Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. ×

Comment Re:Any happy CDN CloudFlare client around? (Score 1) 87

The main advantages are that a) they can take a lot more load that the majority of sites can by both doing pretty good caching and having a lot of geographic redundancy (and the DNS services to handle that), and b) have an operations team that can better respond to DOS attacks than most of their customers are large enough to have (plus network-geographicly distributed resources to hopefully mitigate the attack).

They are probably not going to be faster for small-traffic websites, and they are adding a layer of complexity (so can easily be complicit in bugs, if not outright responsible for them), but if you are going to be hit with a lot of traffic (legitimate or not), then they offer a service that few smaller companies are in the position to match.

Comment Re: Simple Answer (Score 2) 86

"tweak" is a bit of an understatement. They are on their 5th shipping version of modified ISAs (Swift, Cyclone, Typhoon, Twister, and now Hurricane), and the included PowerVR GPU has been increasingly modified from the base technology from Imagination Technology. Where most "tweaking" is in how many cores or what fixed-function units are included, Apple has been playing with the core instruction set to make them more performant (both from power and speed perspectives). This has been how Apple has been at least a year ahead in meaningful performance for at least 4 years now (multi-thread performance is not usually meaningful on a phone), despite having a lower base clock speed than their competitors (thus getting very nice battery savings out of it).

What this article is talking about is that Apple is spending increasing amounts of money directly in R&D, rather than farming it out to their suppliers (which does not count in R&D).

Comment Re:What is the R&D Actually For? (Score 1) 86

The memory limitations that you cite, as well as the driving problem behind slow updates, can be squarely put at the feet of Intel. They have pushed back meaningful updates for a couple of years now. I am not implying that they are doing so deliberately, but rather have been unable to make meaningful upgrades.

To take the memory size limitation, that is because Kaby Lake processors are the only ones to support 64GiB, and the models that Apple would have used were not available (let alone in Apple quantities) until long after the current MacBook Pros shipped. And I think you are a bit mislead about "battery issues". For most workloads Apple's newest MacBook Pro's have 10+ hour battery life. The only place where it is not better than the previous generation is on the 15 inch models on workloads that cause the GPU to kick in. There the battery simply is not enough to really feed that power-hungry GPU. This was an engineering decision (tradeoffs between a better GPU, thinness, and battery life for certain workloads), and real deserves a more thorough understanding than your summary indicates.

A similar conversation applies on the MacPro front. Again, the Xeon processors that Apple used have not been upgraded in a way that justified updates. I wish that Apple had released speed-bumps along the way (and adjusted the bottom-end price along the way), but there was really not enough change since their release to justify a re-work since then. A GPU update might have been nice, but (full disclosure: I worked at Apple, and helped test one aspect of the GPUs) Apple spent a lot of engineer time making those custom GPUs sing on the workloads they were for: FinalCut Pro (not gaming). Likely someone crunched the numbers on sales and determined that it was not worth the expenditure to do that again for a mid-term product. Whenever it is updated again we will see if Apple goes the custom route again.

I also don't think you are evaluating Apple chip work nearly well enough. I you look at the CPU/GPU work they have done on the iOS devices; for the last 4 or so years competitors have been at least a year behind on most real-work testing metrics. Only in multi-threaded tests does anyone else remain competitive within a year timeframe. That is despite Apple being lower-power and lower clock-speed in almost all cases. And the delta has been widening as Apple ramps up on this. They started with nearly off-the-shelf processors, but are now on their fifth version of increasingly modified ARM ISA (Swift, Cyclone, Typhoon, Twister, and now Hurricane), each of which had increasingly custom versions of the paired PowerVR GPUs.

None of that work is about lowering costs, all of it is about improving performance. If Apple only wanted to lower costs they would be using Samsung or Qualcomm licensed CPU designs.

Comment Re: This is not surprising (Score 1) 245

And yet you don't seem to be able to point to a single lie. All of this innuendo, and no facts. If she really did "lie continuously for months", then there would be a clear record of it. At this point it is clear that your position is not about the truth; you have an enemy and you are going to do everything to damage that enemy, even if you have nothing to go on. That is simply prejudice.

Comment Re:Slower than MongoDB, has joins (Score 5, Informative) 21

As a former RethinkDB employee I am more than a little biased, but I don't think that you understand the competitive space around MongoDB. Everything you have sited as an advantage for MongoDB is done better by just about every one of their competitors (RethinkDB included). MongoDB's main advantage is that they were the first big on in the field, and no-one has been able to make something better enough to de-seat them. It is not enough to be better, you have to be noticeably better in order to de-seat a reigning competitor. Think of the phrase "no one gets fired for buying IBM".

And I also don't think you understand the cost of polling, especially for non-trivial (e.g.: not key-lookup) queries. While RethinkDB's `join` queries are not included in `changefeeds`, just about everything else is. So for example if you wanted to keep a leaderboard, say the top 10 scores in a game, you would have to re-compute that every time in most databases (at a minimum scan the index). With RethinkDB it automatically gets modified based on writes in the database, and sent to you. The efficiency improvement is truly huge. And since those queries can be fairly complicated (say: top 10 scores within the week), that gets very expensive with polling.

An example that is in usage right now from a major stock trader: their iOS app uses RethinkDB to get streaming stock-price updates. The app (indirectly through a server) just opens a changefeed on the list of stocks that you follow, and RethinkDB coordinates who needs to get what updates when they feed in the stream of changes of market prices. They don't have a ton of clients constantly polling in order to show them constantly changing feeds of numbers (some change every second, others not in hours), and they can push out changes as fast as they get them.

Slashdot Top Deals

Practical people would be more practical if they would take a little more time for dreaming. -- J. P. McEvoy

Working...