quilciri :
Great, AMD can cool their overheated processors. Leave the cooling to the aftermarket and spend the R&D on a new performance chipset/Socket/processor. Get in the GAME AMD, we need you to compete.
I'd love to see Zen processors outperform Skylake, but I've huge respect for AMD regardless.
With
10% of Intel's budget and personnel, AMD has developed x86 CPU's, ARM CPU's, GPU"s APU's, adaptive sync that does not require additional hardware, their own memory modules (not to mention HBM integration), and an entirely new graphics API that has become the basis of DirectX12.
But of course, no one cares that they only have 1/10th of Intel's money to spread between both their CPU and GPU divisions. They want AMD CPU's to outperform Intel's and cost less.
AMD is very much "in the game", imho.
41 upvotes...LOL. I love AMD, but 42 people seem to be missing the point (including yourself).
You're idea of "IN THE GAME"=
Losing 7B over the last ~15yrs, and I can't even remember the last PROFITABLE FY. Over that time amount of shares have nearly doubled (diluted that is, as they kept devaluing them to get more credit, like the fed keeps printing money we don't have).
NV now has over 81% of discrete (gaining yearly from 63% a few years back).
Intel owns the whole show now on cpu, basically same for apu/server.
AMD margins in toilet (while NV/Intel hitting highs Q after Q, with both having profitable year after year).
Had to sell fabs, land, lease land etc. Only things left is IP really, and that won't be worth much soon if the next gpu/cpu are not a huge hit. At current burn rate, they might make it 2yrs before selling out, or bankruptcy or going further into debt. But that latter deal is difficult considering their now "junk bond" credit rating so to speak.
They are no longer a LEADER in anything, which is precisely why they have ZERO pricing power and margins so low. This isn't AMD employees fault, it is a MANAGEMENT problem. If ZEN is smaller than Intels FULL CPU/IGP die size, they will lose yet again. They needed to take all of Intel's gpu side and pour it into CPU. That would have made for a WINNER, but I fear they went with about Intel's cpu size and will just end up priced to death until Intel can either blow them away again, or maybe just drop higher clocked models to edge out AMD (if even that is needed). They are not going to run away with the gpu race, but they have the first chance since 2000-2003 to run away with cpu since Intel dedicates 1/2 the silicon (more in upcoming chips) of their die to GPU now. They would be left with the same choice as AMD has been recently (going 8 cores, usually not utilized, and running hot/high watts), until they could properly design a HUGE quad again. AMD might have 3yrs to rule and PRICE higher than Intel and possibly get out of debt during that time. From all the data/leaks I've seen though, I fear AMD just tried to match Intel's CPU side (die size I mean) instead of going HUGE to dominate.
If they didn't go at least 1.5x Intel's cpu side, management should all be fired. This is their last chance to produce a winner in a time where Intel has been concentrating totally on ARM for years. Intel isn't prepared to see a quad die the size of their entire cpu/gpu. Even 1/2 of the gpu side would be ok and probably win most stuff, but Intel would simply sell 8c models for the same price until they get something real out the door. But if they went really close to full Intel die size, even the slapping of 2 Intel quads together wouldn't win much (see how well AMD's 8 cores do now vs. Intel quad, due to so little software REALLY using 8 cores efficiently), so AMD would have a few years for Intel to fix things giving a few years of great profits. Before when AMD had a winner they had no ability to really product more than 20% of the market. This time would be COMPLETELY different as they could fab at TSMC/GF and possibly samsung (as they share everything with GF and it wouldn't take too long to tool up at samsung if desired). AMD could really make a HUGE dent in Intel profits, market share, and R&D over a 3yr period or so (maybe longer, as Intel would need a radically different design to catch a really high IPC quad that big). Even going to 10nm would not help Intel win, it would merely lessen the blow due to smaller silicon than 14nm and edge them closer in perf.
DX12 was being worked on about the same time as Mantle (with NV BTW) and NV was chosen to run and XBOX1 game (forza, made for AMD on consoles) to DEMO DX12. DX12 is not based on Mantle (For that matter even Vulkan is FAR different, which isn't to say bits weren't reused, just the whole product is very different).
http://www.pcper.com/reviews/General-Tech/GDC-15-What-Vulkan-glNext-SPIR-V-and-OpenCL-21
Note the main selling point of Mantle was that it was HLSL.
"Khronos is doing something else entirely." IE Shader compilation is completely removed from the driver etc..etc.
