Gearbest Tech
[Gearbest Technology]After 25 glorious years, Intel is about to bid farewell to its important seat in the Dow Jones Industrial Average, and this position will be taken over by Nvidia. S&P Dow Jones Indices officially announced this decision not long ago, and the change will officially take effect before the market opens on Friday, November 8. This change marks that Intel, the former leader in chip manufacturing, is facing new challenges and difficulties.
In recent years, Intel’s leading position in the field of chip manufacturing has gradually been surpassed by competitors such as TSMC. At the same time, its layout in the field of generative artificial intelligence has also lagged behind. It failed to invest in cutting-edge companies such as OpenAI in a timely manner and missed development opportunities. These factors have combined to cause Intel’s stock price to fall sharply this year, with a cumulative drop of 51.5%, making it the worst-performing stock in the Dow Jones Index, and its stock price has also fallen to the lowest level in the price-weighted calculation of the Dow Jones Index.
However, faced with such a severe challenge, Intel did not give up the opportunity to fight back. Currently, the company is intensively preparing for the release of the second generation Arc GP, trying to achieve self-salvation through the key field of GPU as AI technology is changing with each passing day. As an important support for the development of AI technology, GPU’s performance and efficiency are directly related to the breadth and depth of AI applications. Therefore, Intel’s expectations and investment in Arc GPU are undoubtedly an important layout for its future development.
Specifications revealed
Recently, the Geekbench benchmarking platform exposed information about an independent GPU called “Intel Xe Graphics RI”. This GPU has 160 computing units, that is, 20 Xe cores, paired with 12GB of video memory, and an acceleration frequency of up to 2850MHz. While these have lower core counts than the Arc A770 and Arc A750, their boost frequencies are even better.
Additionally, in the graphics logs of the Linux kernel’s Direct Rendering Manager subsystem, Wccftech discovered traces of another new Arc Battlemage graphics card. This graphics card is also equipped with 12GB of video memory, and the memory rate is 19Gbps, which is slightly higher than the Arc Alchemist series. However, since its memory bus width is only 192 bits, the bandwidth may still be lower than the 512GB/s of the Alchemist series, which is about 456GB/s. There is still uncertainty about the number of Xe cores this GPU will have, but leaked information suggests it may be equipped with 14 Xe2 cores.
The most striking part of the leaked information is the presence of Adamantine caching technology. This is a cache technology developed by Intel for CPUs, similar to AMD’s 3D V-Cache, which stacks additional cache layers on the chip. This cache is called “Level 4 cache,” and while it’s relatively slow in the cache, RedGamingTech says Intel plans to equip its flagship chips with up to 512MB of Adamantine cache.
Performance Outlook
At present, we know very little about the performance of Battlemage, and the small amount of leaked information that has been released does not give any clear hints. The only public independent Battlemage GPU test results seem to confirm Intel’s speculation that it is targeting the mainstream market. The card scored 97,943 points in the Geekbench OpenCL test, lower than NVIDIA’s RTX 4060, AMD’s RX 7700 XT, and Intel’s own Arc A770. Considering AMD and NVIDIA are about to launch new generations of graphics cards, Intel may fall further behind in performance.
If early leaks are to be believed, and Intel indeed plans to use 56 Xe cores, the flagship Battlemage GPU is likely to be targeted at the NVIDIA RTX 4070. This is a smart choice because the graphics card market is currently the most competitive in the $400 to $600 price range, with five or six graphics cards competing in this price range.
Initially, there were rumors that Intel’s flagship card would be targeting the RTX 4080, but that now seems unlikely. This statement may be due to the fact that the actual size of the GPU inside the flagship card is similar to the RTX 4080. Still, that doesn’t mean Intel’s card can reach the level of the RTX 4080.
At the lower levels of the product line, things are equally unpredictable. Currently we only have one leaked benchmark from SiSoftware, showing Battlemage performance in a Lunar Lake notebook CPU. Regardless, it seems unlikely that Intel will directly compete with AMD and NVIDIA’s next-gen graphics cards. Based on the current timeline, it looks like Intel is trying to gain a few months of market advantage before AMD and NVIDIA release a new generation of graphics cards.
It is rumored that a discrete Battlemage GPU may never appear in laptops, which means laptop users may never see a Lunar Lake version of Battlemage.
release date
We originally expected Intel’s Battlemage graphics card to be unveiled in the second half of 2024, but now this time has become increasingly uncertain, as Intel has not yet announced any official information. At the beginning of the year, Intel researcher Tom Peterson said that engineers on the graphics team were developing software for the next generation of GPUs, while hardware engineers had already moved on to “the next project.” Battlemage is ready, at least from a hardware perspective.
More evidence comes from a recent shipment listing two Battlemage GPUs. The listing states that the cards are “for research and development purposes only,” which suggests Intel may have entered the final validation phase. Additionally, Intel itself shared a slide with Japanese media showing new GPUs coming in 2024.
However, whistleblowers have different opinions. YouTube user Moore’s Law is Dead cited Intel sources as saying that Battlemage is targeted for release in late 2024 or early 2025. And at a recent event jointly hosted by Intel and Asus, the manufacturer promised to launch Battlemage before the end of the year.
price strategy
Regarding price, we have reason to believe Intel will focus on entry-level and mid-range graphics cards rather than expensive flagships like the RTX 4090.
Based on the known specifications, there are two possibilities for the price. If Intel chooses to go high-end and build a competitive GPU, the flagship card could be priced around $500. On the other hand, if Intel adopts a conservative strategy, its flagship product may be more affordable, around $350. This pricing strategy will allow Intel to compete with AMD and NVIDIA in the mid-range market and attract more price-sensitive consumers.
magicCubeFunc.write_ad(“dingcai_top_0”);
Login with Google