Doesn't seem to stop Intel from putting Lunar and Arrow Lake into high end laptops with Nvidia 8GB slop, while Halo is multiple times better for a myriad of reasons.
You need medication.
'premium gaming handheld' market doesn't exist outside of your fetishistic wet dreams.
Like even $400 is a very steep price for a handheld, and you want kilobuck+ devices for some reason?
80w can easily give you 3hrs on batteryhalo handheld connected to next-gen powerbank
No need.halo handheld connected to next-gen powerbank
That's tiny, cumulative shipments are a mere fraction of Switch OLED, and that one is the least popular SKU by a country mile.High end x86 handhelds above $400 sold millions of units (Deck's most popular model is always the more expensive one)
YES.but a fugly gaming windows tablet is somehow more justified?
That would make it weigh like a kilo+.80W.h internal battery and 30W operation would provide over 2h autonomy.
Waste of silicon.Strix Halo on 30W does wonders.
You're not rolling these parts full-tilt off the battery.Wendel said the 74.5Wh battery lasted 45 min
I am not so sure if it's 8, 9, or perhaps a much higher number of folks who have been waiting for this.we're getting like 5 design wins for those 8 people who somehow need high performance edge clients without having to take them on the go.
Times have changed. There are uses for large-memory GPGPU computing with some money behind, and possibly volume.High performance expensive NUCs don't move any meaningful volume.
NUCs are where good ideas go to die out of unsuccessful B2B marketing, like Intel's GT3e and later Vega M from 12 years ago.
Lest we forget, Strix Halo is requiring a completely new motherboard platform. They can't just drop it into an existing Strix Point/Hawk Point/Phoenix board and use a repainted case. That takes both time and a willingness to take risks. In addition, we don't know of any exclusivity agreements for specific market segments that might exist.
To ease adoption, AMD should have included Memory On Package.
BK was long gone before Lunar Lake was even a concept.OEMs hate Memory on Package though. Lunar Lake was just a BK special since he was obsessed with Apple. You won't see that again in x86.
OEMs hate Memory on Package though. Lunar Lake was just a BK special since he was obsessed with Apple. You won't see that again in x86.
First, the assumption is they have to sell at a loss, and the second is that they wouldn't be willing to do so, when doing that would result in competition getting killed, which is a long-term win.I don't think Nvidia would ever sell at a loss. Especially when they pretty much own the market.
Wdym separate board. What platform flexiblity?In addition to that dGPUs can be on a separate board easing design and reducing TTM, while also having flexibility in choosing platforms.
Isn't part of the picture that Intel used to give a lot of design support? (Besides other 'incentives' which a market leader was able to apply.)AMD has a problem of OEMs not adopting AMD CPUs, or doing only as an afterthought after they spend all their energy and resources on Intel stuff.
In addition they prefer to buy memory from memory makers, rather than from a middleman. Plus more control over inventory I guess.the reason OEMs don't like [memory on package] is because OEM like to rape customers on cost of memory upgrades.
MCM modules are a DISTANT afterthought in the market. Daughter board dGPUs might as well not exist for most of the market. They are almost always tightly coupled with the CPU on the main board for all sorts of reasons.First, the assumption is they have to sell at a loss, and the second is that they wouldn't be willing to do so, when doing that would result in competition getting killed, which is a long-term win.
@ToTTenTranz Not very surprised. Way better attempt, but the same rule applies. They have to keep trying.
@LightningZ71 In addition to that dGPUs can be on a separate board easing design and reducing TTM, while also having flexibility in choosing platforms.
The wrong idea people refuse to give up is that it has to be all-or-nothing. It has to be dGPU-or-nothing, when they can all coexist, serving multitudes of different customers.
Still orders of magnitude over a fugly tablet going for $2K.That's tiny, cumulative shipments are a mere fraction of Switch OLED, and that one is the least popular SKU by a country mile.
This again...That would make it weigh like a kilo+.
You know what else was incoming until it wasn't? Navi 4C.of course AMD will continue halo line, medusa halo incoming