Happy new year and good night!
I'm going to bed. Feels as though I haven't slept all year.
Happy new year and good night!
I'm going to bed. Feels as though I haven't slept all year.
First day of the year...and the coldest day of the year, fuck me right?
Time...line? Time isn't made out of lines. It is made out of circles. That is why clocks are round. ~ Caboose
@evn thoughts on the BIGGGGGGGGGGGGG omg Intel bug?
Whilst it may be unfair to call it the "Intel bug", there is talks even amongst non-tinfoily people on Sweclockers suggesting that Intel and other manufacturers have known about this for at least 20 years and intentionally opted not to plug it to eke out performance. I would honestly not be surprised, lex AMD microstutter.
I don't know whether I myself credit this much, but I believe it's not something they are unfamiliar with, even though this suggestion is probably blowing it up more than the truth.
People are going crazy about this in all the wrong ways... Typically it's fun to bash big companies doing a gigantic fuck up but the people's reactions both defending and attacking Intel aren't rational. I'm not going to claim to know everything about how it works and how the patch affects certain kinds of load performance but still... it's getting kind of ridiculous. Testing irrelevant things or just the usual forum screamers is a bit much.
I really hope POWER9 becomes popular, the work IBM put into it seems amazing.
https://www.olcf.ornl.gov/summit/
https://en.wikipedia.org/wiki/POWER9
https://www.hpcwire.com/2017/12/06/i...ng-doe-google/
https://cloudplatform.googleblog.com...BM-POWER9.html
https://www.theregister.co.uk/2016/0...summit_power9/
And it seems RISC-V is finally getting some action too!
https://fuse.wikichip.org/news/686/e...isc-v-monster/
https://www.esperanto.ai/
I wonder how long it'll take for RISC-V SoCs targeting embedded systems to start appearing. Modems, routers, cameras, TVs, etc. There are multiple things they can make with fairly simple SoCs, then we'll hopefully see it taking on ARM servers too.
Me and my friend were thinking, How fast could you make a computer if you did not have to worry about it's security??
Time...line? Time isn't made out of lines. It is made out of circles. That is why clocks are round. ~ Caboose
Time...line? Time isn't made out of lines. It is made out of circles. That is why clocks are round. ~ Caboose
9ms ping, 180.29Mbps down, 6.09 Mbps up
Faster than 81% of US
Comcast (ugh)
I wonder how SpeedTest calculates the "faster than" percent. I'm sure my upload speed is weighing me down. It would also explain why @Butler Log would be faster than 94% of DE with just 88 Mbps down.
Odd thing is, I signed up (a couple years ago, now) for the 75 Mbps package and haven't changed anything. I did get a notification they were increasing speeds (about one year ago), but it was only to 100 or 120 Mbps. Comcast being generous? That doesn't compute... Also, I've never seen 180 Mbps down in the wild. Highest I've seen is from Steam at around 120.
@apepi - Could that just be noise from one run?
It might be line saturation with apepi, too. Some companies might still assign several households to roughly the same boxy thingy words.
As for calculating "faster than", I did a few runs back in... 2012 on speed test. I think. It's in the "how fast(..)" mega-thread at least. You can use the in-thread search function to find my posts and see.
Upload is very much factored in, as can be seen in tests taken the same day, with both download over 90 (97 and 92?), but upload 80 and 88, and that the higher upload with lower download got a higher score, comparatively.
@apepi
I'll have to upload my college's numbers next time we have a LAN party. :P
Last time (Nov 2017) it was like 220 down, 240 up. Insanity.
Yeah, the nice thing about QAM is that you can always increase data throughput by increasing the number of bits your symbols represent, with 256-QAM being part of the standard since 802.11ac the only worrying thing for normal consumers about wifi is the fact that some locations are full of different networks using the same frequencies which means you can't really use all the time-slots. The 5GHz band is still hugely underused, there are way too many old 2.4GHz only routers being used.
At 5/6 coding rate, a 80MHz channel and a 256-QAM modulation, you can already transfer at ~1.7 Gbps. On top of that you can still go slightly out of the spec with 1024-QAM (broadcom for example supports it, should be in the spec with 802.11ax) and aggregate multiple channels. You can also achieve ridiculous throughput (7 Gpbs) with 802.11ad without using extensions to the spec, since your carriers go at the 60Hz band and you can have higher bandwidth per channel there (the standard uses 2.16GHz wide channels).
Sincerely speaking wifi stopped being an issue in most situations, I think it's still a good idea to run cables when you need/want the reliability and predictability of running isolated networks, but for most people there's not really a difference unless you're running your wifi in bad conditions to begin with.
I hate you all. I live in the country having shitty internet, I am excited with the possibility of even having the potential of being able to broadcast 720p.
Time...line? Time isn't made out of lines. It is made out of circles. That is why clocks are round. ~ Caboose
Are broadcast requirements different than watching a stream? I could watch 720p videos from Youtube with my previous ISP on slow as shit DSL (5.5 down / <1 up). If you're in the 25+ range, then I'd think you'd be able to stream 1080p, honestly. You didn't specify the fps, either, which would have an effect, too.
To think the FCC wants to drop the standards for high speed broadband from 25 Mbps to 10 Mbps... *shakes head*