is that Intel has reduced its maximum number of mobile CPU cores to 6C / 12T in a 15W power supply envelope, instead of 4C / 8T. 14nm processors<a href="https://r.zdbb.net/u/6k9u" target="_blank" rel="noopener noreferrer"><img src="https://www.extremetech.com/wp-content/uploads/2018/07/SEEAMAZON_ET_135.jpg" alt="SEEAMAZON_ET_135 View Amazon AND Trade" width="“135”" height="20"/></a> in the 10th generation, Comet Lake is associated with Ice Lake to complete the field.
On paper, this change should be an excellent choice for Intel. When the company launched 8th generation chips, it significantly improved its performance. Our initial concerns that high-speed dual hearts may be better options than lower-rate quads were unfounded; The low clocks of the 8th generation mobile components did not prevent them from generating excellent results in comparison.
There is good reason to think that this is no longer the case. Here is one of the official Intel slides predicting performance improvements as customers who purchase a new 10th generation processor such as the Core i7-10710U (the six-core variant) can be expected at:
These are important gains for a single product generation. Overall performance up to 16% higher than Coffee Lake, 41% better productivity in Office 365 and the same battery life? Not bad But check the details.
This is from Intel's official disclaimer page. Each numbered entry – 1, 2, 3 – deals with one of the claims we have just shown you. I've highlighted the TDP listed for each CPU in each entry. Note that # 1 and # 2 – both performance claims – involve two very different system configurations. In both cases, the six-core Core i7-10710U was configured to work with a 25 W TDP, while the Core i7-8565U was handicapped by a 15W TDP.
The third data point, however, does not not show this configuration. Here, both chips operate in a 15W envelope. The problem here is that users usually do not have access to a method of switching between operating modes provided by OEM or Intel. It's a decision that the laptop manufacturer does. You can sometimes use third-party utilities or the Intel Extreme Tuning utility to tweak processor configurations, but you can not just switch between 15W and 25W configurations. Regardless of the configuration used by the manufacturer of your notebook, it keeps you away, and this information is usually not published.
We compared backward the launch of the 8th generation in 2017 to see how Intel had handled the messaging in this situation. The 8th generation family has experienced a similar slide compared to the rear generation family of the 7th generation.
We see a similar improvement (although much larger) and a similar footnote. Where does this lead us?
Nowhere good. In 2017, when Intel compared performance between the Core i7-8550U and the Core i7-7500U, it was not necessary to use TDP values to align its performance. The comparison was made with 15W allocated for both processors.
That's the only reason Intel can do that: energy consumption. TDP nominal values are not equivalent to the total CPU consumption and should be not to be read this way, giving a processor more TDP margin allows it to consume more power. When critics spent time with Ice Lake earlier this month, we noted in particular how to give a processor more TDP margin allows it to run faster, as shown below:
<img class=”aligncenter size-large wp-image-297005″ src=”https://www.extremetech.com/wp-content/uploads/2019/08/657642-intel-ice-lake-cpu-tests-pov-ray-1-640×438.png” alt=”657642-intel-ice-lake-cpu-pov-ray-tests "width =" 640 "height =" 438 "srcset =" https://www.extremetech.com/wp-content/uploads/2019/08/657642 -intel-ice-lake-cpu-tests-pov-ray-1-640×438.png 640w, https://www.extremetech.com/wp-content/uploads/2019/08/657642-intel-ice-lake- 300w cpu-tests-pov-ray-1-300×205.png, https://www.extremetech.com/wp-content/uploads/2019/08/657642-intel-ice-lake-cpu-tests-pov-ray -1-768×525.png 768w, https://www.extremetech.com/wp-content/uploads/2019/08/657642-intel-ice-lake-cpu-tests-pov-ray-1.png 911w "sizes = "(maximum width: 640px) 100vw, 640px" />
We do not know how much faster the Core i7-10710U is when you use a 25 W TDP than a 15W TDP. What matters is that Intel gives a false picture of the type of comparison it makes on its 10th generation launch slides. Compare the performance of laptops in two different TDP ranges for your performance indicators, only to reverse and compare what basically constitutes a different Machine configuration for battery life is hypocritical. Switching between the 15W and 25W operating modes may not seem like a problem, but it's not a switch that an end user can launch. When you buy one of these chips, you get either the 25W higher performance version or the lower 15W version, and the builders do not usually communicate the very fine points of their power management strategies. or their SKU selections.
