Comments Locked

37 Comments

Back to Article

  • silverblue - Tuesday, May 2, 2017 - link

    A couple of mistakes... :)
    AMD Q1 2017 Financial Results (GAAP) table: Revenue for Q4'2016 shows $1110B.

    AMD Q1 2017 Financial Results (Non-GAAP) table: second column header shows Q1'2016 and Revenue is missing the M (very minor point!); third column header shows Q4'2015.
  • siddharta - Tuesday, May 2, 2017 - link

    Yes and I do hope they have more than $943 in cash :).
  • Ryan Smith - Tuesday, May 2, 2017 - link

    Like I said, Q4 was very good to AMD.
  • siddharta - Tuesday, May 2, 2017 - link

    Not if they spent $1.26B to be left with $934 :). Thanks for fixing that Ryan.
  • Eden-K121D - Tuesday, May 2, 2017 - link

    I really hope that Vega is not a paper launch
  • edlee - Tuesday, May 2, 2017 - link

    AMD needs to send a game optimizing team to every major publisher, and get them to tune their games for more cores/threads, and you will see ryzen gaming benchmarks in line with intel. They already beat intel in the synthetics, they just need to optimize more for gaming, and it will be selling like hot cakes
  • trparky - Tuesday, May 2, 2017 - link

    Not as long as there are memory compatibility issues, BIOS bugs, and general stability issues. It really comes down to the fact that Ryzen was half-baked at release. If you're looking to upgrade today, go Intel; you won't be sorry. If you can wait a year, see what Ryzen v2.0 is like.
  • EasyListening - Tuesday, May 2, 2017 - link

    lol, so much FUD. Ryzen is doing fine.
  • TheJian - Wednesday, May 3, 2017 - link

    Fine? ROFL. A full quarter of your cpu comeback in sales and you can't break even? Margins up only 2%? Inventory blowing up? What kind of problems are you having here?

    trparky was spot on. They should have waited another month or two for motherboards to NOT be ruining the story at reviews all over the web. That casts doubt on your cpu and ends with lost sales. Your new cpu loses in every game (is there a game it wins in? Close enough to call it ALL IMHO), and we're already discussing gen2 IN reviews from the cpu maker themselves. Multiple company people basically saying well gen1 is only for work stuff, the fixed one with better gaming is coming shortly (16 core x399/X390 chipset, maybe 12core as well). That is not a great way to launch. The games were OK as a problem alone perhaps, but most reviewers complained of board issues (granted bios pretty much, but it's embarrassing) and that made a LOT of sour reviews rather than showing the chips as good as they could have been only a month or so later with FINISHED bios files and better tuned boards. Most of the reviews mentioned how picky memory was also. Again, that's down to another month of testing and proper bios work getting done, but it still adds up to what could have been decent reviews (without those issues) to instead end with, "well, crap, that wasn't what we expected".

    Stock holders took notice too, to the tune of a $3.60 haircut (25%!) on the stock (not good on a $13.50 stock). If you can't make money, I don't care what you say, you've got issues and you had better figure them out quick. Ryzen/amd great comeback amounted to a stock crash. Granted I think it will go back up as server/vega stuff come and hopefully a gen2 desktop (game fixed version) as soon as possible (better be before xmas), but this Q should have been much better than it was. A full quarter of sales of the new cpu and margins suck and no profit either. OUCH. The inventory is worrying since it would seem reviews DID hurt sales and that isn't good for the life of gen1, which is why I say gen2 can't come soon enough. IF the 16 core desktop version that is supposedly fixed works, they need to transition to it ASAP across the full line of desktops. You can't fix the CCX (L3 cache issues) with windows patches etc so games for gen1 are pretty much as is. Game devs are not going to help unless paid (AMD is already short on cash), so better to just fix the tech yourself with gen2 if you can. Most game devs will just figure the hit in their perf is not on enough desktops to be worth fixing (meaning not enough zen chips sold).
  • Tamz_msc - Wednesday, May 3, 2017 - link

    Games, games, games - all you can think of is gaming performance. The ecosystem at launch was nowhere as bad as it was made out to be, considering it was primarily one motherboard manufacturer's whining that set the ball rolling.

