Intel Arc Battlemage Could Arrive Before Black Friday, Right in Time for Holidays

[AleksandarK]

According to the latest report from ComputerBase, Intel had a strong presence at the recently concluded Embedded World 2024 conference. The company officially showcased its Arc series of GPUs for the embedded market, based on the existing Alchemist chips rebranded as the "E series." However, industry whispers hint at a more significant development—the impending launch of Intel's second-generation Arc Xe² GPUs, codenamed "Battlemage," potentially before the lucrative Black Friday shopping season. While Alchemist serves as Intel's current offering for embedded applications, many companies in attendance expressed keen interest in Battlemage, the successor to Alchemist. These firms often cover a broad spectrum, from servers and desktops to notebooks and embedded systems, necessitating a hardware platform that caters to this diverse range of applications.

Officially, Intel had previously stated that Battlemage would "hopefully" arrive before CES 2025, implying a 2024 launch. However, rumors from the trade show floor suggest a more ambitious target—a release before Black Friday, which falls on November 29th this year. This timeline aligns with Intel's historical launch patterns, as the original Arc A380 and notebook GPUs debuted in early October 2022, albeit with a staggered and limited rollout. Intel's struggles with the Alchemist launch serve as a learning experience for the company. Early promises and performance claims for the first-generation Arc GPUs failed to materialize, leading to a stuttering market introduction. This time, Intel has adopted a more reserved approach, avoiding premature and grandiose proclamations about Battlemage's capabilities.



View at TechPowerUp Main Site | Source

 

[Daven]

We have three major releases all at the same time in the second half of the year.

Battlemage - Intel remains below midrange even if they double the performance of Alchemist
RDNA4 - AMD is rumored to only go up to the next gen midrange segment based around $400 and the performance of the 7900XTX
Blackwell - Nvidia is alone at the top. Will they go super high in price?

 

[Bwaze]

I don't even expect Nvidia to ship anyhing this year. They are planning to reveal RTX 5090 and RTX 5080 first, but they have absolutely no challengers in the high end, and they aren't even planned! So they might as well just tease them, with planned availability "somewhere in the Q1 2025, if we feel ike it".

As for Intel, how many postponements did the previous generation need to actually arrive on the shelves? I can't imagine everything running smoothly now.

 

[kondamin]

I don't even expect Nvidia to ship anyhing this year. They are planning to reveal RTX 5090 and RTX 5080 first, but they have absolutely no challengers in the high end, and they aren't even planned! So they might as well just tease them, with planned availability "somewhere in the Q1 2025, if we feel ike it".

As for Intel, how many postponements did the previous generation need to actually arrive on the shelves? I can't imagine everything running smoothly now.

Last time the hardware was sitting on shelves waiting for drivers to be ready.
considering the leaps and bounds those drivers have matured over the past 2 years I don't see why there should be all that many issues this time over

 

[Bwaze]

Last time the hardware was sitting on shelves waiting for drivers to be ready.
considering the leaps and bounds those drivers have matured over the past 2 years I don't see why there should be all that many issues this time over

Last time they also had drivers from the years of making integrated graphics, which their discrete GPU line leans on heavily.

And I don't expect that many issues. Just enough to not really expect them to release a GPU without any delays.

 

[Daven]

Last time they also had drivers from the years of making integrated graphics, which their discrete GPU line leans on heavily.

And I don't expect that many issues. Just enough to not really expect them to release a GPU without any delays.

Uh, did you miss the hot mess of Alchemist drivers at launch? Intel’s experience with integrated GPUs did absolutely nothing to prepare for that launch. It was quite embarrassing.

 

[Vayra86]

1200

Is that your final offer?

Uh, did you miss the hot mess of Alchemist drivers at launch? Intel’s experience with integrated GPUs did absolutely nothing to prepare for that launch. It was quite embarrassing.

Raja even admitted it was a colossal misstep to base themselves on the IGP experience and drivers. It wouldn't scale.

 

[JasBC]

Uh, did you miss the hot mess of Alchemist drivers at launch? Intel’s experience with integrated GPUs did absolutely nothing to prepare for that launch. It was quite embarrassing.

Exactly: there is precedence for Intel fumbling the bag.

The one you're replying to is hightlighting how Intel's launch drivers for Battlemage aren't necessarily going to be usable just because Alchemists' drivers are now, just like how the work on iGPUs didn't translate into competent drivers for Alchemist at launch.

 

[Vayra86]

Exactly: there is precedence for Intel fumbling the bag.

The one you're replying to is hightlighting how Intel's launch drivers for Battlemage aren't necessarily going to be usable just because Alchemists' drivers are now, just like how the work on iGPUs didn't translate into competent drivers for Alchemist at launch.

Well, Battlemage is another iteration, its not something new anymore for them. Its bound to be better.

 

[phints]

Much more interested to know if Intel will release HEDT CPUs on Intel 4 lithography this year. They have fallen so behind.

