zlynx

13 ms·

91.
by zlynx 7y ago
I followed every instruction to configure ROCM and OpenCL in Ubuntu 18 with my Vega 56 card. All the samples and such work but Folding@Home refuses to believe its there. It might be nice if someone at FAH actually tried using it on Linux. I
92.
by zlynx 7y ago
> things like: > screen sharing: - Security problem, untrusted access to display, explicitly denied by design. > hotkey daemons - Security problem, untrusted input, denied by design. > window managers - Messing about with window
93.
by zlynx 7y ago
AFAIK, GPU 1 writes directly to GPU 2 via PCIe. Since GPU 2 is most often an Intel iGPU, that happens to also be CPU RAM.
94.
by zlynx 7y ago
No it can't. Which is why I run mine at 98 Hz.
95.
by zlynx 7y ago
I bet the Office software was riddled with C++ undefined behavior. The Alpha was one of the closest CPUs to the mythical Deathstation 9000 which will destroy you if you make even one mistake in your code. Like Itanium, compiling for Alpha w
96.
by zlynx 7y ago
Although Clear does it with prebuilt binary libraries. No need to compile your own.
97.
by zlynx 7y ago
thermald is good for ordinary users I guess. I hate those things though. I disabled all of those thermal and laptop mode management daemons on my Fedora laptop. I ran a CPU stress while manually tweaking the Intel RAPL settings until I foun
98.
by zlynx 7y ago
That is kind of bad Windows programming but easy to do when writing an app that doesn't need to handle Windows event messages. It probably sits in a loop waiting on socket events and doesn't care if you sent it a WM_QUIT or not. I
99.
by zlynx 7y ago
I do a similar thing with my laptop's swap partition. swapoff, add it to btrfs, then remove it and mkswap again. Always seemed safer than a potentially dodgy USB drive.
100.
by zlynx 7y ago
Shared across the entire state of California: "037-ADD brakes randomly for squirrels."
101.
by zlynx 7y ago
Some backup software gets the modification time of the file, does the copy, then checks the modification time again. If it changed then it recopies the file. If it exceeds some number of attempts it gives up and reports an error for the fil
102.
by zlynx 7y ago
There is always "du | sort -nr | head"
103.
by zlynx 7y ago
Heh. That was one of my thoughts before I saw "cargo." "I sure hope this isn't written in Node."
104.
by zlynx 7y ago
My terrible experiences with thin pools makes me see btrfs as a pool of wonderful, trouble-free and perfect code. Just ask yourself what happens when a thin pool runs out of actual, physical disk blocks?
105.
by zlynx 7y ago
TCP blocking is seen the most on mobile data connections. The LTE could be delivering whole megabytes of correct data but if that TCP connection is missing just one packet it must buffer all the rest until it can get that SACK / ACK ba
106.
by zlynx 7y ago
With the current Linux scheduler, yield() means "I don't need to run and won't need to in the future either." and Linux places it at the very END of ALL SCHED_OTHER processes. Even if they are at Nice 20. So it is no sur
107.
by zlynx 7y ago
Just a nitpick: Rust 1.40 seems to be the stable version for a couple of weeks now.
108.
by zlynx 7y ago
I believe it's Selective ACK aka SACK. It allows TCP to ACK ranges of packets so the sender can see that it only needs to resend an earlier block. I've seen rather good speeds over some bad links from my bittorrent server using la
109.
by zlynx 7y ago
Except libraries like the "third_party" example probably can't use parameter packs just yet. They tend to have to keep compatibility with older compilers. I used to work on a library. You wouldn't believe how long some p
110.
by zlynx 7y ago
A fun one relating to 64-bit and NULL in varargs: If your system defines NULL as 0, which is legal in C++, then NULL in the vararg list only pushes an int-sized (4 bytes) zero, and the vararg reader pulls a pointer-sized (8 bytes) value, wh
111.
by zlynx 7y ago
The 1700X and 1800X had some issues with GCC failures under heavy load. Linux apparently uses more CPU resources faster than Windows does, and caused excess voltage drop in some scenarios, as I recall reading about it. I think various sites
112.
by zlynx 7y ago
For web servers it will be CGNAT in reverse. One overloaded IPv4 load balancer trying to keep up with too many servers behind it. But you'll be able to use its IPv6 address directly.
113.
by zlynx 7y ago
I use Linux, Windows and Android. Mailing lists. Then probably promotional. Notifications from things like Facebook, LiveJournal, Patreon, various web forums, etc. Then probably status reports from various scripts. The least amount of email
114.
by zlynx 7y ago
I like Thunderbird and Mutt. If your system can't handle and categorize IMAP folders of 50,000 messages and 500 new emails per day, don't bother on my account. Have good GPG support. Search is important. My cyrus imap server can f
115.
by zlynx 7y ago
I acquired a Surface Pro X. I can't compare it to a 7 but battery life is excellent so far. One thing that my previous Surface Pro 4 did was Connected Standby. It would usually be fine and wake up after an hour asleep just 1 or 2 perce
116.
by zlynx 7y ago
There are problems with worktree. But yes, I often do use multiple checkout directories. When you do the git clone using the --reference option it gets most of the objects from the referenced directories. But this still results in slow comp
117.
by zlynx 7y ago
Beside clean builds is Git branch checkouts. I hit this at least twice a week. Finish a feature branch and push it. Check out master. Pull master. Make new feature branch. Get an urgent bug fix request, switch to release branch, make a bugf
118.
by zlynx 7y ago
I trust the Reds. Although, as someone else just said, it's an anecdote, not real data like the Backblaze report. I started my NAS in 2012 with two 2 TB Red drives. Later I added two 6 TB Red drives. Some time, around 2016 I think, one
119.
by zlynx 7y ago
Depends on the OS. I don't know if they ever went live with their tablet design but I recall reading LKML emails several years ago using / improving movable memory code so they could easily hibernate or drop chunks of RAM, so thei
120.
by zlynx 7y ago
And corruption actually happens. My Dell has an Sk Hynix NVMe drive that apparently doesn't have safe write on power failure. I know because it somehow woke up in my bag and ran the laptop battery down to 0. BTRFS was not happy with th

More ›