The last reason to suspect that TDP limits CPU performance in this case? The winnings are not great enough Switching from quad to six cores may not be as big an upgrade as going from 2C / 4T to 4C / 8T, but it should still be a basic improvement of 1.5 times, and There are many landmarks that will show it. type of gain – if the chip does not already reach the thermal limits.
Intel is launching a full suite of U and Y class components, as shown below:
Apart from the Core i7-10710U, the improvements are hard to come by. The Core i7-105100U is a 4.8 GHz mono-core boost based on 1.8 GHz, single-core and 4.3 GHz. Intel has not revealed its CPU-focused boost frequencies like the Core i7-8665U, but this processor is a 1.9 GHz / 4.8 GHz processor. The total number of US for graphics and graphic frequency are the same between the two parts. The Core i7-10710U supports LPDDR4X-2933, LPDDR3-2133 or DDR4-2666, while the Core i7-8665U only supports DDR4-2400 or LPDDR3-2133, but these enhancements will only 39, limited value for users. Intel processors are not very tied to RAM bandwidth.
These chips will also carry other improvements made by Intel, such as faster Wi-Fi and support for Intel's dynamic tuning technology. They will collectively target the 7W envelope (Intel 10nm 10nm parts do not fit in values below 9W). They offer a maximum frequency of 4.9 GHz, compared to 4.1 GHz for Ice Lake 10 nm processors. According to Intel, the U and Y series are for customers who want exceptional processor performance, but do not care about graphics. Apart from the new 6-core SKU, all new chips are also quad-core parts.
Our reading of the situation is as follows: Intel is struggling to contain a resurgent AMD by doubling its level in a market where AMD has always been the weakest: the mobile. 10 nm had to be on the market by the 2020 holidays for a host of reasons, but Intel does not manufacture enough chips to engage in a 10-nm top-down refresh in this segment. We now have a combination of 14nm and 10nm components to meet the general needs of the market. The 10nm processors offer a higher CPI and a greatly improved graphics core, but a significantly lower frequency. The 14nm chips will theoretically anchor the product on the market with a "halo" six-core coin.
But this time, the situation is different. When Intel processors went from 2C / 4T processors to 4C / 8Ts in mobile systems, they kept the line on 2C / 4T configurations for multiple product cycles. Indeed, he had the thermal margin to spare. This time, the company telegraphed that its 15W six-core processor was out of breath for a metaphorical air. We do not know what are the real improvements between the Core i7-8565U and the Core i7-10710U, but we can bet that their size is lower than the 16 and 41% cited by Intel. And if, by chance, you get a 25W laptop with a Core i7-10710U, it will not offer a battery life equivalent to the same configuration with a 15W processor unless the manufacturer provides it with A much heavier battery – which means you may get more cores and an equivalent lifetime, but you will have to pay for it with extra weight.
One of the long-standing trends in semiconductor manufacturing has been a steady decline in the number of major foundry players. Twenty years ago, when 180-nm manufacturing was state-of-the-art technology, no less than 28 companies deployed the node. Today, three companies are developing 7 nm technology – Samsung, TSMC and Intel. A fourth, GlobalFoundries, has since left its leading edge business to focus on specialized foundry technologies such as its 22nm and 12nm FDX technologies.
What is sometimes forgotten in this discussion is the existence of a secondary group of foundry make deploy new nodes – but not at the forefront of technological research. China's Semiconductor Manufacturing International Corporation (SMIC) announced it would start recognizing 14-nm revenue from volume production by the end of 2019, just over five years after Intel launched its sales on this node. TSMC, Samsung and GlobalFoundries all have a production capacity of 14 nm, as does UMC, which introduced the node in 2017.
The secondary sources of a node, such as UMC and SMIC, are often not included in manufacturing comparative tables such as the one presented below, as the companies in question offer these nodes after being deployed as advanced products by big foundries. In many cases, they are operated by smaller customers with products that do not make headlines.
The SMIC, however, is a case apart. SMIC is the largest semiconductor manufacturer in mainland China and builds chips ranging from 350 nm to 14 nm. The company has two plants capable of handling 300 mm wafers, but if the transition to 14 nm is an important part of China's long-term semiconductor initiative, SMIC is not expected to have a large 14-nm capacity in the near future. The high utilization rate of the company (~ 94%) prevents it from acquiring a large additional capacity to devote to a production of 14 nm. SMIC is vital to China's long-term manufacturing goals; The country's "Made in China 2025" plan foresees that 70% of the national demand for semiconductors will come from local companies by 2025. It is essential to reach the production of SMIC and put on line new ranges. of products. This distinguishes the company from a foundry such as UMC, which has generally chosen not to compete with TSMC for advanced process nodes. SMIC wants this company – it just can not compete.
