Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
jdsully
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
73 ms
·
841.
▲
by
jdsully
7y ago
Small claims is pretty accessible. This isn't exactly lawsuit of the century stuff - their actions defamed and harmed you monetarily which is exactly the sort of issue our tort system was setup to deal with. Note that if it was your o
842.
▲
by
jdsully
7y ago
I don’t see why he couldn’t sue the bank. Someone should try it.
843.
▲
by
jdsully
7y ago
I’m very excited about high DPI support. I switched to xfce on all my Linux boxes because I like the simplicity but the lack of proper high dpi was a nuisance.
844.
▲
by
jdsully
7y ago
They were actually in a great position until they bailed out Elon’s other company Solar City. Its one of the most blatantly unethical deals I’ve ever seen and the company has been in a perilous position since then.
845.
▲
by
jdsully
7y ago
Even with the mesh the number of hops is variable based on which core is requesting and the physical geometry of the chip. The cores right beside the IMC will have the lowest latency. See this diagram: https://en.wikichip.org&#
846.
▲
by
jdsully
7y ago
That is true with Intel chips as well. In the HFT space people actively work with Intel to determine which cores they should pin tasks to. The speed of light is constant, and some cores will always be a little closer to various resources.
847.
▲
by
jdsully
7y ago
See the architecture diagram here: https://www.anandtech.com/show/14694/amd-rome-epyc-2nd-gen/2 Everything goes through the central crossbar on the I/O die, where Zen1 had memory attached directly to eac
848.
▲
by
jdsully
7y ago
The beauty of digital is you are avoiding the “linear” portion of transistors and are always running them saturated. A ternary machine would require running the transistors in the linear region - which ultra small transistors are terrible a
849.
▲
by
jdsully
7y ago
Yup everything is equally slow now. Kinda sad, but the original NUMA design was treated as a glass half empty situation instead of AMD letting people maximize performance. This change lets them avoid the bad press and everyone is happier de
850.
▲
by
jdsully
7y ago
The pay articles are often more timely and higher quality than the free news outlets. It seems only natural they’d get linked to from an aggregator geared to high quality discussion.
851.
▲
by
jdsully
7y ago
The moon is a category II body, with the only requirement being potential contamination must be documented. This is an area that has been covered by international treaties since the late 60s. https://en.wikipedia.org/wiki&#
852.
▲
by
jdsully
7y ago
Why can’t a user make QoS decisions for their own pages? Root controlled pages should obviously have higher priority. The kernel could still “fairly” evict pages across users - just letting them choose which N pages they prefer to go first
853.
▲
by
jdsully
7y ago
Theres no reason extra data cannot be added to entries in the page cache to make smarter decisions. That’s how Windows and OS X do it in their equivalent subsystems. Nobody is suggesting these pages be pinned which is an extreme measure.
854.
▲
by
jdsully
7y ago
> It costs a library just as much to maintain and catalog a digital book as it does a physical book. Do you have any evidence for this? Specifically e-books do not require expensive floor space and employee time restocking shelves. Whi
855.
▲
by
jdsully
7y ago
Can you point to any other species we’ve eradicated that has resulted in a massive human death toll? From my perspective we seem to be pretty good at killing things and getting away with it.
856.
▲
by
jdsully
7y ago
Because we are tantalizingly close to eradicating it. The low death rate is a sign of success not a signal to stop.
857.
▲
by
jdsully
7y ago
Not every Canadian is admissible to the USA, and as a sovereign nation our policy should stand on its own.
858.
▲
by
jdsully
7y ago
> Of course, you're always free to pony up your own cash to pay for a treatment that isn't covered by your insurance With the big caveat that you cannot pay for "higher quality" of any services that are covered even
859.
▲
by
jdsully
7y ago
The funny thing is vinyl is compressed too. The RIAA curve trades bandwidth in less noticeable sections for bandwidth elsewhere where its better used. Thus different frequencies have intentionally different noise floors.
860.
▲
by
jdsully
7y ago
ZipCPU predates the Risc-V buzz and is more a lightweight alternative to things like the microblaze. Its for when you want a microcontroller on your FPGA.
861.
▲
by
jdsully
7y ago
Canada Computers has them in stock frequently. I’ve been able to get two from them.
862.
▲
by
jdsully
7y ago
At this point I’m fairly certain a trade deal with china won’t happen and we’re just being weaned off our dependency on their manufacturing.
863.
▲
by
jdsully
7y ago
Handles are not kernel addresses in NT and windows will not intentionally leak those to user space. Handle accesses are always through a layer of indirection in the object manager. You will find that windows never makes guarantees of what h
864.
▲
by
jdsully
7y ago
I’m not sure why theres such animosity in your replies. The goal here is not to defend windows but to point out that handles need not aways refer to kernel objects. It would be possible to provide an orthogonal handle based API even if wind
865.
▲
by
jdsully
7y ago
You may be surprised to learn that some windows handles are actually just addresses with no kernel state too. A great example is an HMODULE. In fact windows actually reserves a block of user address space that will never be allocated to di
866.
▲
by
jdsully
7y ago
Kqueue is quite nice. BSD is much better designed and I enjoy working with it more. Linux suffers from people doing just enough to solve their specific problem and then no more. The result is APIs are not orthogonal and follow differing sty
867.
▲
by
jdsully
7y ago
While the Unix philosophy is that everything is a file - that is not the case with Linux. Futexs being a good example. There was an attempt to integrate them but it was done incorrectly with inherent race conditions and abandoned. The VMS a
868.
▲
by
jdsully
7y ago
epoll only works with file descriptors and is therefore not equivalent. The whole point is WaitForMultipleObjects works with many different types.
869.
▲
by
jdsully
7y ago
Thats my point. They should just bite the bullet and implement WaitForMultipleObjects instead of having all these disjoint APIs. Plus sometimes you want to wait for both and this is very difficult with Linux. Its also telling Linux has gone
870.
▲
by
jdsully
7y ago
I like how they won’t say the windows API’s name. WaitForMultipleObjects is quite nice - an epoll that works with all sorts of things not just fds. This is kind of a half assed implementation - sometimes I wish Linux would admit Windows has
More ›