HNHacker News
TopNewBestAskShowJobs

the_real_sparky

107 karma · joined September 17, 2017

submissionscomments
the_real_sparky··on Apple has a secret team working on internet satellites
On-board global SMS would probably destroy a good chunk of the SPOT GPS beacon market. The only advantage GPS beacons/emergency locators might have at that point is battery life, and someone like Apple could probably take that part of the market too with a “low power tracking” mode that only fired up briefly every 10 minutes to transmit your current location.
the_real_sparky··on How tracking pixels work
I was highly amused just a few days ago when a 1080p60 YouTube video was not playing correctly in Chrome (dropping frames), but played perfectly fine in Firefox. Resource utilization was not the problem on that machine. I tried restarting chrome, disabling extensions, etc but it seemed to just be some weird bug.

I will also take a moment to rant about the fact that YouTube allows you to select from a bevy of lower resolutions, but doesn’t allow you to select a lower frame rate. Many times I would rather watch a 1080p30 video on slower hardware instead of the 720p60 I am forced into selecting.

the_real_sparky··on California power outage triggers chaos in science labs
True for inverter based generation (most renewables like wind and solar), but generally not for rotating machines like most Co-gen plants. Some rotating machines (steam turbines generally) won’t have governors for frequency control, but usually in a CHP plant they would have a linked combustion turbine that does. Rotating machines that are paralleled with the rest of the grid are by default in sync with all of the other rotating machines if there is not some sort of fault situation at the plant or on the grid. If you try to parallel a generator with the grid that is out of sync (frequency or phase), the grid will always win and pull the generator into sync (potentially destructively for the generator).

The most likely limitation of a typical power plant operating disconnected from the grid would be if it could “black start” or not. Cogen plants would normally have this capability. Essentially the problem is whether you have enough power available from backup generation to bring all of the plant auxiliary systems online and get the generators initially started and producing power. Once started you set the governor to maintain 60.0Hz (if in North America) and begin slowly bringing on load so you can ramp the generation to match it.

The easiest solution to a future blackout would be to match the load in your island area to the output of your plant and then disconnect from the rest of the grid before or at the time the blackout occurs. This way you have avoided the black start situation completely. The blackout detection, disconnection, and load balancing steps can be automated to occur nearly instantly if the system is designed well.

The main issue for Berkeley is that the cogeneration plant doesn’t seem to be sized to accommodate their islanded load. This is especially problematic if you lose a generator while islanded because then you will have instant frequency collapse if you were already on the edge of your plant’s capabilities.

the_real_sparky··on The One Man MMO Project
RoD absorbed an absurd amount of my early teenage life. Wouldn’t trade it for anything.
the_real_sparky··on Scavenging Russia’s Rocket Graveyard Is Dangerous and Profitable
http://hn.premii.com/#/article/17360420
the_real_sparky··on Bitcoin: A Peer-to-Peer Electronic Cash System (2008) [pdf]
> As Bitcoin is practically useless due to high fees and congestion

I disagree here. Gold is very expensive to get into and out of on a percentage basis, also requires careful storage, and is more difficult to exchange (manual pickup/delivery or insured registered mail). And yet it is worth eight trillion dollars.

In addition, believing that Bitcoin is dead for transactional uses because of temporary issues may be a mistake. It is simply software after all, and there are now hundreds of billions of reasons for the developers and users to get it right. My thought is that the alt-coins are simply testing “improvements” to the core technology, and if any updates are sufficiently proven and begin to threaten Bitcoin they will simply be stolen by Bitcoin’s users and eventually co-opt the alt-coin’s value due to Metcalfe’s law.

