XTX is just the big boye bin.
When cryptoshit crashes and burns again.
Particularly for Nvidia, this time their goods are extra spicy for mining.
AMD is, probably not for a...
Wonder Egg Priority ep1 is something out of this world.
Absolutely nuts given its Wakabayashi's directorial debut.
Who dares to summon the devil during the holy week of Pat G finally asssuming the role he should've had back in 2013.
Expect fucking nothing;...
so you can complain about muh paperlaunch or what.
That'd be very not nice.
Argh, you wish.
Back to ugh, ah yes, Navi2x.
Nah, I gotta get 'em first.
Get in line, I need my 7Fx3s still.
Do you really think I give a shit what flavour of nVidia oven people buy?
I hope but can't guarantee because Milan and RNR would eat a lot of wafers.
AMD marketing truly has the toughest of jobs since people will bitch and moan about everything.
>start the hype early
Damn this is 11th degree cope
YES and YES (if limited quantities).
Well, d'oh, soon it'll all be over.
They lifted like the whole methodology.
It's certified good shit, it's how they crank 4 dGPU dies in a row (over ~2Q).
Renoir iGP reaches 2.4GHz OCed perfectly comfortably.
Referencing TPU DB pre-launch is a bad idea, takes time for them to update the plugs.
That one is a 4SA part.
They've already did it anyway ffs, N10 is a 2560ALU GPU with 64ROPs, while Polaris was 2304 with half of 'em.
Being the best works very well as far as pushing things down the stack go.
That's silly, but the products are good (excellent? depending on how...
Imagine a good ass marketing campaign but without the hustle of actually doing one.
Hell yeah drag them on stage and slit the throats....
That's the whole point here, isn't it?
Consoles do it even harder.
Limited resources <> max IQ possible.
The estimate console number we're running with here is worst case, period.
I wish but alas.
what if pigs flied?
Someone at marketing team spazzed and thought it was a good idea.
C'mon it's the woodscrews company.
Oh wait they also did a video about fans and...
Well uh, they've been timing October launch for some time now.
I buy things in volume.
290x was inadequately cooled which resulted in not very nice temps and sound signature.
Narrative always shifts and goes wherever the market...
Also the cooling actual should be around Sapphire Pulse grade boards (or a bit higher), assuming AMD didn't make some...
Eh, Navi1x one was on-point (maybe even a bit undersold?).
Same with Rome or Matisse or Renoir.
It in general responded well to going wider with more FF to throw around which is what N2x does.
Eh, Maxwell itself is falling off at a rapid...
Allocated statically thus you can safely assume worst case.
PS4p GPU making the thing go jet engine begs to disagree.
Pushing a given...
But even GCN scales with width.
Pretty sure Baby Polaris > Small Polaris > Middle Polaris show pretty sane scaling with SE count.
(well nothing ever scales linearly, but getting close is trve art)
Also might as well say 8SAs because SEs actual play a weird token role...
Yeah, it is early October for the event and mid-late October for product availability.
I guess we'll have to wait and see; but given that nV had a 103 die planned (and 106 being a fair but under 104 this time to fit mobile properly)...
I think Apple had the full configs.
They just wanna scale the same SE n times over without some specific part choking so tons of "extra"...
ROPs don't spill out of SA and Tonga was 4SE with 384b GDDR5 and that was years ago.
Well duh, they do.
Doubt, N10 has 64ROPs and gets not-much perf from mem OC.
Same for that class nV GPUs.
GA104 is what, 64 ROPs running...
It's good shit.
The real question is not the IP or the chips actual, but prices, memory configs, dates and availability.
From the block diagram...
If only N10 didn't have 64 of those.
Quite literally >50% more bandwidth over N10, and that's just offchip.
They've just redesigned the RBEs for...
Directly with ALU count while being overweight, yeah.
You see, they're scaling out, not up here.
Issues only arose in Fiji when they've plugged...
No lol, exists neither in products nor parts' lists.
I think you should read on what fused multiply-add is.
It has 20 DCUs and 40 CUs.
It does have 80 scalar units, yeah, 1 per SIMD, which each CU now has two of.
It's a dual compute unit.
Navi21 has 40 dual compute units.
For 80 single compute units.
Each CU still does 64 32b FMAs.
And since FMA is 2 ops,...
No, this is talking WGP.
N21 has 40 WGPs.
Engineering is always full of compromises.
Because it's not really a net win until you plop 4+ stacks.
The fastest one is 2.54Gbps which isn't really *that* fast.
Yeah. between CoWoS-STAR...
FS2020 is cheating, it's the new Crysis.