Not trying to be anti-AMD here, just laying out facts you seem to be completely ignoring. Financial reports don't lie. Market shares don't lie etc. You can say it's amazing they're still here (testament to great engineers in the face of moronic management decisions), but not that they are IN THE GAME so to speak. They are losing their butts in everything cpu, gpu, drivers, gsync (while expensive) better than freesync, mantle killed (well duh, poor guy can't start his own API with 20% share to push it), pushing console mistake (far too low of margins. should have passed like NV), 3rd party for Geforce Experience competitor (again not as good), etc. Money spent on consoles totally stalled R&D on PURE CPU/GPU which is why they lost the cpu race and lost all that share of gpu to NV while drivers suffered too. This is exactly what Jen said as the reason why they passed. It would take too much away from R&D on core products. Fools took that as "butt hurt NV over losing consoles". No, VERY SMART management realized bean-counters were right, margins were not worth stopping R&D in CORE products and support for those core products.
I really hope AMD capitalizes on the first time in a decade they could produce an actual winner (have 2x the size of Intel's cpu core to work with on same process essentially) and easily capable of producing massive amounts of chips between multiple fabs unlike any other time in their lifespan. I'm astounded AMD pushed the PURE CPU for moving up GPU. Their GPU will be running against a highly competent NV chip, while their CPU (IF HUGE) could have wasted Intel in perf for at least a few years. If they cpu is HUGE, they can still win for a while, but it should have came BEFORE gpu, as NV/AMD are both working from the same place and neither has a way to dominate everything like you could with a 2x cpu die size that takes YEARS to design (4-5 usually, like russian roulette, only you don't find out your dead for 5yrs). Intel can maybe speed that kind of thing up to 3-3.5 but not more than that no matter how much money they throw down. Intel's only option for years would be to plop two down (running hotter, more watts until gpus stripped), or shrink to more mhz with 1 or 2 quads sans gpus. Neither way works to win, just lessen the pain. You can't really price to death a king in perf either. Intel can't afford to do much of that for 3yrs while trying to keep samsung, TSMC, GF and all of ARM off their backs. AMD could price a huge die at or ABOVE Intel current pricing at all levels. Intel shareholders would not stay in the stock if they saw cpu margins tank (along with profits) to price AMD to death. A loss in billions of Market value gives Intel far less competitive room to move against all others mentioned. They would likely have to keep pricing pretty high in the hopes most wouldn't go for AMD massively at higher prices. It would be interesting times for Intel for sure no matter how you slice that...LOL. APU's have no shot either, as they are priced to death by ARM coming up, Intel moving down. The HUGE cpu is the only thing that can make AMD real profits for a few years straight.
Again, I really hope I'm wrong about what MANAGEMENT chose to do with cpu die size (already wrong pushing it out for pulling in gpu). The Zen CPU needs to be 1.5-2x Intel's cpu side or they blew it. I'll buy AMD for a WINNER even at $100 over Intel, but I won't buy AMD for a tie or anything close with no sign of another product that will make serious cash to keep them afloat for years to come. I feel sorry of AMD employees who keep having to follow incompetent leaders who sabotage their ability to produce winning products. I hope management went for the CLEAR WIN (and pricing power for years), not merely matching Intel cpu die for silicon savings which amount to nothing in a tie and instantly bring price pressure to stop AMD from gaining share or making profit. Being competitive isn't enough, they have to have a WINNER this time (see being semi-competitive for the last decade, it got them NOTHING but loss after loss yearly). Dirk Meyer was right when he said AMD needs a KING first before trying other crap (consoles, etc), as he left the company in ~2011. I hope they stripped the GPU side to make a KING cpu, rather than some silicon savings for a tie or worse. Stripping gpu and making cpu marginally bigger probably just gets close to Intel, but that won't give a billion in profits for the year (can't save that much from smaller silicon as Intel moves to 10 and prices down again) for years like a HUGE die that actually WINS would. A silicon savings (on such small share anyway) would only raise margins a few points and only until Intel hit 10nm shortly after, but a winner could raise them 25 pts and produce huge profits for ages.
I hope people get the point. I hate management at AMD, not AMD workers or the company. I'd really like to buy their stock again one day too! But not until I see a ZEN die size that amazes me. You're not IN THE GAME if you're losing money, market share, buildings, R&D dropping for 3yrs etc etc, while the exact opposite is happening for NV/Intel.
- 
Hoping AMD change their mind...
Facebook
Twitter
Instagram