Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
AMD's share price rises 52% in 24 hours, the sharpest increase in the company's history
It's been quite a week for computing giant and perpetual NVIDIA adversary AMD. After releasing their Q1 2016 quarterly report the company's share value skyrocketed by over 50% over the course of a single day's trading.
http://www.pcgamesn.com/amds-share-price-rises-52-in-24-hours-the-sharpest-increase-in-the-companys-history
http://www.pcgamesn.com/amds-share-price-rises-52-in-24-hours-the-sharpest-increase-in-the-companys-history
Comments
With it coming back down the way it has it was someone getting a big chunk out of the way, but not ready to take the further plunge just yet :P
http://forums.mmorpg.com/discussion/449736/amd-expects-1-5b-in-future-revenue-for-three-new-gaming-processors-but-what-are-they/p1
Stock is holding at around 3,70$
http://www.nasdaq.com/symbol/amd/interactive-chart?timeframe=5d
Reasons:
1. Better than expected Q12016
2. China deal of 293m$+licencing fees
3. Console deals - PS "Neo", Nintendo and presumably new XBox1.5
4. Polaris and Zen (AMD CPUs will be deployed in China in supercomputers, presumably its Zen)
2. Much better than expected (upgraded) Q22016 (and onwards) estimates
80% on Nvidia users have GPU weaker than PS4.
Only 30% of Nvidia users have upgraded to Maxwell.
There is huge number of people who are potential GPU upgrade candidates and vast majority of those are <350$ (aka mainstream) segment.
High end is only a small fraction of the market.
Building huge chips is hard. Building early chips on new process nodes is hard. Adopting new memory standards is hard. Building new GPU architectures is hard. Doing all of them at once is really, really hard, to the degree that your chip is likely to be a catastrophic failure. Think Radeon HD 2900 XT or GeForce GTX 480 as the examples here.
If you start with a small chip on a new process node first, things that go wrong don't go as wildly wrong. You can then learn from your mistakes and build a bigger chip later that fixes a lot of the things that would have gone wrong if it had been your first chip on the new process node.
There's also the issue of memory types. If you want GDDR5X or HBM2, then you're waiting for it to be available, and that delays everything until around the end of this year. The lower end cards are fine with GDDR5, so you can release them sooner rather than making everything wait. AMD is certainly going that route; it's unclear whether Nvidia will do likewise.
That means if Intel wants to keep its current "must have" status they will have to go <14nm soon which is unlikely because they just released a new 14nm series. So AMD will gain market shares and increase its value which will increase stocks.
Downside, people bought Intel CPUs for their PC and unless something major happens they will keep their current CPUs for some time. But you could counter that with increased console sales and AMD going for the next gen consoles.
On the GPU market there isn't enough information to make a serious prediction imo. We know what AMD is planning and have a rough estimate of how good the new chips will be but Nvidia, to my knowledge, hasn't released anything of that kind yet.
But even without that information I assume that AMD will lose on the GPU market again. Because whenever I read posts about AMD I see stuff like "their drivers suck", "Nvidia has better frames", "They are just worse since 20xx". Without any offense, Nvidia does have a huge group of fanboys. No matter if Nvidia lies about specs or pays for preferred compatibility (you think those Nvidia logos on a loading screen are there by chance?) people seem to prefer the company so based on that I assume Amd will stay the 2nd.
Honestly, I think the idea that AMD is magically going to pull a rabbit out of their ass and release a CPU that's on par with Intel's as far as perf/watt is a pipe dream. They simply don't have the R&D budget that Intel has and it would take essentially an act of God (or corporate espionage) to pull that off.
Now, being that real competition is good for the consumer, I will GLADLY eat crow if that ends being the case and I'm wrong, but I doubt I will be.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
Sandy Bridge has been out since 2011. That's plenty of time for AMD to figure out what made it so good and mimic it. Intel has only gotten marginal improvements out of their CPUs since then, so I think it's very plausible that AMD could catch up or at least come close. If Zen is 90% or 95% as good as Kaby Lake by whatever metrics you like, that would be more competitive than AMD's CPUs have been in a decade, and probably good enough.
ARM ended up in a dominant position in mobile because Intel didn't even compete in that space, and by the time they decided to it was too late and would require too much money to compete. Even still, they're doing great with the atom processors and getting them into lower end tablets.
That being said, I should clarify I was referring strictly to the desktop PC market (I figured that was a safe assumption to make about this discussion). Frankly ARM/mobile is something I could really care less about.
I think things are going to start evening out in the mobile space pretty soon here. Mobile has reached the point that desktop did a few years ago, where the processing needs of the average user is far below what the top end processors are capable of. With the sole exception of some of the mobile games, people don't need quad core processors to load and run chrome and youtube and such on the phone.
If you look back 20 years, a run of the mill PC was easily $1500+ and that speed could be justified in "everyday" tasks, if loading up your word processor, internet browser, etc, etc was significantly faster, it made the whole experience better. Now, on a desktop, the difference between a 5yr old i7 and a brand new i7 is .5 second vs .6 seconds. Essentially it's at a point where the difference is under the radar of what the average human could detect or would even care about.
