shaklee3

78 ms·

931.
by shaklee3 7y ago
Browsing Reddit is perfectly fine on satellite internet. It's obviously not as fast as cable, but it's not the satellite internet of 10 years ago people judge it on. TMF is probably the most accurate consultant in this industry if
932.
by shaklee3 7y ago
The ground system is what will make or break starlink. Every phased array antenna to date is at a price point that would put starlink selling exclusively to enterprise customers. This means consumers won't be able to get it unless you&
933.
by shaklee3 7y ago
No, it's not. Streaming media is by far the largest consumer of bandwidth on the internet today. Latency doesn't matter at all for those. Please see this blog for approximate cost estimates, or the MIT study comparing leo constell
934.
by shaklee3 7y ago
Thanks. I completely missed that legend. Does it actually scale linearly by cores? I would think there's a plateau after a certain number of cores.
935.
by shaklee3 7y ago
You are referring to the ISL in these satellites, that are not in the first batch, nor are there any plans to change that soon from what I've seen. So without the ISL, their latency is certainly not lower by adding more satellites. The
936.
by shaklee3 7y ago
I'm well aware of the scale of starlink, but the speeds delivered by starlink won't be any faster than the existing GEO satellites at the time, not will it be cheaper. The OP talked about high speed internet, which, in most cases,
937.
by shaklee3 7y ago
Satellite internet has existed for decades. Why is spacex the only way to provide high speed internet?
938.
by shaklee3 7y ago
Building more satellites doesn't reduce latency. Are you referring to the ISL?
939.
by shaklee3 7y ago
To illustrate this, let's assume the GEMM used avx512 perfectly. Roughly that's 2.5e9 (clock) * 16 (ops per cycle) * 2 (FMA per cycle) * 64 (cores). That puts Intel's peak performance over 2TFLOPs, which they're showing.
940.
by shaklee3 7y ago
Something is still weird with the gemm benchmarks. That link you sent has BLIS winning all around, but the absolute value of the BLIS is poor. MKL gets around 2TFLOPS on xeons, and this blis benchmark is only hitting around 30-50GFLOPS. Are
941.
by shaklee3 7y ago
Has anyone seen similar benchmarks, but with BLIS? OpenBLAS is interesting, but BLIS would possibly perform better.
942.
by shaklee3 7y ago
I think it would have been useful if the author benchmarked the actual time taken to parse a large json file, and did a sanity check to make sure the time difference made sense with ipc/clock factored in.
943.
by shaklee3 7y ago
Aphyr is the author of jepsen itself, and I believe all the tests as well.
944.
by shaklee3 7y ago
They're probably speaking about epyc. In general, the availability of epyc, and the number of servers made for it, are very small.
945.
by shaklee3 7y ago
I think they were intentionally behind. None of their products used/supported pcie4, so it was to their advantage to limit their competitors.
946.
by shaklee3 7y ago
Intel had always said they were behind because they wanted to be. In other words, they didn't think 100G was on the rise. I tend to believe them since nic asics are much easier to make than an x86 processor.
947.
by shaklee3 7y ago
FPGAs are in a tough place. Like OP said, most people writing RTL make asics, or at least an asic that's programmable. The FPGA target market is getting slimmer, since we have programmable Asics, like GPUs and tpus, that are as perform
948.
by shaklee3 7y ago
While that's true, this particular company doesn't seem to be targeting anything that would improve containers (OS optimizations, new CPU). So I think the OP was correct in that simply changing the BMC or making it more secure on
949.
by shaklee3 7y ago
It has a fakespot rating of F, for what it's worth: https://www.fakespot.com/product/usb-c-wall-charger-ravpower...
950.
by shaklee3 7y ago
I think AMD needs to take a page from Nvidia and really get their software together. The CPU and GPU libraries are a mess. There are multiple versions of libraries where in some cases both are maintained and it's not clear which to use
951.
by shaklee3 7y ago
Openblas is significantly slower than mkl for most things, and doesn't have all the same features (only the blas part). It's also not developed by AMD. I think they need to take their software more seriously in general, since the
952.
by shaklee3 7y ago
That's absolutely not true. It's the reason movies like Coco are still on Netflix and not Disney plus. You can't just renege on a previous contract because you started your own streaming service. These things take time.
953.
by shaklee3 7y ago
I would say that AMD should make their own mkl if they want to compete. It's fine that it runs on AMD, but it is, after all, a closed-source library made by Intel. At some point amd needs to take the software side seriously. Same goes
954.
by shaklee3 7y ago
Not sure, but it works by storing files on the remote server under ~/.vscode, and this is where all the normal code files are stored. When you connect to a remote system it loads these files over ssh.
955.
by shaklee3 7y ago
See comment below. Just because it wasn't true for you, doesn't mean it's not true. I'm not talking about mocking services to run it; I'm talking about the development environment and sharing with a team of 20 or so
956.
by shaklee3 7y ago
I'm speaking for my individual case and others I work with where we have codebases of over a million lines of c++, with many header-only libraries. On an 80-core server make -j can still take 3-4 minutes, and that uses all the resource
957.
by shaklee3 7y ago
I tried it on the first day it was released and haven't looked back. It's made me significantly more productive than with the old ssh plug-ins.
958.
by shaklee3 7y ago
Outside of web development, sometimes you need special hardware or workspaces configured to do development. Usually this can be done in a container, but that comes with its own annoyances. Having a central development server for large compi
959.
by shaklee3 7y ago
What is missing from vscode for remote development? MS released their remote development extensions earlier this year, and everything from browsing to searching feels native.
960.
by shaklee3 7y ago
I thought wealthfront and the like aren't exactly the same type of insurance. They don't put your money in a single account, and if you ever needed the insurance you have to go through a third party that made a deal with the actua

More ›