No it’s not. I was stating that I don’t understand how you came to the (surprising) conclusion you did.
You could have just explained, but you chose instead to throw a defensive tantrum. OK. Whatever.
I work in the processor industry, though not for Intel or AMD. The Intel goal the last I heard it is for under 100 failures per million during system assembly. I’m sure AMD is the same. So I’m betting the parts aren’t bad.
However, there are so many configurations, I think you hit a subtle bug that got turned on by your configuration and almost no one else’s. Processors ship with known bugs, that they usually work around in firmware on the system or else they hope no one notices.
I heard a paper from Dell about how they load the software on the systems they build because that is the only way they can test the systems they ship. There are so many millions of possible configurations that making sure they all work is impossible.
Welcome to the wonderful world of system test, which is a black art.
It’s highly unlikely some new processor would just come from nowhere to compete with Intel in the high end desktop space. It’s not like opening up a new restaurant or law agency, fabs alone cost billions of dollars to build, the research involved would likely run into the hundreds of billions not to mention about a decade worth of research. And that’s only to get to where Intel is now while Intel is 10 years more advanced at that point. Not to mention, 9 years into your 10 year journey, Intel can just drop their processors down to non-price gougy levels and wait until you run out of money and fail before raising them again. There’s no way a pricing opportunity would encourage a new player to enter into the market.
That being said, Intel’s biggest threat right now isn’t AMD, it’s ARM. Intel’s terrified of ARM commoditizing them from the low end and they’re throwing money left and right to claw incremental percentages of marketshare in the tablet and phone space. If Intel had some magic secret developments waiting patiently in the lab, it would have been deployed by now.
That’s my point exactly.
Some other manufacturer would tweak an existing processor to compete in the desktop (non-Windows) space. IBM or ARM perhaps? The whole landscape would change, but it wouldn’t mean $5,000 video cards…
That’s true, but it’s a drop in the bucket in terms of total CPUs. The Xbox 360 sold around 80 million units in a few years. The new one is in single digits. The Playstation 4 has sold 22 million.
Annual smartphone sales are in excess of 1 billion. And they’re basically all ARM.
Why would anyone want to compete in a shrinking market? Remember, Apple went for Intel also. And Intel is both not opening new fabs and also selling capacity in existing fabs.
ARM has two big advantages. The first is that it has a solid grip on the mobile space. The second is that it is an IP, not a manufacturing, company, and so doesn’t have to worry about building fabs.
I believe AMD isn’t building fabs wither, but is buying fab space. They used to buy it from IBM, I don’t know about now.
Microsoft said in March that they had shipped 10 million XB1 units. That’s not “sold” but they shipped with processors inside (I hope!) so that’s around 32 million AMD processors going out the door. The total number of console sales will probably start going up as publishers taper off production for the last generation – Ubisoft has said it’s done developing for 360/PS3 aside from some niche title (some dance game?).
I wasn’t trying to compare them to ARM, although my using your quote probably gave that impression. I was just saying that there was more to the processor/GPU market than desktops and laptops.
yeah, this is true, and in more ways than one. Intel is basically giving away Atom-based SoCs, and for the life of me I can’t figure out what the value proposition would be for hardware manufacturers. the one thing that keeps x86 going in the desktop/laptop space is the decades of Windows software legacy that people want to keep using. Apart from Windows tablets, there’s no such legacy to speak of, so I don’t get why anyone would choose an Intel SoC over any of the countless (commodity or bespoke) ARM SoCs on the market. Especially when the Intel SoCs have to be on a newer, more expensive fab process node just to run in the middle of the pack compared to what’s out there on the ARM side. x86 has no benefit whatsoever in the handset and tablet market, and ARM is continually working its way up the performance ladder, as shown by the exceptional Apple A7 and A8.
AMD used to own two fabs, one in Austin (25) and one in Germany (36.) They let 25 go when they spun off their flash manufacturing as Spansion, and spun off 36 which became Global Foundries.
So when HASN’T Intel dominated AMD in processor sales? Even at AMD’s best, Intel still had something like 2/3 of the desktop and server markets, and for the most part, the Intel processors still beat the AMD ones performance-wise within a category. And the IT saying that “Nobody ever lost their job for buying Intel” explains why just about every corporate environment I’ve been in was 100% Intel (and I suspect they also account for the majority of sales as well).
AMD was always the value brand, and sometimes had better overclocking capabilities, but at no point did they outsell Intel, at least that I’m aware of. And the aforementioned growing irrelevance of CPU power and growth of mobiles and tablets means that AMD’s niche has gotten smaller, and they don’t have anywhere to go.
GPU-wise, it was closer, but hasn’t Nvidia always been a little out ahead of ATI/AMD? I don’t remember it being as perpetually lopsided sales-wise as the CPU market, but I always got the impression that they were either dead-even or Nvidia was slightly ahead.
I’ve bought AMD for a long time because of price, momentum, and no complaints about performance. Another reason I like AMD is because they tell you the number of cores right in the name of the processor. For example “AMD Phenom II X6”, I know it’s a 6-core processor because of the X6. Contrast that with Intel, names like “Intel Core i7-2600K” or “Intel Core i5-2500K” tell me nothing about the number of cores (no they aren’t 7 and 5 cores), which is the first thing I want to know when CPU shopping. Nitpicky I know, but just the name “Core i7” annoys me for some reason.
My latest machine is both AMD processor and AMD GPU, and I have so far been able to play every game I have purchased on max settings, with no issues. And I certainly haven’t purchased top of the line, either- I suppose my hardware won’t last for many more generations of new games, but for now I am pleased with both the price I paid and the performance it has given me.
a big part of AMD’s problem was that a lot of that “value” came at the cost of performance or stability. The K5 was terribly slow compared to the Pentium. The SS7 CPUs (K6/-2/-3) were much better, but they relied on third party platforms from e.g. VIA and ALi which were utter garbage compared to Intel’s chipset families. Even AMD’s own chipsets had crippling bugs like the inability to use AGP 2x with many GPUs, or the PCI bus corruption issue between the AMD750 and VIA southbridge, or the AMD760MP’s broken USB controllers, etc. They basically only got their shit together with the K8 chipsets.
at no point in its existence did AMD even possess the ability to outsell Intel. Intel has always had a lot more production capacity than AMD.
Massive bribes essentially. The race right now is to see whether Intel can get more power efficient faster than ARM can get faster. But Intel loses the war immediately if legacy apps get compiled only for the ARM instruction set. By keeping enough x86 chips in the market, it forces OS makers to keep their compiler cross-platform features acceptably bug free and for app makers to fix any x86 only bugs in their apps.
Intel doesn’t care if it bleeds cash in the short run as long as it keeps them viable in the tablet/phone space in the long run.