Posts by caesium

China's 7nm chip surprise reveals more than Beijing might like

Economically the article's arguments are irrelevant

Let's assume for the sake of argument that China NEVER ever figures out EUV. Furthermore, let us assume that China is banned from importing any processor from any other country forever, and smuggling has been eradicated from existence.

What kind of application could they not achieve with sufficient number of 7nm chips? Well, they'd lose the computer manufacturing market for software pros, which including servers, laptops, and some more niche gizmos like VR sets - but everything else would be fine, including just about all industrial applications and all military applications. These applications haven't even reached 22nm in mass use! Even cloud hosting would be fine - the CPU is mostly abstracted away anyway, so they'll just use more processors. Even smartphones may be fine (most users already have enough computing power for their uses), and frankly, most light computer users can do with 7nm (I'm writing this comment on a 7nm). So basically we're talking about very limited economic damage if there were an import ban, which there isn't.

The proper takeaway is that for every practical purpose China has the chips it needs, even if it's not the smallest and newest node. Instead of chips, the West should concentrate on arms to match PLA growth.

Checkmate: DeepMind's AlphaZero AI clobbered rival chess app on non-level playing, er, board

Re: Bitter people are bitter...

"Stockfish won’t benefit from additional hardware."

Of course it would, especially given the small time limit. Deeper search translates to stronger play with alpha-beta all else being equal. It might not have mattered significantly if the games were using a standard time limit.

"Stockfish being configured not using an opening book is to compare engine to engine strength"

Which is incompatible with how this was advertised. If one of the ideas is to compare machine-learning with human tuning, than we need to use all of the relevant strategy's resources, not ban a certain type of human tuning (which Stockfish was designed to play with) and than declare victory...

"jumping through the hoops that the waste-of-space detracters invent."

No hoops here, just use standard tournament rules, vanilla SF configuration and publish _all_ of the games. Or else we'll conclude Google spiced up their (already interesting) work.

SQL Server 2017: What's new, what's missing on Linux, and what's next?

Re: Price point with Oracle? Think lower...a lot lower!

MS probably expects to profit from big Oracle installations which may change to SQL Server. These people can pay. If the choice is between earning 5,000,000$ for install from 1,000 users or 1,000$ from 50,000 (numbers are random), obviously the first is far more profitable for MS.

Besides, it would be difficult to compete on a different basis on Linux than Windows, while keeping feature parity - or do you expect SQL Server to compete on features on Windows, while on Linux to competes on price?

Re: Does its own memory management and threading??

SQL databases - at least high performing one - are always like that. A kernel simply doesn't know enough about the database, so it would tend to make dumb decisions like employing read-ahead when you wanted to select only one record, trying to cache parts of the database in memory competing with the database's cache, giving priority to the wrong queries based on 'fairness' (assume for simplicity each query has its own 'thread') and so on.

A database knows so much more about current and future use so it needs to override much of the libc and kernel's default handling (open all files directly, use light-weight 'threads' and so on). Eventually the functions are abstracted to utility functions and from there to a layer between an OS and the database. SQLOS is probably a misnomer - AFAIK SQL Server doesn't use a kernel module even on Windows.

Mud sticks: Microsoft, Windows 10 and reputational damage

Re: The author misunderstands Microsoft completely

"generate worse code"

"Older tools will continue to work in exactly the same way .."

You are just rephrasing what I said. Older tools won't get added support for new processor features, meaning they generate worse code *than newer tools*. And lack of support for features can and does make development harder, unless an #ifdef mess is your idea of good code.

"And yet 'backwards compatibility' is often claimed as one of Windows best features"

It used to be. Then Microsoft changed direction.

"it requires changes to suit the ever changing Microsoft OS, such as rewriting everything..."

Well, yes, it does. But other software houses have the issue as Microsoft: competing with their own older versions. They may feel rewriting isn't all that bad if they get to sell more.

"List some if you can."

Already did. The BSDs (Frree/Net/Open/Dragonfly) and Apple. GCC and binutils did not drop support - rather the projects refused to upgrade versions due to licensing considerations. Just trawl through the mailing lists and see what issues it causes.

Re: The author misunderstands Microsoft completely

Microsoft's moves will generate a reaction. They already do, judging by the way Microsoft representatives are talking about $PRODUCTs in open and closed meetings, and also by that Microsoft is making sure to support hybrid options (at least for now).

But the overall result can be nonetheless profitable to MS. Apple is indeed a good example. They all but patented the 'forced upgrade' treadmill, yet their software business is still very profitable and they maintain a loyal fanbase. Microsoft could well end up similarly successful financially or even more so with the low road. I hope more users will know they have options too...

Re: The author misunderstands Microsoft completely

"Do you think that the old tools stop working?"

Sure, one can keep using older tools. But this introduces other problems. Older tools won't support new language constructs (say, latest C++), may generate worse code and will not build code for latest Microsoft OS forcing one to rely on compatibility (or making sure that the source can be built with several versions of build tools). And often there's a need to maintain an entire stack to keep the older tools working. Then again, it is necessary if only to be able to test the result. Of course, this all is far from impossible, just has its costs. And if enough developers don't pay these costs, than there's a pull for newer versions of Windows.

