Bondrewd
Veteran
Obviously, they're having a first in their history mobile push proper.
Ding-ting-ding-ding.
They've pitched CZN + N22/23 and the feedback was great.
Speculation: GPU Performance Comparisons of 2020 *Spawn*
- Thread starter eastmen
- Start date
- Status
They've pitched CZN + N22/23 and the feedback was great.
Bondrewd
Veteran
Eh, the more patient ones are expecting SKU wars, and they'd be right on target.
So if I can summarize...
1. AMD will beat the 3080/90
2. But they don’t really care about all that
3. They’re not planning to make money on discrete GPUs anyway
4. They’re all about CPUs and APUs now.
Am I close? I guess it makes sense. That means they can sell a few really fast expensive chips, claim the crown and go back to building CPUs while everyone is busy refreshing Newegg waiting for new GPU wafers to come in.
1. AMD will beat the 3080/90
2. But they don’t really care about all that
3. They’re not planning to make money on discrete GPUs anyway
4. They’re all about CPUs and APUs now.
Am I close? I guess it makes sense. That means they can sell a few really fast expensive chips, claim the crown and go back to building CPUs while everyone is busy refreshing Newegg waiting for new GPU wafers to come in.
Bondrewd
Veteran
Kinda.
Depends on how are they going to push the 'efficiency' angle.
Ye.
Ye.
Hell yeah, RNR has been the quickest mobile ramp in their entire history.
They've exited Q2 with RNR accounting for something like >25% of their mobile revs.
C&G is posting record quarters while their AIC share is shrinking each Q.
Fun.
Ding-ding-ding.
Huawei-freed wafers will be delivered late Q4 to early Q1 so they'll be finally able to supply GPUs en masse too.
Last edited:
Because we don’t know! We just know that if they had a model it would work!
Looking at Nvidia revenue forecast, I would say that “news” is ... erroneous.
Bondrewd
Veteran
Bingo, SS is free slots galore.
Thanks for the explanation.
Cool, i remember fusion frenzy, fun little game, very nice graphics also for the time.
Great times, the good thing with competition, AMD is also going to drop 20TF+ GPU's with advanced features. Only benefits us gamers.
In the GPU market there is much of profits to be made, much higher then the console APU industry for amd. I think they will try to compete.
Bondrewd
Veteran
That doesn't mean it's their focus in wafer allocs.
They're a CPU vendor, an explosively successful one at that.
Then you look at quarterly Intel x86 revs and weep.
GPUs are funny toys, but if you're a kickass CPU vendor you gotta do what you gotta do.
And they want to be successfull in the GPU market too im sure, just like with their CPU's being 'bad' before (FX8xxx), now their basically more popular then intel (even though not the fastest at the high end). RDNA3 might be very intresting and be here sooner then later.
Bondrewd
Veteran
Yes, but when you have a choice between selling more CPUs and maybe selling more GPUs you're picking the former every single time.
x86 CPU TAM is gigantic while Intel is weak.
Oh, yes.
But they're gonna shoot #2 arty first.
Let's not get ahead of ourselves. AMD is enjoying a little sunshine right now but they're far off from dethroning Intel. And as we've seen before it can (and probably will) turn back in Intel's favor soon enough. So they shouldn't give up on the lowly gaming market just yet.
The commentary is an accurate near-term profile of AMD’s market position right now, and I don’t think it is “getting ahead of ourselves” in any way.
Sure, revenue diversification matters in long term with AMD itself as the testament, but so as maximising RoI aka best bang for the bucks. With the finite amount of TSMC production capacity (after fulfilling client deliverables), why would they prioritise their traditionally weak discrete GPU products, over driving further their momentum on server and consumer CPU/APU that are the bigger money pies?
IMO it is not about giving up, but picking fights for the bigger fishes to fry. It is not life and death for them to deprioritise dGPU production/launch at this very moment. That and also Nvidia not being on TSMC 7nm gives them plenty of leeway, if they truly have a great product stack in waiting.
Last edited:
arandomguy
Veteran
The bigger concern to me with AMD from a consumer stand point is more so the Enterprise/Consumer dichotomy as opposed to CPU/GPU. If demand is high across the board the DIY consumer market is likely going to be low if not the lowest in the priority chain for allocation. If we look at their last statement there was a distinction in that they specified Milan would be shipping in 2020 while they only used the more vague term of launching in 2020 for everything else. Allocation is almost surely going to prioritize to corporate customers, whether it's Epyc, APUs for OEMs, or semi-custom. Not just due to immediate contractual and profits motivations but to work on those relations over the fickle DIY segment.
In regards to whether or not waiting is sensible I think it's more complicated than it might seem. In normal circumstance given the timelines it would be but how confident are people in the actual shipping situation assuming an AMD announce in November? Given the high demand in all sectors currently I'm actually personally skeptical that AMD is going to have high volume for DIY for either CPU or GPU this year. But the bigger concern to me is the risk factors from not only Covid but also the seemingly trending upwards movement in crypto prices again. There could very well be a risk of higher prices across the board despite the higher competition.
In regards to whether or not waiting is sensible I think it's more complicated than it might seem. In normal circumstance given the timelines it would be but how confident are people in the actual shipping situation assuming an AMD announce in November? Given the high demand in all sectors currently I'm actually personally skeptical that AMD is going to have high volume for DIY for either CPU or GPU this year. But the bigger concern to me is the risk factors from not only Covid but also the seemingly trending upwards movement in crypto prices again. There could very well be a risk of higher prices across the board despite the higher competition.
Bondrewd
Veteran
A little?
En route.
No, not a single chance.
Absolutely, which is why they're heavily pivoting their client uArch into laptop/mobile.
No, they're shipping everything not Cezanne and N23/24 this year.
Even then AIC N23 might just about make it.
Last edited:
Even if i got the choice of selling a wafer full of Athlon 200/3000G vs. selling a wafer full of A100?
I agree in general, that CPUs generally have a much bigger TAM and you have to have massive supplies to cater to the hyperscalers so every wafer counts for you. But there's one more angle: Intel is vulnerable right now and probably will be for at least a year in Desktop and Server, where Nvidia comparatively is in a much better position.
Bondrewd
Veteran
Not bleeding edge.
Yeah assuming that wafer of Athlons gets me OEM cred as needed.
AMD has no GPU cred as is, especially in DC so fabbing say, MI100s beyond most basic inventory buildup would be a missstep.
Until 2023 best case.
Product-wise nah, inside people's brains yeah, by a long shot.
Intel has been forfeiting their mainstream clout since around Broadwell mess.
Leoneazzurro5
Regular
Well, Athlon 2000/3000G are not supply constrained as they are not on 7nm AFAIK.
- Status