snaky

82 ms·

841.
by snaky 8y ago
> getting sponsored for adding features (doesn't seem to scale all that well) And that's a problem.
842.
by snaky 8y ago
That's a rules of the game. Every department want to grow its importance in the company, to get more power and resources. Well, sometimes this end up in failing company, but they have no choice - if dep X decide to stop it, other won&#
843.
by snaky 8y ago
That is and will stay impossible no matter how many resources for simulation you have, actually. > This technique, when applied to the real world, is sometimes useful, but can sometimes also lead to self-deception. This technique is call
844.
by snaky 8y ago
The state of Firefox and Chrome shows more than anything else that the many cores is good thing and especially that RAM size is the king actually. Modern smartphones with 6 and 8 GB of RAM make a difference.
845.
by snaky 8y ago
The logical way forward is then to actually just hiring a candidate as a short term contractor for a week or two to solve real work problems on site, in a team with other employees and see how that goes .
846.
by snaky 8y ago
Okay, I should have clarify the 'niche' thing 10 years ago was not GPU per se, but of course GPGPU, the GPU as a specialized coprocessor for general purpose computation, not restricted to graphics processing and output.
847.
by snaky 8y ago
New Snapdragon 845-based Chromebook is on the way. > It comes as Microsoft continues its work with Qualcomm to optimize Windows for devices powered by Qualcomm's Snapdragon chips, including the forthcoming Snapdragon 850, which Sams
848.
by snaky 8y ago
Yes, but speedup can be huge. For example, for NFA processing, using (a couple of) 2W FPGA against 200W GPU "GPUs underperform FPGA by a factor ~80-1000x across datasets" http://people.cs.vt.edu/~xdyu/html&#x2
849.
by snaky 8y ago
The two huge companies being merged is more often considered fail than otherwise. > a 2004 study by Bain & Company found that 70 percent of mergers failed to increase shareholder value. More recently, a 2007 study by Hay Group and th
850.
by snaky 8y ago
There are some interesting ideas about programmability, like automagically offloading some computation parts to FPGA. http://conal.net/blog/posts/haskell-to-hardware-via-cccs http://conal.net/blog&
851.
by snaky 8y ago
> The problem is to end up at "noindex site" thinking that the "noindex site" is a competitor site And there is no solution for that, it seems. The solution for the problem <red in address line>'hey, that&#
852.
by snaky 8y ago
If there was a spinlock, the problem stay. > the rule simply is that you MUST NOT release and immediately re-acquire the same spinlock on the same core, because as far as other cores are concerned, that's basically the same as never
853.
by snaky 8y ago
Not at all. Specialized coprocessors are mainstream actually - from network cards offloading to small chip in every iPhone, and for a reason. The obvious way forward is universal specialized coprocessors, reprogrammable for the task(s). Bet
854.
by snaky 8y ago
Back button is not the only way to end up at noindex site.
855.
by snaky 8y ago
> Clearly, they don't have an answer to AMD at all They still have a huge opportunity for CPU+FPGA, they bought Altera for the purpose.
856.
by snaky 8y ago
> noindex So, the browser extension indicating (with big red fonts) that this site is noindex could be a simplest solution? For not power users who don't know about any extensions that would be not so easy though. If that function w
857.
by snaky 8y ago
> the lock was being acquired and released ~49,000 times and was held for, on average, less than one ms at a time. But for some reason, even though the lock was released 49,000 times the Chrome process was never able to acquire it. Well,
858.
by snaky 8y ago
> cmov's can make a difference. You can use them to avoid polluting your branch prediction tables, you can use them to make code smaller, and you can use them when they simply just fit the problem really well. It's just _not_ t
859.
by snaky 8y ago
And not without a reason. https://yarchive.net/comp/carry_bit.html
860.
by snaky 8y ago
Actually this may be enough for modern IoT sensors, adding capacitor and doing transmission less often helps. Even flashing the LED (5mA pulses for 1ms) is only 5uA on average (once a second, 0.1% duty cycle) or even 1uA (once a 5 second, 0
861.
by snaky 8y ago
You are comparing 2016 China numbers with 2018 Netflix numbers. Talking about paying users only, maybe Netflix is still on the top - not for long, of course. > According to research firm IHS Markit, video streaming in China will more t
862.
by snaky 8y ago
Rockchip has hardware Hi10p support since RK3229 IIRC.
863.
by snaky 8y ago
On the US market. Which is not only market in existence, nor even biggest one. Take China for example - where neither Netflix nor YouTube are present. > Compared with the US, China still has plenty of room to grow in a country with a pop
864.
by snaky 8y ago
By the way, there is a project to "Implement a front-end for CakeML that accepts OCaml or Haskell syntax" > An involved version of this project would be to write a new verified parser for an alternative syntax. A simpler versio
865.
by snaky 8y ago
That's doubtful that high-assurance Haskell is in much demand actually. People usually prefer much more strict (in both senses) languages for that, CakeML in particular comes to mind.
866.
by snaky 8y ago
That's a shame we have triggers as the only tool for this job, because e.g. CHECK constraint "cannot contain subqueries nor refer to variables other than columns of the current row"[1]. Triggers are very low-level thing actua
867.
by snaky 8y ago
Integrity (consistency) was the biggest point of relational database theory itself. > The true importance of relational theory is highlighted by the title of the original (and considerably shorter) version of Codd’s first paper. That ver
868.
by snaky 8y ago
That's 80-20 economy which is growing more and more, going hand in hand with "economy of scale". If tens of millions are less than 20 per cent of users - so be it, nobody cares. What does matter is 80 remaining per cent.
869.
by snaky 8y ago
Quote from Linus, 2003 >><rant> >>Why do the file systems have to be so tightly integrated in the "ring0" core? This is one subsystem that screams for standard callouts and "ring1" level. >><&#
870.
by snaky 8y ago
If it pops up on AOSP.

More ›