Zhao Haijun and Liang Mong Song, Co-CEOs of SMIC, published a report declaration on the company's 14-mile ramp, as follows:
FinFET research and development continues to accelerate. Our 14nm is in production at risk and should generate significant revenues by the end of the year. In addition, our second generation FinFET N + 1 has already started engaging customers. We maintain consistent, long-term cooperation with our customers and seize opportunities emerging from 5G, IoT, automotive and other trends.
Currently, only 16% of the semiconductors used in China are built in this country, but the country is increasing its semiconductor production capacity faster than anywhere else on the planet. The company is investing in a $ 10 billion manufacturing facility that will be used for dedicated 14-nm production. SMIC is already installing equipment in the completed building. Production should therefore increase in this plant in 2020. Once online, the company will have much more capacity of 14 nm (SMIC's main known customers are HiSilicon and Qualcomm). In the past, Texas Instruments had already built with the company (it's not clear if it still does), as well as Broadcom. TSMC and SMIC have been the subject of several lawsuits for misappropriation by intellectual property; both cases were settled amicably with substantial payments to TSMC.
Despite these expenses, analysts do not expect SMIC to immediately catch up with the main players in the foundry in other countries; Analysts told CNBC that it would be necessary a decade for society to bridge the gap with other major players. The exact dimensions on the 14nm node of SMIC are unknown. Foundry nodes are defined by the individual company and not by a global standard organization or by reference to a specific metric. Those seeking additional information on this topic find it here.
Wednesday August 7th, AMD launched the 7nm update of his family of Epyc processors. These new hearts are not limited to Intel in one category, they bring huge improvements in each category. AMD has reduced its prices by heart, increased the IPC and promises to provide many more processor cores than an equivalent Intel socket.
It was only another time that AMD nearly defeated Intel so decisively: the introduction of the Opteron dual core and the Athlon 64 X2 in 2005. The launch of Epyc this week seems more important. In 2005, AMD's dual-core processors matched the number of Intel processors, outperformed Intel's core clock and core processors, and were quite expensive. This time, AMD is targeting the trifecta, with higher performance, more cores and lower per-core pricing. This is the most serious attack against the high-end Intel Xeon market, launched by the company.
Industry analysts have already predicted that AMD's server market share could double in the next 12 months, reaching 10% by the second quarter of 2020. A larger share of the data center market is essential objective of AMD. A larger share of the corporate and data center market will not simply increase AMD's revenue, it will help stabilize the company's financial performance. One of AMD's critical weaknesses over the last two decades has been its reliance on low-end PCs and retail channels. Both markets tend to be recession sensitive. The low-end computer market also offers the least revenue per socket and the lowest margins. Business cycles are less affected by slowdowns. AMD briefly achieved its goal of a substantial market share for businesses in 2005-2006, when its market share for servers had broken by 20%.
Fans like to focus on the performance of AMD desktops, but apart from games, overall PC sales are declining. Growth in narrow categories such as 2 in 1 was not enough to offset the overall decline in sales. Although no one expects the personal computer market to fail, it is clear that the 2011 economic downturn was not a shock. AMD still has an interest in fighting to expand its share of the desktop and mobile market, but it makes more So it makes sense to fight for a share of server space, where revenue and unit shipments have increased over the past 8 years. The year 2019 may be a difficult year for server sales, but the general trend of migrating workloads to the cloud shows no signs of slowing down.
In our discussions on Rome, we focused primarily on the Epyc 7742. This graph, derived from ServetheHome, illustrates Epyc's performance against Xeon on more SKUs. Viewing at the bottom of the pile:
A pair of AMD Epyc 7742 costs $ 13,900. A support of 7502 (32C / 64T, base of 2.5 GHz, amplification of 3.35 GHz, $ 2600) is equivalent to $ 5,200. The Intel Xeon Platinum 8260 processor is a $ 4,700 processor, but there are four in the system with the highest score, for a total cost of $ 18,800. AMD processors worth $ 13,900 earn you 1.19 times more performance than Intel processors worth $ 18,800. The comparison does not improve with the falling of the pile. Four E7-8890v4 would cost nearly $ 30,000 at list price. A pair of Platinum 8280 costs $ 20,000. The 8676L is a $ 16,600 processor at list price.
But it's not just the price, or even the price / performance ratio where AMD has an advantage. Intel heavily subdivides the features of its product and bills much more. Consider, for example, the price difference between the Xeon 8276, 8276M and Xeon Platinum 8276L models. These three processors are identical, with the exception of the maximum amount of RAM supported by each. The price, however, is anything but.