 

[Gmr_Chick]

Article: "This time, Intel has adopted a more reserved approach, avoiding premature and grandiose proclamations about Battlemage's capabilities."

They've only adopted this new approach because Raja is no longer there with his choo-choo train of hype. Thank god.

 

[Vya Domus]

you all just have unrealistic expectations for a first time gpu maker.

Except they're not first time GPU makers, they had some know-how from iGPUs. Unlike CPUs every new GPU architecture is pretty much a blank slate, new design, new ISA, new compiler, new drivers, etc, somehow AMD and Nvidia manage every time, Intel couldn't. I didn't expect them to but even after all this time new games still suffer from embarrassingly bad performance or are just outright broken on launch, it is no longer unrealistic to expect better from them.

Look at this, their fastest GPU is the same as a 5700XT, a mid range card from 5 years ago.

 

[Darmok N Jalad]

Maybe I'm just missing it, but has there been much talk about net performance gains with this next gen? I mean, something like a 20% generational uplift is normally a giant win, but that isn't exactly going to get them very far.

 

[HisDivineOrder]

We have three major releases all at the same time in the second half of the year.

Battlemage - Intel remains below midrange even if they double the performance of Alchemist
RDNA4 - AMD is rumored to only go up to the next gen midrange segment based around $400 and the performance of the 7900XTX
Blackwell - Nvidia is alone at the top. Will they go super high in price?

If 7900XTX performance is priced at $400 next gen, that's going to be amazing. I think Lisa's going to look at that performance and think it's looking more like $700 because Nvidia will be sitting around that price, too, for probably the same kind of performance and she rarely goes aggressive on GPU pricing.

Thing is, Intel is doing good work keeping everyone honest on the lower to mid range. Hopefully, they can break into the high end eventually. We know AMD won't be competitive until Intel forces them to be. They've proven that over the last two generations.

 

[Minus Infinity]

I don't even expect Nvidia to ship anyhing this year. They are planning to reveal RTX 5090 and RTX 5080 first, but they have absolutely no challengers in the high end, and they aren't even planned! So they might as well just tease them, with planned availability "somewhere in the Q1 2025, if we feel ike it".

As for Intel, how many postponements did the previous generation need to actually arrive on the shelves? I can't imagine everything running smoothly now.

They will have no incentive to launch so early. RDNA5 is very late 2025 at best. Even if RDNA4 N48 is 7900XTX levels of raster and only $450-500, 5080 is not competing in that segment so there is no heat on Nvidia. Why not wait and launch a full product stack for once: 5060 right through to 5090 say in Q2 2025.

BTW if N48 is 16GB and does achieve 7900 XT/XTX raster and is at least 35% stronger than 7700XT in RT I would buy one to replace aging 2080 Super for sure.

 

[watzupken]

While the cadence in which Intel is updating their dGPU sounds ok since it matches both AMD and Nvidia, but the truth is that they are already 1 generation behind. Current Arc dGPUs were meant to compete with RDNA2 and Nvidia’s Ampere. Fortunately for Intel, AMD did not move the needle significantly this gen with RDNA3. But against the Ada range of GPUs, they are less efficient and not competitive in terms of performance. So with Battlemage replacing Celestial, unless we are getting a significant step up in performance, otherwise, it’s likely going into the budget bin like it is now. And when it comes to budget GPUs, they will have a tough fight with AMD. Therefore, if they are launching their new GPUs around the same time AMD and Nvidia releases or announces their next gen GPUs, you can be sure the announcement will get “crowded out” by Nvidia.

 

[GodisanAtheist]

"Intel's next GPU will arrive in time for that special day when retailers discount their trash products that haven't sold all year so they can get a tax write-off before their fiscal year closes out"

Doesn't sound as good I guess...

 

[Solaris17]

I welcome it! Its great having a third player. I just wish they would release it sooner. Its more of a gut feeling, but I dont think releasing with everyone else is the right call. Maybe they need to do it later in the year to give celestial the miles it needs to get completed.

I just dont think its a great stance from a market presence perspective. I was impressed by alchemists performance for a first go, but it will take a few generations before they start scratching at the high level cards. Releasing next to them I cant imagine helps them.

Maybe they are going for the cost angle. If someone releases a card for 800 and you are doing 450 it gets your eyes I guess.

In any case sweet! Cant wait for battlemage and celestial what a wild time, Intel descrete cards.

 

[Firedrops]

RDNA4 - AMD is rumored to only go up to the next gen midrange segment based around $400 and the performance of the 7900XTX

This just reeks of MLID nonsense. He rumours that for every generation.

 

[Onasi]

This just reeks of MLID nonsense. He rumours that for every generation.

It’s theoretically possible, but almost hilariously unlikely. That would be a Pascal level price-performance uplift, even more so, actually, in this specific case. And we all know that the GPU market just doesn’t play like that anymore. I mean, current 400-ish AMD offering is the 7700XT (well, depends on location and discounts, but I’ve seen it at 400). The 7900XTX is what, almost twice as fast? So we’re talking potentially 80-90% gen on gen improvement at the same price point? Nah fam, that’s whack. Would be cool, but even Pascal wasn’t this big.