the_real_sparky··on Bitcoin: A Peer-to-Peer Electronic Cash System (2008) [pdf]
I think that the mistake here is believing that the miners control the network, when most likely the miners simply work for the decentralized network. When the miners become centralized there is little they can do that the users/nodes can’t unwind with a software update, even up to the point of firing those miners completely if necessary (changing the proof of work so that their current hardware is worthless). I think that long-term the incentives are aligned for them to simply do their job.
the_real_sparky··on Bitcoin: A Peer-to-Peer Electronic Cash System (2008) [pdf]
This has turned into an interesting question since some coins now are specifically designed to be resistant to ASICs. The downside for those coins is that botnets or general purpose hardware (that can be purchased or rented for an attack and then resold for other uses later) are a serious threat to the network. A potential benefit for Bitcoin is that the network may actually be more secure with custom ASICS because the people investing millions into developing the custom hardware stand to lose everything if they attack or seriously harm the network. That equipment would be useless for any other purpose, and that same dominating hardware makes the Bitcoin network immune to attack from general purpose hardware.
the_real_sparky··on Bitcoin: A Peer-to-Peer Electronic Cash System (2008) [pdf]
I’d be interested in everyone else’s opinion about the odds that Bitcoin becomes an asset class to rival gold (say at least $2T market cap vs the $8T gold currently is). Right now I’m guessing 20%. Those odds still make the current price a positive expected value (since the 1/5 future has such a handsome payoff), but with a 4/5 chance of failure it wouldn’t be very smart to do this with more than a few percent of net worth. Thoughts?
the_real_sparky··on Micro Scisky: A 32-bit Open-Source Brushed Flight Controller
The brushed designs have been completely outclassed by the new breed of micro brushless quads. The Hoverbot Nano is one of the best examples of lightweight brushless design.

https://hoverbot.io https://www.rcgroups.com/forums/showthread.php?2806620-Hover...

the_real_sparky··on HP Updates Z8 Workstations: Up to 56 Cores, 3 TB RAM, 9 PCIe Slots, 1700W
I just assume that the larger Z8 chassis might handle the heat a little better? The Z6 works for my use, but the CPUs are definitely running on the hot side.
the_real_sparky··on HP Updates Z8 Workstations: Up to 56 Cores, 3 TB RAM, 9 PCIe Slots, 1700W
Electric grid dynamic/transient event (disturbance) simulations. It's essentially modeling the response of electromechanical equipment (generators and large loads), large solid state devices (utility scale renewables and fast-switching reactive devices), lumped load models, and protective devices to events on the electric system. The simulations run in ~1ms timesteps over 10-30 seconds of real world time (which can take 4-20 minutes of CPU time). Each event is independent and they parallelize pretty much linearly. Python multiprocessing is the lifeblood of this work.

Edit: I guess I should be clear and say that the simulations themselves are running on a commercial software package (Fortran based) and I'm simply setting them up, spinning them off, and then processing all of the resulting data in Python.

the_real_sparky··on HP Updates Z8 Workstations: Up to 56 Cores, 3 TB RAM, 9 PCIe Slots, 1700W
Only 8x8GB since I'm working on CPU limited simulation tasks. One of the CPUs and 4 of the DIMMs is on a daughter board in the Z640, which makes the whole thing even more impressive. It's a fairly compact machine. Storage is just a single SSD and single HDD. Graphics is a low end Quadro just to get the 4x mini-DisplayPort. The HP Z8xx series would probably be a better choice if a person is going to go crazy on the rest of the hardware. But that's just conjecture on my part based on the fact that my current processors are running near TDP (but aren't being throttled down based on the monitoring I have done.)
the_real_sparky··on HP Updates Z8 Workstations: Up to 56 Cores, 3 TB RAM, 9 PCIe Slots, 1700W
HP prices are... high. It's easy to spend $10k in processors alone for a dual-Xeon system. The ones mentioned in the article will be over $13k each: https://ark.intel.com/products/120498/Intel-Xeon-Platinum-81...
the_real_sparky··on HP Updates Z8 Workstations: Up to 56 Cores, 3 TB RAM, 9 PCIe Slots, 1700W
I've got an HP Z640 workstation with dual E5-2690 v4's and it is basically silent under full load. It has no problem running all 28 cores at 3.2GHz which is the max all-core speed for that processor.
← PreviousPage 2 of 2