AMD, which lost over $2.8B in 5 years, takes a hit after new report
arstechnica.com
arstechnica.com
AMD is punching way above their weight class lately. It's causing competitors (or really competitor since Intel is their main one) to have to scramble to counter them.
Considering what AMD went through, I indulge in a bit of AMD fanboyism at times, I try to avoid it. Some articles compare the latest Intel releases with AMD, and they often end up saying "ok Intel still has a 10% perf lead in half the tests, but .. twice the price". Am I the only one to consider what AMD did a miracle ? is it Intel friend network turning this into a "meh" ?
I have never been a fanboy of either but for my use case it's a nobrainer right now. Even if I was a fanboy I'd have a hard time justifying getting an intel chip.
But what's up with the single channel ram in the ibm one? Both of the others have dual-channel.
There's no ryzen based laptop atm (regional limits might apply).
One Asus gaming laptop that is basically a desktop that you can carry around, with effort, and that has no battery life to speak of. Not available here.
Two Lenovo laptops (Thinkpad A series and Ideapad), both seem restricted in terms of CPU (aka not the 2700), don't come with the latest GPU or memory (single channel or unknown or limited to 8 GB). None of these is available.
The HP x360. Fancy, limited to 8 GB as it seems. Also doesn't come with the latest APU (only announced for 2500U). Not available yet.
The Acer Swift 3. Limited to 8 GB as far as I can tell. Aaaand not available.
Bottom line: You cannot get a laptop with either the mobile chips or the monster gaming thing now. They exist somewhere, but not here. Most of the announced devices have crippling limitations (memory) and are - for me, someone who deeply sympathizes with AMD - basically DOA. Which probably reinforces the manufacturer's opinion that they should just create Intel based products...?
I want one. I can't get any of these. All of these are kinda bad / quite limited / would be a "I want them for the AMD label, although they suck in some ways" decision.
Fortunately, I can live without such a laptop, for now. I'll just keep an eye out for updates.
Market dynamics I suppose.
More and more i felt the place could be rechristened Apple Technica.
I still love their article on setting up one's own mail server. That was probably the high-water point for their quality.
Both are owned by Conde Nast...
MIT Technology Review remains superb, however: www.technologyreview.com
What does the fact that they picked a report that had an impact on a company (in other words a newsworthy story) with their "journalistic quality"?
Even if the report is back, it had an impact, and that should be reported. You'd rather they didn't covered stories involving upsetting things?
Plus, though slow consoles and crypto will hurt a little, that's really only their existing business. Their future is in their new and competitive product lineup from this year. Epyc particularly should be very good for AMD if it can gain momentum: expensive server CPUs that AMD can manufacture cost-efficiently, and which are competitive with or straightforwardly better than Intel in certain use-cases.
(full disclosure: I indirectly own AMD stock because I expect next year will be better.)
Sony profits rise 346% thanks to strong performance from PlayStation gaming unit
Next week:
New Xbox and new Playstation will be released.
I don't get how they can claim that consoles will slow down.
But their pricing is on par with intel's offerings, so why should I buy the AMD product? For almost my use cases extra cores aren't helping. Outside of some edge stuff at work, the same applies at work.
When I saw AMD's pricing it more or less justified my decision not to buy their stock. I don't think this article is negative. Its just, from an investor's perspective, AMD stock isn't very attractive. I have no idea if a major price cut would help with this, but I think they need to be more competitive with intel in regards to pricing. None of offerings, per core, really beat my 2011 2500k i5. Why wouldn't I just replace it with the newest i5 or i7?
This is just a market correction from ryzen/zen hype. AMD's stock rose way too fast for their actual market offerings. Congrats to those who rode this wave and sold at the right time, but now we're starting to see some sanity here.
Also the article is little more than blogspam of another article that gives a simple summary of the Morgan Stanley report.
https://www.cnbc.com/2017/10/30/amd-shares-crater-after-morg...
The Morgan Stanley report's summary isn't rosy for AMD:
---
Moore lowered his price target for AMD shares to $8 from $11, representing 32 percent downside from Friday's close.