Mobile processors and products are very close to that same point. I just recently got an LG G5 from a G3, and while it is clearly faster, it's not something that if I were forced to go back to the G3 that I would cry into my cereal in the morning. Now, make me go back to the HTC Incredible. Yes, big difference, however that was also 5 years ago. So there's that.
As far as Sandy Bridge vs now, while yes, each iteration has been marinal, the overall difference between sandy bridge and skylake is between 35 and 45% clock for clock. That means a skylake at 3ghz will perform on average 35-45% better than sandybridge. Factor in that skylake is clocked significantly higher stock and boosts quite a bit higher, and it amounts to quite a lot of difference.
Now, is that difference meaningful in a gaming sense? No, not really. There is little justification for a skylake upgrade for a gamer. However if you do other things with your PC that isn't able to use GCN or CUDA to do the compute, than there are HUGE gains to be made moving to skylake.
As to your last point, yes, *if* Zen is 90-95% the performance of Kaby Lake, then that would be a massive win, and is definitely in the area of a difference that most people wouldn't care, even gamers a 10% processor difference might amount to a 1-2% fps difference (and likely not even that). So if AMD could sell a chip that performed 90% as well as intel and at 80% the cost, they would have a massive win on their hands.
Frankly I just don't see it happening, but again, I will gladly eat crow if I'm wrong. Intel has basically been operating as an unopposed monopoly in the desktop space (slight exaggeration but you get my point) and some real genuine competition from AMD would be fantastic.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
For that matter, if a larger R&D budget can create anything, how do you explain that Intel's GPUs are still so far inferior to AMD and Nvidia?
If the larger R&D budget always wins, how do you explain AMD being clearly superior to Intel in the Athlon 64 days? And that was back when Intel's process node advantage mattered a ton for desktops. Today, it still matters when you're power limited, but it's nearly irrelevant in desktop processors.
I'm not claiming that AMD Zen is going to destroy Intel, or even that it will certainly be highly competitive. I'm only claiming that it's plausible that Zen will be competitive. We'll have to wait and find out.
Sorry, you're right, i'm an idiot, I didn't realize snapdragon's were ARM also.
Regardless, point still stands that Intel didn't make a serious push into mobile processors until it was far too late and they've realized that now to compete they would have to dump a stupid amount of money into R&D as well as trying to take more of the pie from ARM.
It's a lot easier to be in at the start of a market and become a major player, than it is to enter an existing market and take a majority share of that market. Look at Microsoft, Oracle, Intel, etc. These were all the companies who were there at the beginning, they built the market which is why they have the large slice.
Insurance is another example, when you look at companies like Allstate or State Farm who have been around for 100 years, it's hard for a new company to come in and take a majority portion of that market.
Back to the desktops, things were a lot different back in the Athlon 64 days, and what you're not accounting for is that Intel was fat and lazy at that point because they had no real competition prior to AMD. PowerPC processors were useless for just about everything a normal PC user did which is why it almost put Apple out of business (for example).
So what you had is a fat, lazy Intel who was like a complacent old lion that was busy resting on their laurels when AMD came in and bit em on the ass without them knowing. Once that happened Intel kicked into high gear and started hosing them consistently, and that has purely to do with $$$, in so much as R&D budget, as well as them being able to afford to hire the best people and entice the best people to come to the company.
"The surest way to corrupt a youth is to instruct him to hold in higher esteem those who think alike than those who think differently."
- Friedrich Nietzsche
"We all do the best we can based on life experience, point of view, and our ability to believe in ourselves." - Naropa "We don't see things as they are, we see them as we are." SR Covey
That's my biggest gripe with new consoles being released every other year and all these reiterations of Windows and "new" pc's. They sell them based on some game or bit of software that most want(or need for business purposes) and then leave us all dangling waiting for more, only to be supplied with yet another reincarnation/port of something we've had for years(that now will only work on the new hardware/framework).
I for one refuse to buy anymore new consoles/tablets/whatever until the games/software for them add up to enough in my mind to justify the purchase and that doesn't mean 10-20 remakes of hockey or shooter games, I'm talking real rpgs like we were given for the PS1 and 2. The only reason I update my pc as often as I do is because I seem to burn out my motherboards on a bi-yearly basis, by the time my warranty replacement dies it's usually time to replace the whole cpu anyway. Some things should last longer than a new year's resolution... just saying.
Also they don't need to do corporate espionage to get the technology. They are all pulling from the same scientific discoveries, just engineering their products differently to get different benefits and drawbacks. Take the FX and Core i7 for example. The FX has a lot more memory bandwidth and larger caches. So something like 3D rendering which would use all that memory bandwidth is faster with an FX processor. On the other hand the Core i7 can do something like composite faster because it accesses its lower amount of memory faster.
Intel has been cutting back and it makes sense to prepare now for a couple years when they will be trading benchmarks with AMD.