    Ryzen has barely been in the market for 1.5 months at the time of this report.
  • mapesdhs - Thursday, May 4, 2017 - link

    Check TheJian's past posts, he's more aware of most than non-gaming issues (I find his replies are some of the ones actually worth reading).

    Fact is, reviews did make a lot of the poor showing of Gen1 in gaming, and no amount of AMD or anyone else saying well it's not really meant for that (yet) is going to bring people back who've already decided to buy Intel instead. Nuanced factors such as having more than high enough frame rates anyway, or it being a non-issue at higher resolutions, are lost on most people; a great many people don't read these articles in full, they jump straight to the conclusions page, and there in many reviews one will find, ok so it's good for pro tasks, producitivity and media, but bummer about the gaming! Oh the gaming! They do ham it up a bit.

    He's also absolutely right about the mbd/BIOS issues, it's why I decided not to buy a Ryzen system yet (I was on the verge of getting a 1700X, but got tired of trying to obtain info about which mbds actually existed, vendors were listing models that did not exist on manufacturer web sites, and forums were full of moaning about BIOS problems, especially RAM, the very thing that would help present sensible performance).

    Lastly, AMD should have been open from the beginning about how the use of fast memory is important for Gen1 performance, though given the launch timing that might have highlighted the mbd/BIOS issues even more. Better that though than reviewers finding out on their own about CCX functionality, etc., as it all makes entry level and midrange RAM look horrible for Ryzen, which increases the cost of a system that has RAM fast enough to properly exploit what the chip can do.

    All of which makes me wonder, why did AMD launch the chip when they did? What was on the cards in terms of competition that meant their waiting another month or two couldn't be tolerated?
    Or was it more about just getting the thing out to help the perception of their financial position? Beats me.

    Ian.
  • nirolf - Wednesday, May 3, 2017 - link

    Full quarter you say? Have you read the article?

    "The Ryzen 7 launch in turn isn’t meant to turn AMD’s fortunes around overnight – especially when it was only on retail shelves for one month of the quarter"
  • zodiacfml - Thursday, May 4, 2017 - link

    Yep, I saw the news of the stock price drop. I think it is probably due to the leaked Vega benchmark and the sales of the Ryzen might be settling down after the influx of early adopters who bought it.
  • Timoo - Thursday, May 11, 2017 - link

    "Fine? ROFL. A full quarter of your cpu comeback in sales and you can't break even? Margins up only 2%? Inventory blowing up? What kind of problems are you having here?"

    The thing you are missing here, is that Ryzen 7 only hit the market at the beginning of March. This means that sales only count for 1/3 of the period and revenues/cash flows for even less (30-day payment delays). Ryzen 5 only got launched on the 11th of April, so those revenues are not even reflected in these numbers.

    So, if Ryzen impacts this balance sheet, it is only for their fair share: 30% of the period and only related to Ryzen 7.
  • edlee - Tuesday, May 2, 2017 - link

    I agree that launch motherboards should have been tested and tuned better, but I would not exactly call Ryzen half baked at luanch. Most of the teething issues have been addressed. if I was building a new gaming rig, i would go Ryzen 5 1600 + Gigabyte G3 B350 + G.Skill Trident 16gb DDR4 3200mhz + nvidia GTX 1080

    Im still using an i7-3770k OC 4.9GHZ, but looking for a change this summer, when Vega hits
  • trparky - Tuesday, May 2, 2017 - link

    What do you call the fact that motherboard manufacturers didn't have working BIOS for their boards? OK, they had "working" BIOS versions but they were buggier than a roach motel. I'd call that half-baked.

