Thursday, July 4th 2024
Intel Arc "Battlemage" Xe2 GPUs with 448 EUs (56 Xe cores) Spotted in Transit
Intel very much does intend to make discrete gaming GPUs based on its Xe2 "Battlemage" graphics architecture, which made its debut with the Core Ultra 200V "Lunar Lake-MX" processor as an iGPU. With its next generation, Intel plans to capture an even bigger share of the gaming graphics market, both on the notebook and desktop platforms. "Battlemage" will be crucial for Intel, as it will be able to make its case with Microsoft and Sony for semi-custom chips, for their next-generation consoles. Intel has all pieces of the console SoC puzzle that AMD does. A Xe2 "Battlemage" discrete GPU sample, codenamed "Churchill Falls," has been spotted making its transit in and out of locations known for Intel SoC development, such as Bangalore in India, and Shanghai in China.
Such shipping manifests tend to be incredibly descriptive, and speak of an Arc "Battlemage" X3 and Arc "Battlemage" X4 SKUs, each with 448 execution units (EU), across 56 Xe cores. Assuming an Xe core continues to have 128 unified shaders in the "Battlemage" architecture, you're looking at 7,168 unified shaders for this GPU, a staggering 75% increase in just the numerical count of the shaders, and not accounting for IPC increase and other architecture-level features. The descriptions also speak of a 256-bit wide memory bus, although they don't specify memory type or speed. Given that at launch, the Arc A770 "Alchemist" was a 1440p-class GPU, we predict Intel might take a crack at a 4K-class GPU. Besides raster 3D performance, Intel is expected to significantly improve the ray tracing and AI performance of its Xe2 discrete GPUs, making them powerful options for creative professionals.
Source:
Tweaktown
Such shipping manifests tend to be incredibly descriptive, and speak of an Arc "Battlemage" X3 and Arc "Battlemage" X4 SKUs, each with 448 execution units (EU), across 56 Xe cores. Assuming an Xe core continues to have 128 unified shaders in the "Battlemage" architecture, you're looking at 7,168 unified shaders for this GPU, a staggering 75% increase in just the numerical count of the shaders, and not accounting for IPC increase and other architecture-level features. The descriptions also speak of a 256-bit wide memory bus, although they don't specify memory type or speed. Given that at launch, the Arc A770 "Alchemist" was a 1440p-class GPU, we predict Intel might take a crack at a 4K-class GPU. Besides raster 3D performance, Intel is expected to significantly improve the ray tracing and AI performance of its Xe2 discrete GPUs, making them powerful options for creative professionals.
30 Comments on Intel Arc "Battlemage" Xe2 GPUs with 448 EUs (56 Xe cores) Spotted in Transit
Specs look promising for a big uplift at the top end of the stack.
The A770 was comparable to a 4060 I believe, but lower priced.
As you say Drivers are likely the biggest make or break here, but I'd predict a significantly stronger launch (driver wise) than Alchemist was, assuming what Battlemage is builds upon Alchemist.
4070 Ti performance at, say, 350-400 bucks would be a decent offering. Regular 4070 level - not so much, that would have to be more in the 250-300 ballpark. RT-wise they are already in a decent place, so good pricing is what will be necessary for them to succeed, that, and, as you mentioned, getting drivers right out of the gate. This go around it won’t be excusable to have them at Alchemist launch level.
In average frames in 1080p difference between A770 and Nvidia RTX 4070 is 80%! 75.8 > 136.5 = 80.1%.
However I'm still excited to see how Battlemage will do in games. I found the experience alright, even in older titles.
If Nvidia and AMD start release their cards November and this gets released same time or later, then it will fail.
So they need to push and get it released ASAP, before September.
Intel has been selling some their products at cost or even cheaper, this has been documented and it's an undeniable fact. They've also basically bribed OEMs with "joint development funds" to keep AMD's chips out of the best laptops....this is EXACTLY the type of stuff Intel will do in the dGPU market. Again, they're not gunning for Nvidia like some of you have deluded yourselves into believing, they're gunning for AMD, and the mention of the console business in this article reminds us of that. Any success in dGPUs at AMD's expense will be further used by Intel to disadvantage AMD further in x86 (i.e. Intel will just tell OEMs that they'll give them the dGPU at cost or even less if they if they use Intel CPUs and not AMD CPUs in their prebuilts).
I know a lot of you think that Intel's entry into dGPUs will somehow usher in a a new era that's better for consumers....but that's just not going to happen. Basically what's happening is that another company, with a proven history of anti-competitive and anti-consumer practices (that it still does at present) is entering the market, not some brand new startup up committed to spending the status quo, Intel IS the status quo. I'm sure some of you will try and say "It's better than nothing", but even that cannot be assumed, and you can basically justify anything no matter how bad, with such thinking. Mark my words....this will NOT be a net gain for consumers, enthusiasts, or the dGPU market as a whole.
At the end of the day they are selling parts and wanted to make profit out of it, the reason why they sell their first gen cheap is because its underperforming even with the superior spec (256-bit memory bus, 16GB VRAM etc). AMD should've put more effort to consumer GPU market now rather than CPU. We all know now that RDNA3 is a flop, the upcoming RDNA4 is what RDNA3 supposed to be, but even so we haven't seen any RDNA4 parts today. I just hope with Intel getting better in consumer GPU market, it will accelerate AMD effort to make theirs better too. Lately they only focus on AI bullshit.
Intel and AMD both have scored supercomputer deals recently. But I think such computers tend to work around software limitations that would burden other market segments, so I can't tell whether AMD or Intel is really meaningfully outdoing each other relative to Nvidia.
On the consumer side, www.tomshardware.com/pc-components/gpus/stable-diffusion-benchmarks showed intel had surprisingly competitive performance compared to amd. But, there's a few caveats. They were using Windows and all the AMD SD users I know claim the windows performance is a disaster compared to using AMD + Linux + SD. I haven't heard if the same is true for Intel because I don't know anyone who has even tried. On Nvidia it doesn't really matter, the performance between linux and windows for SD isnt that significant.