GPUs in Laptops, why even bother.
Intel has finally unveiled its upcoming ARC Alchemist series of graphics cards, or more rather they’ve unveiled the unveiling of their upcoming ARC Alchemist Graphics Cards as the company intends to fulfil its promise of “launching” something within the first quarter.
Now what exactly will Intel be officially launching as Intel hypes up its “A New Stage of the Game” event on March 30th? Well Intel will be unveiling their mobility range of ARC Alchemist products.
How exciting. Lisa Pearce, VP and GM of Intel’s Visual Compute Group officially broke the cover of Intel’s fore coming announcement and answered a variety of questions Intel themselves have asked with the general response being typical marketing fluff.
In response to why exactly Intel have sought to release their ARC Alchemist for mobile oriented products first rather than the traditional DIY PC market the answer is quite obvious. Money.
Laptops are horrid devices, especially mobile devices that actually do come with a discrete graphics card solution, they are woefully inefficient piles of junk with performance in comparison to traditional desktop PCs that are beyond an embarrassment.
I much prefer traditional mobile devices, efficient devices that don AMD Ryzen APUs rather than Intel Core CPU’s paired with an NVIDIA GeForce discrete offerings that just chew through battery life and your wallet.
However, it’s these cancerous BGA laptops and notebooks that just generate a constant flow of cash for corporations such as Intel, NVIDIA and AMD as the mobility market is ever growing each and every quarter.
So it makes perfect sense for Intel to prioritize a market that’s basically a guaranteed sale for Intel and the OEM partners who choose to manufacture mobile devices packing Intel’s ARC A370M.
The second question that Intel asked themselves is regarding the sort of performance consoomers can expect from the illusive ARC A370M, with the general answer being vague and just another massive marketing fluff statement full of bullshit and deception, so par for the course with Intel.
While no specifics were shared regarding ACTUAL performance figures because of course that bombshell won’t be dropping until the end of the month, or rather the beginning of the second quarter.
The only takeaway here is that Intel will be marketing the ARC A370M as offering “twice” the performance compared to current Intel integrated graphic solutions such as their “Iris Xe MAX” DG1 solution which is already regarded as being a slow pile of absolute shit, especially when Intel had the bright idea in turning an iGPU into a dGPU.
Considering how Intel’s Iris Xe integrated solutions are convincingly thrashed in terms of gaming performance by AMD Ryzen 5000 series APUs containing RX Vega integrated graphics and much more so by the current crop of Ryzen 6000 series of mobile processors featuring RDNA 2 based 600M graphics.
It’s fairly safe to assume that the ARC A370M would probably be based on one of Intel’s lower-end SKUs of their “DG2”, particularly around that of 128 Execution Units in total. (1024 shaders)
Considering how Intel’s marketing fluff piece describes the ARC A370M as being twice as fast compared to Iris Xe solutions at “1080p medium” on Metro Exodus compared to the Intel Core i7-1280P, such performance gains would still put the ARC A370M below that of the AMD Radeon 680M integrated graphics on Metro Exodus.
Across other various titles providing that similar performance scaling can be achieved the A370M would either barely match the performance set by the Ryzen 7 6800H APU or come out ahead with a solid enough performance gap but is undoubtedly guaranteed to fall behind NVIDIA’s MX550.
An entry level mobile dGPU featuring 1024 CUDA Cores and is based on the previous generation Turing architecture.
But I digress, providing that Intel can acheive an illusive 2x performance gain over its Iris Xe based iGPU solution across the entire board then and only then would it actually be comparative to NVIDIA’s three year old scrap, but comparative nonetheless.
With Intel’s disclosure describing a total system TDP of just 40W with the ARC A370M pared with an Intel Core i7-12700H it’s quite apparent by now that while Intel’s ARC Alchemist is pretty damn slow by todays standards and even yesterdays standards it is pretty damn efficient. The Ryzen 7 6800H comes in with a TDP of around the same on its lonesome so I’d rule in favor of Intel in that regard.
While the performance of a 128 EU Intel ARC Alchemist graphics card is more than certainly going to be much better in a desktop configuration than a mobile one, the performance of what’s believed to be one of Intel’s “better” entry-level solutions will more than certainly continue the trend laid out by DG1 as being a massive pile of shit in comparison to other mobility products and it certainly would be when Intel finally decide to drop a deuce on the desktop market.
Intel’s entry level A370M with what’s expected to contain 128 EU’s will be challenged by NVIDIA’s four year old TU117 based MX 550 and current AMD Ryzen APUs. Intel have a lot more planned for the market such as a separate mid-tier product containing a maximum of 256 EU’s (2048 shaders) along with the flagship DG2-512 SKU containing a rather infamous figure for Raja Koduri, 4096 shaders along with a cut down 386 EU (3072 shaders) variant.
While we’re going to have to wait a little bit before proper performance figures and specifications of the Intel ARC A370M is properly unveiled on the 30th, initial performance figures from Intel’s deceptive marketing tactics aren’t looking all that enticing.
And with how many times I’ve stated personal expectations for ARC Alchemist, with initial performance targets of an RTX 2070 Super for the flagship DG2-512 from insider information years prior, considering how Intel’s “128” EU mobility graphics card is going toe-to-toe against NVIDIA’s own entry level product and an AMD Ryzen APU rest assured as Intel expand upon “ARC” with subsequent product releases the consumer disappointment for the market’s “third player” will grow exponentially.