    Then let's not forget that IPC benchmark numbers are less than stellar, they only meet the equivalent of a two year old Intel chip (Haswell). For you, an owner of a 3770k, and me, an owner of a 3570k, would only see marginal improvements in single-threaded tasks. Sure, you'd see a huge boost in multi-threaded tasks but for single-threaded tasks there's really no improvement to speak of. AMD kind of screwed the pooch here.

    Ryzen v2.0 seems like it's not going to be that much of an improvement over v1.0 since they are projecting only a 5% boost in IPC which is pretty pathetic since an Intel Kaby Lake chip wipes the floor with Ryzen in single-threaded tasks.

    Face it, Ryzen didn't live up to the hype, that's all there is to it and it shows in how AMD's stock took a nose dive. Someone needs to buy AMD right now or they're done. Perhaps Samsung can buy them and infuse some serious amounts of cash into AMD's R&D budget.
  • Meteor2 - Tuesday, May 2, 2017 - link

    Ryzen may not have lived up to 'hype' but it matched what AMD promised.
  • Alexvrb - Tuesday, May 2, 2017 - link

    The main issue at launch was low memory clocks. This has largely been resolved via microcode in the form of AGESA updates (which have been integrated by mainboard vendors into their BIOS releases with pretty good efficiency), but they're continuing to tune things and will be releasing more updates. Other than that, there's an extremely rare FMA3 bug (Intel has experienced similar bugs recently too) that probably will never occur in real-world usage, but they're fixing that in a microcode update as well.

    Was there any real stability concern that you wanted to mention, or just felt like beating up on AMD with a sideways and deceptive "general stability" comment?
  • prisonerX - Wednesday, May 3, 2017 - link

    Looks like Intel is paying Russian troll armies to try to smear AMD now.

    Постарайся, товарищ!
  • prisonerX - Wednesday, May 3, 2017 - link

    Yeah buy Intel, pay top dollar for an i7 that is 60% IGP, what a deal!
  • Amoro - Tuesday, May 2, 2017 - link

    don't look at their stock right now...it's nose-diving.
  • edlee - Tuesday, May 2, 2017 - link

    it should have never reached 14 or 15 a share in the first place, it should have stayed 11-12, they are on the right path, and with current market growth they should be more valuable in a year or two, stupid investors were pushing their stock up where they should have never been in the first place
  • Shadowmaster625 - Tuesday, May 2, 2017 - link

    lol. RIP
  • fanofanand - Tuesday, May 2, 2017 - link

    It's called profit-taking and is quite normal. This *gasp* even occurs with Apple, Google (er, alphabet) etc.
  • mapesdhs - Thursday, May 4, 2017 - link

    Yup, ride the hype train, help it climb, jump ship when it starts to fall. Not stupid at all, all perfectly normal; stupid would just be those who bought thinking it was not going to go down.
  • versesuvius - Tuesday, May 2, 2017 - link

    By the time AMD manages to release Vega there will not be many users left who have not upgraded their graphic cards. And the past experience with AMD shows that the cards will not be widely available even for those who want to upgrade. Naples is another story of course. It is best for AMD to put all its efforts and money in bringing many many core Naples chips to market. Vega can wait until the next cycle or even skipped altogether.
  • Meteor2 - Tuesday, May 2, 2017 - link

    ?. Vega is launching this month. I don't recall much in the way of supply problems for Polaris GPUs.
  • versesuvius - Tuesday, May 2, 2017 - link

    There were problems. Unlike Nvidia where the market is flooded with cards from every manufacturer immediately after launch, AMD cards were only available from a very few with limited availability. Polaris is a good card and did fairly well. A dual GPU RX 480 never materialized. It could give 1070 a good competition and even take some sales of 1080 away from Nvidia. Perhaps Vega is only as good as a dual GPU RX 480, which is not good enough. Let's hope that is not the case though.
  • Alexvrb - Tuesday, May 2, 2017 - link

    Dual GPU cards are extremely niche, to say the least. While I personally would have liked to see something like that, I find it hilarious to see you mention "limited availability" and then advocate an oddball card like that.