Working with old tools can be a problem in non-Microsoft lands too. Just take a look at those projects which refused to upgrade GCC and assorted tools due to not accepting FSF's GPLv3 (Mac, BSDs, etc.). They all ended up signing on LLVM/clang (or relenting on GCC) because relying only on an old GCC version was introducing serious issues.

Re: The author misunderstands Microsoft completely

As I point out in my reply to Mr. Plinston, quite a few software houses have as much motive as Microsoft to sign on the upgrade treadmill. After all, they are competing with their own apps too.

These software houses are too small to build a cloud for themselves, but if they join in with Microsoft, they can use its OS changes as an excuse to sell new versions. And if enough such applications exist, then users will be forced to upgrade.

One must note also that developing for old OS is becoming harder and harder. For example, new development tools simply do not support Windows XP or Vista.

P.S. Speaking of the cloud, here's the title of another story in the sidebar:

"Amazon WorkSpaces two years on: Are we ready for cloud-hosted Windows desktops?". There's of course not much real reason to be 'ready' for it. But it will make a lot of money for MS and whichever OEM will host the cloud. And if MS agrees to not doing the hosting themselves, OEMs will have an out path and could choose to keep collaborating with Microsoft... Dell Cloud, anyone?

Re: The author misunderstands Microsoft completely

Your comment (and the one by ecofeco) are perfectly correct that this is a bad deal for users (well, at least Desktop users running non-Thin clients - which are the majority of Microsoft's users). Most of us would have been willing to keeping running Windows 7+ version for a good time. However, Microsoft, being a commercial company, is committed to its bottom line, not us, the users. I hope I was clear about Microsoft choosing its own interest over its customers' interest. Will MS's scheme work? IMHO, there's a pretty good chance of that.

* The point about users still using old software is exactly what I was making. However, it takes only a critical amount of new software which won't run on old OSs to make a case for switching.

* It would be smart to design one's software to run on many OSs including many versions of Windows.

But not all developers will do that. Some will be tempted by Microsoft's various shop offerings, which don't offer content for W7 etc. Development tools will support older versions less and less. Developers are encouraged to upgrade their own OSs with cheap offerings (e.g. MSDN) - at which point some will be too lazy to test on old OSs. And other software shops have the same logic and motives as Microsoft to sign on the upgrade treadmill.

* OEM are screwed here. But where will they go? Margins in the PC biz are razor thin anyway. Most alas don't have the courage to risk signing on Linux.

* The customers are screwed too. Duh, that's the entire point. I am sure MS will offshore a part of the OS to the cloud too (can't offershore everything - too many angry enterprise customers). There'll be some excuses here (some backup service for example) and it will be free at first. Sooner or later a paid premium service will be offered...

The author misunderstands Microsoft completely

I agree with the author that the W10 UI is still unpolished. Yet, he misunderstands Microsoft's position, leading to (IMHO) completely wrong conclusions on how this came about.

Simply put, Microsoft is no longer in the PC-selling business. MS changed UI on windows 8 to a UI which is much less suited to desktops. *This was done entirely on purpose* (more accurately, the new UI being worse on desktop was an acceptable downside). The choice of new UI had nothing to do with user input, and everything to do with Microsoft's bottom line (and if we follow the stock reports, we can see it was a very successful choice).

The old MS model was exactly as the author thinks it was: Sell Windows in order to sell PCs which sell Windows (and of course, other related MS products like Office/Exchange/etc.). Yet that model has an obvious downside: Microsoft was essentially competing with itself. Every time a new version is out, a consumer is asking himself whether to upgrade. Since the old version worked fine, and will work fine as long as his PC 'lives', he may well choose to stick with it for a few years - leading MS to lose lots of money.

How does one solve this problem? The standard way is to innovate. But innovation is risky, costs money, and each time Microsoft has to do it all over, all in order to compete with itself. Instead, MS chose a different path, with two long-term gains, at the cost of short-term pain. MS can do that since it has lots of money (this is exactly what they did with XBox. Lose money on the short-term in order to try to gain long-term). The path is:

* First, one answer to competing with your own products is to create intentional incompatibilities between versions, so that eventually users have to upgrade in order to run new software. An app written to the new W8 UI will hardly run fine on W7, and eventually, users will have to upgrade. Profit! But this would have allowed any new UI. The new UI was chosen because:

* The other answer is to move your products to the cloud, at which point, the company can start charging based on (for example) use-time, size-storage, etc. and stop caring about competing with its older versions. We can see Microsoft's future with Azure, Office 365 and other products, all of which have a very similar UI.

At the time Metro was decided, it was known that all these cloud products would need a UI, and that there's no practical way to make the Web emulate Aero. So instead, Microsoft's choice was to design a web UI (which is what Metro is really suited for). Then change the desktop UI to train users towards Microsoft's de facto Web UI. Sure, Microsoft will have worse Windows sales, but this hardly threaten their marketshare. And in training users to switch towards its cloud products, it gains a much more reliable revenue stream, which would be very difficult for regulators to break. Also, making the UI match removes a de facto competitor to its web offerings.