Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
SoapSeller
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
14 ms
·
91.
▲
Estonia e-Residency is now open for applications
(apply.e-estonia.com)
6 points
by
SoapSeller
11y ago
|
2 comments
92.
▲
by
SoapSeller
11y ago
Hmm, It's 9$ for the bare board, no hdni/vga. The charge 15 for the adapter which should cost much less to produce. Also, it's look like allwinner's version of broadcom's rpi, so they mught get the various chips dir
93.
▲
Mystery person puts shuttered music site Grooveshark back online
(arstechnica.com)
3 points
by
SoapSeller
11y ago
|
0 comments
94.
▲
by
SoapSeller
12y ago
First, for youtube, right-click -> Stats for nerds. Watch the "Dropped frames" counter, with hw acceleration it usually stays close to zero("-"). For youtube&everything-else, you got two major tools: 1. CPU usage,
95.
▲
by
SoapSeller
12y ago
The missing phrase in the post is "hardware acceleration", which is is completely broken for VP8/9(In Firefox & Chrome, IE&Safari doesn't support either). I'm using an extension[0] to force youtube to serve
96.
▲
by
SoapSeller
12y ago
This project(at least for me) is a little bit misleading. Because it isn't rendering HTML using WebGL. It using html2canvas[0] to rasterize the browser rendering into a canvas, upload it to WebGL, apply effects and patch-back events(i.
97.
▲
by
SoapSeller
12y ago
There are some efforts out there to do full JPEG decoding in CUDA[0], and hybrid approaches[1]. Actually, AMD supports JPEG decoding via OpenCL implementation in their drivers[2] and Intel has native hardware decoding in their latest GPUs[3
98.
▲
by
SoapSeller
12y ago
I did, with the AKG K1000[0](They are very uncomfortable, btw). The problem with soundstage is that most recordings are made for speakers. Binaural recordings + headphones results in much better and much more realistic soundstage. With the
99.
▲
by
SoapSeller
12y ago
OpenMP is terrible. While you can argue whether introducing pragmas as control structure is a good idea or not, the performance issues with high load(>80% total cpu) and the caveats(two OpenMP versions in same app, multiple threads runni
100.
▲
by
SoapSeller
12y ago
Previous discussion: https://news.ycombinator.com/item?id=4755288
101.
▲
Powerful, highly stealthy Linux trojan may have infected victims for years
(arstechnica.com)
273 points
by
SoapSeller
12y ago
|
99 comments
102.
▲
A machine that turns Coca Cola back into water
(dezeen.com)
3 points
by
SoapSeller
12y ago
|
0 comments
103.
▲
by
SoapSeller
12y ago
Yandex and Baidu.
104.
▲
by
SoapSeller
12y ago
From the "Start" page: If you didn't type it correctly, you might not see anything. If this happens, make sure that you've copied the example code exactly This is definitely a big no-no if your goal is "to ma
105.
▲
by
SoapSeller
12y ago
It's somewhat sad that AMD can no longer compete head-to-head with Intel so they have to produce processors based on ARM designs. I'm really hoping that someone will compete with Intel on the high-end CPUs, maybe IBM's PowerP
106.
▲
Cable companies: Wer'e afraid Netflix will demand payment from ISPs
(arstechnica.com)
3 points
by
SoapSeller
12y ago
|
0 comments
107.
▲
OnePlus One review: Big, fast, cheap, open, efficient, and unlocked
(arstechnica.com)
5 points
by
SoapSeller
12y ago
|
0 comments
108.
▲
by
SoapSeller
12y ago
30-day active users is somewhat problematic for a service like Drive. I am accessing(reviewing) on weekly basis some docs that other people have uploaded to Drive and shared with me. So I'm considered "active" user while I do
109.
▲
by
SoapSeller
12y ago
Yep, ICC's output is much bigger: clang 29K gcc 36K icc 89K
110.
▲
by
SoapSeller
12y ago
I've used the same flags like the OP have suggested in the source header comment.
111.
▲
by
SoapSeller
12y ago
Not for this example. On E5-2680 v2(Ivy Bridge): ./magic_forest_gcc 617 655 606 7.22s user 0.36s system 100% cpu 7.580 total ./magic_forest_icc 617 655 606 7.00s user 0.34s system 100% cpu 7.340 total ./magic_forest
112.
▲
California company Aerofex accepting preorders for $85,000 hoverbikes
(arstechnica.com)
4 points
by
SoapSeller
12y ago
|
0 comments
113.
▲
Amazon’s latest patent is sillier than the peanut butter sandwich patent
(arstechnica.com)
4 points
by
SoapSeller
12y ago
|
0 comments
114.
▲
by
SoapSeller
12y ago
I'm using ASUS VivoTab Note 8 for a couple weeks now and I'm very happy with it. It has proper(IIRC Wacom) digitizer and the Atom processor is is handling OneNote touch + web browsing quite well. The only let down is screen reso
115.
▲
A rescue plan that might have saved space shuttle Columbia
(arstechnica.com)
252 points
by
SoapSeller
13y ago
|
152 comments
116.
▲
by
SoapSeller
13y ago
Didn't Google won the the "Oracle vs Google" case on the ground that APIs aren't patentable? What prevent MS(or any other big player) to re-implement GMS?
117.
▲
by
SoapSeller
13y ago
When I interviewed with Amazon a few years ago, almost all of their recruiter emails ended up in Gmail's Spam folder. Clicking on the "Not Spam" didn't seem to fix this, and I ended whitelisting amazon.com. At the time I
118.
▲
by
SoapSeller
13y ago
"for Ubuntu"? really?
119.
▲
by
SoapSeller
13y ago
This is the most annoying hn post I've encountered. Ever. Every single link here(post & comments) redirect me(on my android phone) to the Play store listing for their "Wait" app, that doesn't contain any useful infor
120.
▲
by
SoapSeller
13y ago
JIF is already taken by JPEG[0], so GIF, with a hard G. [0] http://en.wikipedia.org/wiki/JPEG_Interchange_Format#JPEG_fi...
More ›