The analyst predicts cryptocurrency mining driven sales for AMD's graphics chips will decline by 50 percent next year or a $250 million decline in revenue. He also forecasts video game console demand will decline by 5.5 percent in 2018.
"We expect cryptocurrency to gradually fade from here, consoles to decline, and graphics to be flattish," he wrote. "To be clear, we admire what the company has accomplished on a fraction of its competitors' budgets in both microprocessors and graphics - our cautious view is based entirely on the current stock price, and the limited potential for upside in 2018 and beyond."
---
Also from a market facts perspective:
AMD today: $10.92
AMD a week ago: $14.25
>That report is so negatively biased it's suspicious.
Well, no one is buying the stock for x reasons so the price is falling. How is that not factual?
Moore might be right that $8 is AMD's rational market price considering their current offerings and sales projections. I don't see how anyone can look at AMD's stock performance in the past month and decide that it isn't destined for a downward slope until it hits a new floor.
This is not a pretty picture for a company that just released its latest and highly anticipated offerings, after years of delays and promises, a few months ago:
edit: oh well, downvotes from fanboys who will be offended at anything but "AMD IS AMAZEBALLS"
You got downvotes for being wrong.
> But their pricing is on par with intel's offerings, so why should I buy the AMD product
This generation AMD is killing the price/performance ratio. Starting off a long winded argument with a factual inaccuracy isn't the best way to get the audience on your side.
> None of offerings, per core, really beat my 2011 2500k i5
Funny, I replaced my 2011 2500k i5 with an R5 1600x and the performance difference is incredible (having 3200Mhz DDR4 helps as well).
You are guilty of the same bias that you're accusing the downvoters of.
Again, synthetic benchmarks that show atypical tasks like video encoding aren't helping sales. No one buying this stock because sales projections are poor. Turns out per core performance is what the market is biased for due to the benefits of it on typical computing tasks. AMD's big gamble with lots of cores didn't pay off like some assumed, thus the market correction we're literally seeing.
The report is fully factual, maybe it gets some causes wrong, but people are not buying this stock, they're selling it off, thus the price drop and the price drop in the future to a perhaps $8 stock price. Considering AMD floats around $4-5 in general, $8 seems like a decent place to be for an AMD slightly off life support. Thus far you have provided ZERO evidence on why a $14 AMD stock price is rational when all evidence points to it being the height of a hype cycle and nothing more. Or any real criticisms of Moore's analysis. Are you claiming his projections are incorrect? If so, what evidence do you have?
Not to mention, the data they have on seeing the big cryptocurrency GPU buys starting to taper off, on top of miners also buying NVIDIA especially the 1070 which has a strong price/performance ratio for certain types of mining.
Worse, AMD's console bread and butter is only getting worse. 5% reduction may actually be optimistic.
And that's on top of Coffee Lake eating their lunch.
https://www.pcgamesn.com/intel-core-i7-8700k-review-benchmar...
https://www.pcworld.com/article/3230369/components-processor...
Again, if Ryzen/Zen is so amazing, where are the sales? All I see are breathless fanboys being especially rude to anyone to calls out Ryzen's flaws or the pros of Intel's offerings. It seems like you guys are giving emotional arguments on what you want to happen in the market and not actually what is happening in the market. You can't will AMD into a proper intel competitor with a strong stock with just angry and misleading social media postings.
I suspect a lot of people are actually looking at video processing performance now that everybody has a 1080p camera in their pocket. I don't do that personally but I recall reading it a lot.
> Again, if Ryzen/Zen is so amazing, where are the sales?
I'm under the impression that Ryzen is selling great. They had to sell Ryzen 7s as Ryzen 5s to meet demand.
Ryzen has a doggy memory controller, as of today, for quite of 2400Mhz RAM sticks, you simply can not install 4 sticks.
Many users are still running chips that is known to segfault when highly parallel workload such as parallel gcc compilation.
Technical praises? Maybe solve the memory compatibility issue first?
Ryzen is a good chip, but it may not be terribly competitive outside of edge cases, specialist builds, or as typical with AMD, budget builds with binned chips with failed cores sold as 3 or 4 core versions.