    Anyway, for the most part there were minimal supply issues for Polaris past initial high demand for the reference design. I don't suspect they'll have any problems with Vega at this point, beyond the initial demand spike. It's not even a new process for them.
  • versesuvius - Wednesday, May 3, 2017 - link

    "I find it hilarious to see you mention "limited availability" and then advocate an oddball card like that."

    Of course I advocate it but maybe limited availability of the chips was one reason that AMD did not make a dual GPU RX 480? It would be ridiculous to have a card where only a few thousands of it could be made. And more importantly it would not make good business sense to have a card from previous half generation that competes with your flagship card after its launch. Anyway the closer it gets to the launch of Vega I become more convinced that it sits somewhere between 1080 and 1070 with an asking price of around %10 more than a 1070. Let's hope that AMD does not make the same mistake that it made with Fury and Nano.
  • Meteor2 - Tuesday, May 2, 2017 - link

    Right, how can a company have a positive margin, but a net loss? What costs aren't they booking against product sold?
  • ckbryant - Wednesday, May 3, 2017 - link

    Margin's can be positive and still make a loss; the products margins are defined as either gross margin or net margin. The main idea of margin is the (Net Profit - Cost)/Net Profit. It doesn't account for overhead or anything else when you categorize things into different market segments like the GAAP numbers. They seperate out Computing, Graphics, Custom IP and it's possible for a company with high ASP and high 35% margins on products to not meet profitability. Even in restaurants if you aren't making 50% margins you probably aren't making any money by the time you factor in overhead and etc; when I managed a restaurant we preferred something termed "keystone" margins which is about 250% on food; and due to waste and etc there were times when 212% margins and something like a fridge or freezer going out would prevent profitability. Margins are the best frame of a companies profitability; Debt to Income structure is a little bit better. I'm a tad bullish on AMD right now; they can't go much further down in terms of their Server CPU sales theres room to go since ARM hasn't been the huge leap in the Server market that investors and etc thought.
  • ckbryant - Wednesday, May 3, 2017 - link

    aren't the best frame*** no edit
  • mapesdhs - Thursday, May 4, 2017 - link

    (yeah, bummer about the edit thing; come on Anand, we're practically into the where's-my-jetpack era and still no edit function??)

    In general, anything less than a 100% margin in business is a bad idea.

    Talked to a lady recently who'd visited a function event at a place that had it's own cafeteria, they were charging 8.50 UKP for a baked potato & cheese. :D
  • harobikes333 - Tuesday, May 2, 2017 - link

    Still Bullish on AMD. Their Vega should turn out pretty decent & the Server Market share should increase. Purchased on their nosedive today... I believe they'll upright the ship in the coming couple months.
  • mapesdhs - Thursday, May 4, 2017 - link

    Should, should, believe. Hope is not a plan or a rationale.

    What you ought to be saying is that AMD should learn from Ryzen, make sure Vega doesn't have any equivalent launch issues, ensure adequate supply, double check its power consumption, heat and noise levels don't suck, make sure it behaves properly and for dear god's sake make sure the drivers don't suck (the main reason I've avoided AMD cards in my own systems; I have some anyway for benching, always more issues, especially with CF, and I don't think DX9/CF was ever fixed despite repeated promises).

    I'm not saying you'll be wrong about a turnaround, but atm such a belief cannot be based on anything more than hope and optimism.

    I'd want to see something better than just pretty decent though if it's to pull sales away from NVIDIA. If AMD did want to make a big splash, it wouldn't be that hard, just make sure the cards are reliable and have strong FP64 or small int for compute and deep learning, areas where NVIDIA have ditched their old style strong showing entirely.

    Ian.
  • mapesdhs - Thursday, May 4, 2017 - link

    (by the latter I meant what used to be potent performance in cards like the 580, original Titan, etc., and even the 10x0 have greatly restricted small int)

Log in

Don't have an account? Sign up now