 

[Vayra86]

It’s theoretically possible, but almost hilariously unlikely. That would be a Pascal level price-performance uplift, even more so, actually, in this specific case. And we all know that the GPU market just doesn’t play like that anymore. I mean, current 400-ish AMD offering is the 7700XT (well, depends on location and discounts, but I’ve seen it at 400). The 7900XTX is what, almost twice as fast? So we’re talking potentially 80-90% gen on gen improvement at the same price point? Nah fam, that’s whack. Would be cool, but even Pascal wasn’t this big.

Yep.

Its virtually as unlikely as AMD abandoning their high end proposition, but we have seen them do it before, the company seems to work on a different kind of logic than most. There is almost no reason for AMD not to push their GPUs further, not in the least because they need them to keep powering consoles.

 

[Dr. Dro]

Yep.

Its virtually as unlikely as AMD abandoning their high end proposition, but we have seen them do it before, the company seems to work on a different kind of logic than most. There is almost no reason for AMD not to push their GPUs further, not in the least because they need them to keep powering consoles.

At this point it's not a matter of reason, but a matter of ability. RDNA 3 shows that there's some serious issues with their architecture, on paper, the 7900 XTX should be much faster and much more efficient than it actually turned out to be. Taking one step back and carefully reworking and reevaluating their technology and position over the course of this generation may very well turn out to be the wisest move AMD could make, especially if they can't pull a GPU that is 100% faster than the 7900 XTX at the same power envelope out of their hat. With N31's real world performance equating the smaller, leaner, and much more efficient AD103 at best, and a prospective 5090 being at least twice as fast as a 4080 Super (provided they achieve a worst case scenario of +70% over 4090, which is what it did over 3090 - on average, it's 100%), they'd have to to even hope to keep up performance on the halo segment. Top-end GPU performance doubling every generational cycle is still a thing, as amazing as it sounds.

IMO, just work on the tech, on the drivers, and then make a big boy card

 

[Macro Device]

I mean, something like a 20% generational uplift is normally a giant win

1080 Ti is faster than 980 Ti by about two thirds. About 50 percent faster per dollar.
2080 Ti is faster than 1080 Ti by about a third, also being more expensive. It was and still is a horrible $ per FPS release.
3080 Ti is faster than 2080 Ti by about 50 to 60 percent. It's mediocre at best considering the even more so increased price. About 40% $ per FPS improvement.
4090 is faster than 3080 Ti by about the same 50 to 70 percent. Nothing impressive about that, either, since 4090 is so ridiculously expensive.

20 percent is just "you'd better have done nothing" kinda uplift. NV with the major help from AMD are doing their best to gaslight us into thinking it's okay for a next-gen $600 GPU to only outperform last-gen $600 GPUs by a dozen percent. 20 years ago, it was a thing for $100 GPUs to outperform last-gen $200 GPUs by about 100 to 300 percent depending on a game.

she rarely goes aggressive on GPU pricing.

Never. She never does.

I really do want to have an Intel GPU but none of them fit the description of "6700 XT is a snail compared to them." 'haps the Battlemage will deliver... Knowing Intel, it's a really high possibility. Also knowing Intel, there could be more than a million "buts."

 

[Vya Domus]

RDNA 3 shows that there's some serious issues with their architecture, on paper, the 7900 XTX should be much faster and much more efficient than it actually turned out to be.

People keep saying this but there is actually nothing at all to suggest this is the case, for instance 4090 has roughly 35% more shading power yet is only 15-20% faster than the 7900XTX in raster performance, 7900XTX also uses 15% less power just as expected given the performance gap, I don't see anything abnormal here. RDNA3 is less efficient the lower the graphics load is because of the chiplet design, this is unavoidable and Nvidia chips are going to behave the same when they'll switch to MCMs.

 

[Dr. Dro]

People keep saying this but there is actually nothing at all to suggest this is the case, for instance 4090 has roughly 35% more shading power yet is only 15-20% faster than the 7900XTX in raster performance, 7900XTX also uses 15% less power just as expected given the performance gap, I don't see anything abnormal here. RDNA3 is less efficient the lower the graphics load is because of the chiplet design, this is unavoidable and Nvidia chips are going to behave the same when they'll switch to MCMs.

View attachment 343010

Just look at its fillrates, compute, amount of execution resources... efficiency is a bit of an umbrella term, but in this context it's more than just performance to power consumption ratio, it can also be about making the best use of the hardware`s available resources. It seems obvious at this point that the RTX 4080's full potential is extracted, this is clearly not the case with the 7900 XTX right now. Theoretically it should smack the 4080 into the ground, not be a tie to slight win in raster but a murder scene in RT.

Past a point scaling becomes less and less efficient, that is normal. But AMD's architecture hits that wall far, far sooner, you just need to look at the 7800 XT and the 7900 GRE for the supreme example