Now that Coffee Lake is released Intel also got a hexa-core against the Ryzen 5 1600, that one was eating their lunch until now. But even that i5 is not much stronger.
The Ryzen 3s + the smaller Ryzen 5s (their 4 core offerings, 3 core offerings don't exist this time) are competitive against the i3s, though the Coffee Lake i3-8100 wins against the Ryzen 3s at least. No, this is not the strong suit this time, it is the middle class with the Ryzen 5 1600 (in all areas) and productivity with Ryzen 7.
This is exactly why the mobile chips are so promising: This generation the single core performance is very good, energy-wise Ryzen looks nice as well, and the graphics AMD always did better. Raven Ridge APUs should make for very strong laptops.
There are ten-year high levels of shorts on AMD, so there's speculation about how much of the bursts of negative press you see is basically market manipulation to unload a position. It's one of the more volatile stocks today, which would seem to make it a possible candidate for that sort of thing.
I mean, obviously, it stands to reason that the right actor with the right investment portfolio and control of the right media outlets could pull something like this off, but it seems that there are very, very few people/companies in that position.
In fact, the Jeff Bezos / WaPo / Amazon nexus is perhaps the only example I can think of where this sort of thing is even possible, let alone carried out.
Do you have examples? I'd genuinely love to learn.
https://en.wikipedia.org/wiki/Subprime_mortgage_crisis#Inacc...
I think it's easiest to logically consider everything in reverse. Let's assume you're correct, that this sort of thing never happens. So that entails two possibilities. The first:
- No reporters, or individuals with the ability to influence what is written, have vested interests.
I think the above can be pretty easily dismissed as false. But if it's false we can assume at least some people have vested interests. Which leads to the next possibility:
- Those with vested interests choose to never actively utilize their position to further their interests
And again, I think we can pretty safely declare that's false. So it's not a matter of does it happen, but to what scale it happens. And that is going to be pure speculation. Although it would logically follow that the more people that a vested interest in event x, the more common the occurrence of biased media in favor of provoking x would be expected to be seen.
That is not why she went to prison.
> for lying to federal investigators about a stock sale that she called ''a small personal matter.''
The why was most likely insider trading allegations.
The Citron Report's author often finds stocks that have gone up, places shorts, and then promotes the idea that these companies are scams with no fundamental value. He's made a considerable amount of money doing this.
This seems a clear case of bashing .. and if so, it means the article is a compliment to AMD.
This guy runs a blog about companies that he thinks are over-valued, and he shorts their stock. He's transparent about all of this. His readers / investors / the public know that he's likely short these stocks and that he makes a business of both using this investment strategy and blogging about it. They follow his advice at their own risk, with all of the information fully available.
There's a HUGE difference between that and getting into a position that you don't disclose, manipulating a media outlet into running coverage that appears objective but which is actually designed to do nothing but help your position, and then cashing out.
Their product line up is very exciting and only going to get better. Mobile, Server, Workstation, and even still Consumer level CPU/GPU products are all very competitive and priced to sell. The only segment that sells quickly after something comes to market is really enthusiast consumer facing products. It takes time for a product line to build confidence and make a name for itself. Especially after nearly a decade of being the other guys with the mediocre at best option. The last AMD system I owned as a Phenom II X6 and it was really quite nice. I actually owned an Athlon MP (hacked XP chips) all throughout school a decade ago. Jim Keller is a brilliant CPU designer, and he designed Zen.
Granted the last time AMD took a significant market share Intel made many missteps. I don't know if Intel will have another Itanium like failure here.
Personally, I greatly enjoy my 1950X, as it provides just about double the performance per dollar compared to Intel's offerings. But I can see how that assessment would not be true for others, especially gamers and servers.
But just look what the last year brought us in terms of performance and price from both AMD and Intel, after several years of stagnation from Intel, and duds from AMD. AMD revived the CPU market, by providing competition. We need AMD.