In this case, "Maximum Memory" includes Intel Optane. The 4.5 TB RAM memory assumes that 3 TB of Optane is installed next to 1.5 TB of RAM. For comparison, the Rome 7nm processors offer up to 4TB of RAM support. It is a standard feature built into all processors, simplifying product purchases and future planning. AMD not only offers chips at lower prices, it is also interested in Intel's market segmentation method. Good luck justifying a price increase of $ 8,000 for additional RAM support when AMD is ready to sell you a 4 TB addressable capacity at base price.
One of AMD's talking points with Epyc is how it delivers the benefits of a 2S system in a 1S configuration. This table of ServetheHome shows the differences:
The advantage of AMD here is that it can simultaneously hit several Intel weaknesses. Need a lot of PCIe lanes? AMD is better. You want PCIe 4.0? AMD is better. If your workloads evolve optimally with the hearts, no one sells more cores per socket than AMD. Intel can still claim some benefits – it offers L3 unified caches much larger than those of AMD (each AMD L3 cache actually is 16 MB, with a slice of 4 MB per core). But these benefits will be limited to the specific applications that respond to them. Intel wants suppliers to invest in creating support for its persistent Optane DC memory, but nothing is said. how much do it. The current low prices of NAND and DRAM have made Optane's competition on the market much more difficult.
The move to 7nm has given AMD an edge in terms of power consumption, especially when you plan to end the server's life. STH indicates a single-threaded power consumption on a Platinum Xeon 8180 at ~ 430W (wall power), compared to around 340W at the wall for the AMD Epyc 7742 system. They note however that the high number of cores on the latter AMD processors will allow them to remove between 6 and 8 Intel Xeons 2017 sockets (60 to 80 cores) to consolidate workloads into a single AMD Epyc system. The energy savings from removing 3-4 dual-socket servers are well above the difference of about 90 W between the two processors.
Features like DL Boost can give Intel a performance advantage in AI and machine learning, but the company will fight hard. Until now, the data we have seen suggests that these factors can help Intel. match AMD as opposed to beating him.
The catalog prices we quoted for this story are the official prices that Intel publishes for Xeon processors in 1K units. It is also notorious that they are inaccurate, at least as far as major OEMs are concerned. We do not know what Dell, HPE and other vendors actually pay for Xeon processors, but we know that it is often well below the list price, which is usually only paid for by the retail network.
The gap between Intel's list prices and actual prices may explain why Threadripper did not have a lot of market penetration. Despite the fact that Threadripper processors have offered a lot more cores per dollar and better performance per dollar for two years, OEMs sharing sales information, like MindFactory, report very weak sales of Threadripper and Skylake-X. However, Intel has not shown any particular interest in lowering the price of Core X. It continues to position a 10-core Core i9-9820X as a suitable competitor for chips such as the Threadripper 2950X, despite AMD's superior performance in this game. This strongly implies that Intel has no particular problem the sale The 10-core processors to OEM partners who want it, despite Threadripper's superior price / quality ratio and AMD's share of the workstation market, is quite limited.
While Intel has cut its HEDT prices (the Core i7-6950X at 10 cores was worth $ 1723 in 2016, compared to $ 900 for a Core i9-9820X today), it has never tried to make a price / performance comparison against Threadripper. If this bulwark is to collapse, Rome will be the processor that will do it. Ryzen and Threadripper will be considered more credible workstation processors if Epyc begins to penetrate the server market.
Intel can reduce its prices to meet AMD in the short term. In the long term, we will have to challenge AMD directly. This means that more cores will need to be delivered at lower prices, with larger amounts of memory supported by socket. Cooper Lake, which is built on 14nm and includes additional support for new AVX-512 instructions focused on AI, will arrive in the first half of next year. This chip will allow Intel to focus on some of the markets it wants to compete with, but will not change the base count differential between the two companies. Similarly, Intel may have difficulty setting up a $ 3,000 to $ 7,000 premium for the support of 2TB to 4.5TB of RAM, since AMD is willing to take over up to 4TB of memory on each processor socket.
We do not know yet if Intel will increase the number of central servers with Ice Lake servers or what types of designs it will market, but ICL in the servers is in at least a year. By the time the ICL servers are ready for delivery, AMD's EUV 7 nm designs are also ready. After launching the mother of all refreshment cycles with Rome, AMD's challenge over the next 12 to 24 months will be to demonstrate the continued pace of updates and continuous performance improvement. If so, it is truly able to create the kind of stable business market that has been desired for decades.
