gnufx

65 ms·

961.
by gnufx 6y ago
Indeed. Infiniband was even meant for "datacentre" use, and I have better experience of it in compute clusters than Ethernet apart from the performance.
962.
by gnufx 6y ago
The node interconnect is crucial. However, things like V100s (and presumably the CORAL2 GPUS) are clearly built for high performance computation, if not necessarily "HPC" in this sense; arguably also the nodes used by Sierra/
963.
by gnufx 6y ago
Well, the sort of technique in https://news.ycombinator.com/item?id=23071459 long pre-dates things like that (but not sophisticated resource managers, of course) before we knew it was so difficult.
964.
by gnufx 6y ago
Well, I can manage a system running hundreds or thousands of heterogeneous stateless nodes consistently from an image essentially simply managed "by hand" (maybe with judicious help from things like package auto-updates). Otherwi
965.
by gnufx 6y ago
And sshd is a fairly common thing to fall over in my experience (though particularly on badly-managed clusters). Presumably more so on otherwise-inaccessible systems, by the laws of life. The agent saved systems isolated in places like x-
966.
New privacy threat combines device identification with biometric information(news.liverpool.ac.uk)
1 points by gnufx 7y ago|0 comments
967.
by gnufx 7y ago
I'm not sure I entirely agree, but Ross Anderson deserves to be read on the topic: https://www.lightbluetouchpaper.org/2020/04/12/contact-traci...
968.
by gnufx 7y ago
GEMM is a fairly special case, but the case is unproven as far as I can see. The OpenBLAS kernels are certainly bigger than they need be, but it's not clear to me they satisfy the request (especially addressing "holistic" tr
969.
by gnufx 7y ago
What server resources to you need for those 40-50 users with Jitsi/Matrix? I haven't seen estimates of required resources when I've looked, but I assume there's some doc somewhere. (Thanks for the good work.)
970.
by gnufx 7y ago
I confess I didn't think it was vulnerable to that. If it is the reason LineageOS dropped it, the drop was a long time coming (2019).
971.
by gnufx 7y ago
It uses ecloud for the portal, defaults to the provider DNS (but you can change that), and treats any behind-the-scenes contact with Google (and others) as a bug to be fixed. I don't think the support range is misleading at all, lookin
972.
by gnufx 7y ago
That can't be right, as e.foundation sell S7 and S9 preloaded. I haven't done it, but flashing instructions are at https://doc.e.foundation/devices/starlte/install for example.
973.
by gnufx 7y ago
https://e.foundation supports more devices (still based on LineageOS, but actively de-Googled). My Nexus 5 is one such. They also sell refurbished phones with /e/ pre-installed (in Europe).
974.
by gnufx 7y ago
Indeed, over 25+ years that I've known, and I'd be surrounded by that in the office, not VNC (but some X2go). The single rubbish X server I've known is (or was) Xceed (or however it's spelt) which people inexplicably use
975.
by gnufx 7y ago
We had effective direct-manipulation GUIs long before OpenGL, so that can't be it, though we did have DMA and graphics buffers in the '80s and DRI wasn't relevant.
976.
by gnufx 7y ago
I didn't notice performance was a problem on something like a SPARCstation 1+, though it was definitely slower than the blindingly-fast GEC4000-driven vector displays we needed to replace.
977.
by gnufx 7y ago
My "hyperTeX" documents under NeWS looked pretty much the same on screen as printed, modulo hyperlinks, but I wouldn't want to write documents directly in PostScript. I'm unconvinced generally about the trades-off of mar
978.
by gnufx 7y ago
The LaserWriter was also, at least in principle, faster than our super-minis at the time, as we only mused.
979.
by gnufx 7y ago
Thanks for all this stuff, some of which I wish I'd been aware of at the time, like LispScript. I don't remember the details, like how it was used, but I recall the metacircular evaluator in HyperNeWS (or do I mean HyperLook?) bei
980.
by gnufx 7y ago
If you can find a library copy, The NeWS Book (e.g. https://www.waterstones.com/book/the-news-book/james-gosling... ) might also be of interest for a static view, but I don't know how much was actually impleme
981.
by gnufx 7y ago
You need recent ssh-keygen's -o to make the passphprase effective (default for ed25519). That was covered here at some stage, but I don't have a reference.
982.
by gnufx 7y ago
Does anyone know whether what was raised in https://news.ycombinator.com/item?id=21419015 was addressed, as the thread didn't get updated?
983.
by gnufx 7y ago
I don't know how omp simd is related to "blocking", but it can slow code by a factor of two compared with GCC optimizing the loop normally, because you get AVX(51)2 but not FMA. (Observed with a generic C GEMM, which got abo
984.
by gnufx 7y ago
For those of us not expert on the standards, could you explain in what way Fortran is weaker, and its rounding control deficient?
985.
by gnufx 7y ago
This "leagues ahead" is simply not true experimentally. I ran the Polyhedron benchmarks on SKX with profile-guided optimization. Pre-release gfortran 10 was (insignificantly) faster on the bottom line than beta ifort (from oneapi
986.
by gnufx 7y ago
There will be cases for which it's not so, but every time recently that I can remember someone saying ifort/icc vectorizes some numerical code profitably, and gfortran doesn't, and I had the code, I got GCC vectorizing it wit
987.
by gnufx 7y ago
The Fortran standard just doesn't talk about "aliasing" (in that context), and "defaulting" suggests you can turn off the storage association rules somehow. I maintain from long research computing experience (at lea
988.
by gnufx 7y ago
I'm well aware of Fortran runtimes. The question was getting at whether the front end is completely independent of the middle(?) end.
989.
by gnufx 7y ago
But it is a nice enough language, at least in its sphere: "ALGOL 60 is alive and well and living in FORTRAN [sic] 90." -- Tony Hoare
990.
by gnufx 7y ago
Every time Fortran is mentioned, this seems to come up. In one thread I referred to the usage figures for the UK "tier 1" system, for instance.

More ›