- Mar 3, 2017
- 1,623
- 5,894
- 136
That has been his MO the entire time, yes.I little cryptic for me. Arrow lake is better than Zen 5 ???
I wonder how he'll spin it.That has been his MO the entire time, yes.
Just one month until the meltdown.
Plenty.How much will Zen5 technical superiority translate to availability of laptops?
Define successful.Someone said that Intel offers much better support to OEMs, and that that is the reason Intel is so much more successful in laptop space.
Lmao no QC efforts will forever remain a badly done/badly timed joke.I imagine Snapdragon would be an additional pressure to force them to do better here?
Aye. Lisa said in the earnings call that they're experiencing a faster ramp and adoption with Strix when compared to Hawk. And that's just the premium part.Plenty.
It does comply with the 45 TOPs requirement, yes.Kraken is going to have enough TOPS to have the AI PC sticker?
Define successful.
I wonder how he'll spin it.
I'd say if AMD actually climbs to 33% during Z5's lifetime, I'll call it a victory. 40%+ would be amazing.85% revenue in laptop market
I'd say if AMD actually climbs to 33% during Z5's lifetime, I'll call it a victory. 40%+ would be amazing.
But I don't have much hope. I suspect the real reason is just that the INTC servicing is Enhanced Massage Parlor tier. You get everything lubricated and stimulated until you're satisfied.
The lubrication being money and the stimulation being Intel boys doing half the integration job so laptop makers just plop the CPU in where they're told and get sent what they need to do and execute it.
Feels like AMD is lacking in terms of "customer service" towards OEMs.
Also, if we could get an OEM worker to talk, I'm pretty sure there would be very loud complaints about the usual AMD problems. Docs being a labyrinth of "go to document X->document Y->document Z->document X->repeat circle", multiple calls to AMD personnel with "we'll get back to you" which comes back after weeks and the answer is "errr we're not really sure", and so on.
For all the loudmouthing he does, George Hotz did illustrate that with detail. And I've heard many many complaints over the years that "AMD does all Open Source, but the docs are so empty nobody but themselves understands anything".
I think you're making a false equivalence here.All this is just bad mouthing that was spread through the net by some viral marketers, if just 1% of this FUD was accurate AMD wouldnt have captured 33% of the server market, that is, the most demanding customers market.
You got it wrong here and wrote a wall of useless text because you re not aware of the reality, AMD servers CPUs are only used by those big names and they had no offering till recently for the low end market, wich they finaly adressed with their Siena plateform.I think you're making a false equivalence here.
First off, the laptop market has extreme competition. Many models, many companies, many clients, meaning questionable margins. The more restaurants at the fair the less it's possible to overprice.
This isn't the case at all of the enterprise/server market, where the buyers are not many, and I mean sure you have a lot of smaller corpos using some servers, but they're peanuts compared to AWS, Meta, Google & co. These heavyweights have a lot of work to put in to optimise their server farms, so they will actually have the competent personnel do all the necessary jobs, while a laptop maker will not be able to pay for the most competent people since their margins don't justify these experts' compensations.
Second, server is really about the whole RAM/mobo/CPU performance over cost. Also the cooling systems and necessary space for the racks, which are bought in bulk. Not for small enterprises but for big ones yes. Actually the biggest cost is in RAM because they have insane amounts, not in CPU. And for the OEMs like HP, Dell, Supermicro, their entire job is to offer the best perf/highest value system they can. And they make the margins to justify the cost of integration over what will be a few systems per vendor, while laptop vendors essentially throw a new laptop out every 2 weeks.
The margins are very high provided you can offer the best CPU/RAM combo.
It makes perfect sense that AMD, who relies on the OEMs to integrate and evaluate, would do very well in server, but poorly in laptops. The margins justify highly paid integrators and long periods of evaluation for server lineups, giving ample room for EPYC to show what it has. Not so in laptops, both in margins being lower and models coming out muuuuuuuch more often. AFAIK you basically have one server lineup per gen of Xeon or EPYC. It's nothing like that at laptop, where they have to cram the design and push it out the door quickly so the next one goes to design.
This is where Intel's fat butt is still firmly sitting on their excessive amount of employees that AMD isn't competing with. Chewing for the server OEMs is neither necessary nor wanted, they'll do the job with ample time and margins. Baby laptop makers would rather get all the help to get a semi-decent, semi-worthwhile laptop out quick, not waste any margins and to send to market.
Just enjoy the movie, friend. If it crashes, it was still a fun ride.Damn, I try not to let the hype get to me, but it still feels like the AMD hypetrain has yet-again set up quite insane expectations.
I feel like if Strix Point is released and only offers 15% ST gain, 15% GPU gain and 10% more battery life, people will be massively dissapointed.
I had no idea what was on the picture ... until i read the filename
sssshhhh. Don't spill the secrets. You'll see more of it soon.I had no idea what was on the picture ... until i read the filename
It wouldn't win any slots like that.I feel like if Strix Point is released and only offers 15% ST gain, 15% GPU gain and 10% more battery life, people will be massively dissapointed.
No it's going off Turin scores.The ridiculous 40% SPECint figure was entirely the creation of a severe misreading of the well-known Keller graph.
In actuality, it's closer to 30%, and adjusted for clock speed, 23-27%.
Probably should've lurked a bit longer.The ridiculous 40% SPECint figure was entirely the creation of a severe misreading of the well-known Keller graph.
In actuality, it's closer to 30%, and adjusted for clock speed, 23-27%.
I will tell you in July when I get mine !!Probably should've lurked a bit longer.