When AMD launched the dual-core Opteron and its consumer counterpart, the Athlon 64 X2, the company finally felt that come. A little over a year later, Intel launched the Core 2 Duo. AMD spent the next eleven years roaming the wilderness. Later, the leaders would admit that the company had gone out of sight and was distracted by the acquisition of ATI. A series of problems followed.
The simplistic assumption that P4 Prescott was a disaster for which Intel could not recover had proved inaccurate. In the past, attacking Intel has often been likened to hitting a rubber wall with a Sledgehammer (pun intended). Distorting the wall is relatively easy. To destroy it completely is a much more difficult task. AMD may have the best opportunity to take market share in the company it has always had with Epyc 7 nm, but server sharing construction is a slow and cautious process, not a wind sprint. If AMD wants to keep what it builds this time, it has to play its cards differently from 2005-2006.
But that being said, I do not take lightly phrases like "golden age". I use it now. Although I do not make any predictions as to its duration, the 7nm Epyc's debut officially formalized it: welcome to the second golden age of AMD.
Intel may have launched Cascade Lake relatively recently, but another refresh of the 14-nm server is already on the horizon. Intel has lifted the veil on Cooper Lake today, giving new details on how the processor integrates into its product line with the 10-nm Ice Lake server chips supposed to be queuing for the deployment in 2020.
Cooper Lake features include support for Google's bfloat16 format. It will also support up to 56 processor cores in a snap-in format, unlike Cascade Lake-AP, which can scale up to 56 cores but only in a welded BGA configuration. The new take would be known as LGA4189. There is reports that these chips could offer up to 16 channels of memory (since Cascade Lake-AP and Cooper Lake use multiple chips on the same chip, Intel could run up to 16 channels of memory per socket with version double chip).
Bfloat16 support is a major addition to Intel's artificial intelligence efforts. While 16-bit semi-precision floating point numbers have been defined in the IEEE 754 standard for over 30 years, bfloat16 changes the balance between the format used for significant digits and that used for exponents. The original IEEE 754 standard is designed to give priority to precision, with only five bits of exponent. The new format allows a much larger range of values but with less precision. This is particularly useful for artificial intelligence and deep learning calculations, and is a major step on Intel's path to improving the performance of artificial intelligence and deep processor learning computations. Intel has released a White Book on bfloat16 if you are looking for more information on the subject. Google says that using bfloat16 instead of the conventional semi-precision floating point can generate significant performance benefits. The society written"Some operations are related to the memory bandwidth, which means that the memory bandwidth determines the time spent in such operations. Storing the inputs and outputs of memory bandwidth-related operations in bfloat16 format reduces the amount of data to be transferred, improving the speed of operations. "
The other benefit of Cooper Lake is that the CPU would share a socket with the upcoming Ice Lake servers in 2020. A theoretically important distinction between the two families is that Ice Lake servers at 10 nm can not support bfloat16, while 14nm Cooper Lake servers will. This could be the result of increased differentiation of Intel's product lines, although it is also possible that this reflects the difficult development of 10 nm.
The introduction of 56 cores as a base indicates that Intel expects Cooper Lake to expand to more customers than the Cascade Lake / Cascade Lake-AP target number. It also raises questions about the type of Ice Lake servers that Intel is going to market and the possibility of seeing 56-core versions of these chips as well. To date, all of Intel's 10-nm Ice Lake messaging has focused on servers or mobile devices. This may reflect the strategy used by Intel for Broadwell, where desktop versions of the processor were scarce, and where server and server components dominated this family – but Intel says later the fact of not publishing Broadwell desktop was a mistake and that the company had gaffed by skipping the market. Does this mean that Intel keeps launching an Ice Lake desktop or if the company has decided to no longer use its desktop computer? made understand that this time is not yet clear.
Cooper Lake's attention to AI treatment means that it is not necessarily meant to go with AMD's next 7 nm Epyc. AMD has not talked much about AI or machine learning on its processors and, although its 7nm chips add support for 256-bit AVX2 operations, the company's CPU division does not tell us has not yet hinted that a particular goal is the AI market. AMD's efforts in this area are still based on a graphics processor and, although its processors will certainly work with AI code, it does not seem that the market is at the same level as that of Intel. Between the addition of a new support for AI to existing Xeons, its products Movidius and Nervana, projects like Loihiand plans the data center market with Xe, Intel is trying to build a market to protect its high-performance computing and high-end server operations, and to address Nvidia's current dominance of the industry.