So you need either dynamic binary translation, or a processor that can support both x86-64 and whichever ARM instruction sets. Maybe it's possible, but there are two hurdles. Doing that for a gigantic instruction set (x86-64) on a processor with a modest instruction set is going to be challenging. For dynamic binary translation, it's correctness and speed. For supporting two ISAs (somehow), it'd be crazy hard to get the micro-codes right for everything.
And most applications aren't multi-threaded, so you're limited by single-core performance. And that has been x86's bread and butter for decades (also POWER/SPARC I guess), whereas ARM has been power efficiency, multiple cores, specialised cores/silicon. (Edit: This also explains why neither Intel nor ARM has any incentive in making a chip that supports both ISAs, even if it took the dreaded mode-bit)
But if it could be done, then yes, probably patents. IBM brought Transitive Corp, who's tech Apple had licensed for Rosetta (dynamic binary translation PowerPC -> x86, although even that didn't support PowerPC 970). Of course, IBM then did fuck-all with the tech and people of that acquisition, but they still have that IP.
My next desktop system will be a Threadripper for sure. But here's to hoping they can make a move in the mobile market too; would love a multicore X1 carbon.
I should have demanded the dodgy Asus Strix X99 board be replaced at with the first CPU replacement (it was 99.999% likely the Asus' board's fault, not Intel's CPU - even though they got fried), alas now it's too late.
I have weird issues such as not being able to get it to boot every time if Hyper V is enabled, so I have to run Docker using Virtualbox. I'm really quite over the Asus QC and will go e.g. Threadripper + Gigabyte next time.
Here's the Q&A which happened during the call:
Joseph L. Moore - Morgan Stanley & Co. LLC:
Great. Thank you. I was interested in your comments that the sequential growth in the Computing and Graphics business was driven primarily by graphics. How literally should we take that? And I guess, if graphics is up close to $150 million sequentially, is that business now on par with the CPU business? Can you just give us a general sense of the size of the two businesses there in that segment?
Dr. Lisa T. Su - Advanced Micro Devices, Inc. :
Yeah. So overall, the growth in Computing and Graphics, when you look over the past few quarters has been very strong. And we've seen growth both on the Ryzen side, particularly in the desktop side of the business as well as on the graphics side. So in terms of size of the business, again, I think we stated in the prepared remarks that the GPU business had a record quarter for us and we're seeing very strong growth.
We're seeing strong growth as a result of the new product launches. So the Vega product actually did very well for us in the quarter as well as overall Polaris in both gaming and blockchain markets. But yes, we're pleased with the graphics performance. But I'll also say Ryzen did very well in the quarter. We look at the progress that we're making in the desktop channel when you look across retailers and e-tailers across the world. And in the Ryzen 5 and Ryzen 7 segment, we're seeing significant share gain in those parts of the business. So I think both parts of the Computing and Graphics business are doing well. And we continue to expect growth as we go forward.
Joseph L. Moore - Morgan Stanley & Co. LLC:
And is it possible to size the blockchain portion of that?
Dr. Lisa T. Su - Advanced Micro Devices, Inc. :
I think the blockchain tends to be, again, it's hard to separate because it goes through some of the same channels as gaming does. I will say that blockchain sort of behaved as we expected in Q3. So we didn't see anything that we didn't expect. We did see some benefit of channel restocking. So if you look at our channel inventories today compared to July, we had healthier channel inventory levels. And we expect that consumer blockchain will level off a bit as we go into Q4. But there's also commercial blockchain component that we believe is interesting and likely to continue into the medium term. As we look into Q4, though, we also see growth from just the OEM side of the GPU business as we start ramping Vega into OEMs.
So the part in question is this:
> And we expect that consumer blockchain will level off a bit as we go into Q4.
AMD is doing very well and its chips are only going to become more and more competitive with Intel's. That's all you need to know really. The market is irrational, so it's best to ignore it. I don't know why all the tech writers are making such a big deal out of this. They should stick to chip reviews.
You address the 2TB NVMe SSG memory as a filesystem directly from the card itself - no PCIe system transfer involved. 24.6TFLOPS half-precision.
You can program using OpenCL or HIP (opensource cuda replacement )
https://pro.radeon.com/en/product/pro-series/radeon-pro-ssg/.