R is Such a Fun Letter

I might identify it as a letter, but I am actually referring to the R statistics environment, something I taught myself to use a bit more than a year ago, and continue to put find new ways to use it and learn more.
This past week I did a number of things with it, including putting together some complex graphs. Had to fight with it to get what I wanted, but once I had the results, and threw in some more data, I think I see a pattern that is intriguing, at least to me. Still have a lot more data to collect, but what I have at least serves to develop the scripts I will use.
Something else I figured out how to do just today makes one step of processing the heart-rate data I collect easier. For the Body on Games videos, I keep the recorded data and upload the CSVs for people to view, if they are interested. To help look at the data, I make a column that shows the time in the video the specific measurement is for (helps the measurements are made every second). Have I have been doing this involves loading up Excel, typing in a couple entries, and dragging the pattern down. I figured there had to be a way to do it in R, which I already use to generate a graph and histogram data, and again I had to fight some, but ultimately ended up with a script to do the job for me. It is just so much more convenient this way. The reason I had to fight with R is because of the, to me, convoluted way it handles time, always requiring a data. All I wanted was a hh:mm:ss sequence starting from zero and counting up by one second, but that is not something it wants to do. I figured it out though and am glad I did if only because I enjoy finding ways to move away from Office and Office-like software to more powerful, if more complicated/hard to use, FOSS solutions.

Awesome Opportunity Taken

I've been busy the past few weeks, and the primary work consuming my time has been working on a couple reviews, and one of them has gone live! I was very fortunate that I was able to receive on loan some Lenovo hardware to review, including the Lenovo Explorer, the company's Windows Mixed Reality headset. It has been a pleasure to work with it, in part because of how easy it is to get working. There is no special set up to speak of, and in all of the SteamVR games I have tried it in, it just worked and worked well. It leaves me looking forward to the WMR platform continuing to develop, including receiving formal SteamVR support, as it is currently in beta.
So cool, but I have to get back to some other work now, and post something on the Body on Games blog.

Finally Upgraded!

The final parts arrived and now I am finally rocking a new build featuring an AMD 1950X CPU, so I have 16 cores/32 threads running at 4 GHz. It is combined with 32 GB of DDR4 3200 MHz CL14 RAM and is running so nicely. Video processing is much faster and the games I've played thus far have also improved nicely. It is that video processing I wanted so much though (why I went for the Ryzen Threadripper over just a Ryzen 7) and it is really awesome to have a system capable of over 2x encoding speed at x264 Medium, compared to about 0.1x before. I am very much looking forward to putting it to some serious work soon, but right now, I have a few other things to take care of first (like this post).

Comment on Threadripper

Just days ago a video was put out showing that the four dies in an AMD Threadripper CPU are in fact all Zeppelin dies, but two are disabled (very possibly defective dies). Since then I have seen a number of people commenting on how they wonder if this means we could see a 32-core Threadripper in the future, by just having these two dies be active. Well, from what I understand of the Zeppelin die and the Infinity Fabric that ties them together in both Threadripper and Epyc, this cannot happen and will not happen.
As we know from the single-die Ryzen processors, the Zeppelin dies are complete onto themselves, critically with their own memory controllers. The controllers are designed for dual channel, but by combining them with Infinity Fabric we can get quad-channel (Threadripper) and eight-channel (Epyc) memory support. This is also why you would not see a Threadripper with more than two dies (or less then two dies) active because the platform has no support for six-channel or eight-channel memory. Sure, in theory Infinity Fabric could allow another die's memory controller to be used, resulting in higher latency for both dies involved. It would require an entirely different platform that will provide access to RAM for these extra dies to be usable, and I cannot imagine AMD would want to introduce a fourth desktop platform. Currently we have Ryzen 3/5/7, single-die platform, Threadripper, two-die platform, and Epyc, four-die platform. Crossing products between these makes no sense and adding another platform to that already healthy mix is also nonsensical.
So, chances are the only way to see an even higher core count Threadripper is to also see a higher core count die (which means a higher core count Core Complex/CCX). That will certainly come, but I would think it would not be until Zen2 and the 7 nm shift. At least, that is my reasoning on it.

Not my Preferred Price, but I Got One

I didn't get it at launch, but I was lucky and caught some Gigabyte Vega 64s on up for order Amazon Friday morning at the MSRP for the Black pack. That means instead of the $499 price of just the card I paid $599 for the card with two games. Well, one of the games I definitely want, so in effect I'm only paying $40 more than I would have preferred (I want Wolfenstein II which will be $60 for a time).
Now, I do want to comment on the Vega pricing story, which personally I am seeing more as a non-story that a few actors have put forward and the Internet is running with. The idea is that AMD gave an MSRP of $499 for Vega 64, but had only little intent to sell many at that price point, so 'in effect' the actual MSRP is $599. Basically people are upset there were so few at the $499 price, so instead of being rationale and patient they are attacking AMD and every second AMD does not respond to their liking is another reason to despise the company. I am very glad I interpret the world differently than this as I'd quickly exhaust my emotions. Here's my take on it, based on some of the available facts:
The card-only and Radeon Pack versions of Vega 64 have different SKUs, meaning they are supplied differently and possibly also ordered in bulk differently. (Not an expert on SKUs but they would be ordered differently, it is more I am not sure how much the wholesale prices differ.)  So a company may have gotten 100 total Vega 64 GPUs, but only 10 of the card-only SKU and 90 for the more expensive Radeon Packs. This means you have a lower chance of getting the cheaper SKU and better chance of getting the more expensive SKU (which happened to me, but I'm not upset about it). AMD has issued an official statement saying they are trying to restock all the SKUs, so with patience you will see the card-only SKU come back in stock, but they might be bought up quickly. Meanwhile, there are a lot of people, including crypto-currency miners trying to buy them up as well, which is making some of the gaming community super-salty, because it means less product for them, which drives up costs.
Now for my, I think rational take on this. AMD had stated the Radeon Packs were meant to counter the hordes of miners trying to purchase up the cards. Personally I don't think the packs were very interesting (most people will probably have the CPU and motherboard already, and a monitor too while games we already expect to be free), but I want to look at the intent here because that is much simpler to consider (I do not want to guess about the internal politics to cut the deals necessary to arrange the deals between the different manufacturers). AMD likely made a calculation that assumed miners just want the card-only version while gamers would be more accepting of the more expensive packs that give them something extra (I personally would have rather saved the $100 and purchased the games separately, but I'm not complaining because I do want one of the two games, and will probably enjoy both. Plus, I chose to purchase now instead of waiting and hoping for a better offer in potentially months) and that calculation skewed how many GPUs went to the SKUs. If they already wanted to target gamers over miners, then they would put more into the Radeon Packs.
One's gut might say AMD was wrong about balancing the SKUs, but let's add a dash of reality to this. They were going to sell out fast, across the board, and once that happened the prices would go up. This has already been happening with GPUs and it was going to happen here. It didn't matter what the balance was, this would happen. AMD, according to my guess, was hoping the extra $100 would dissuade enough miners to leave some for gamers, and while it may have worked for some, there are two things that worked against them. One was rumors that Vega was going to be amazing at mining (looks like this isn't true, but then not all of the hardware features are currently driver-enabled either, so we'll see what happens, plus AMD has long had a powerful compute side, so there would be interest anyway) and the other is the prices were going to skyrocket regardless. New tech almost always sells-out immediately because it is new and exciting and in limited quantity. Those who want it are going to get it, whether it is a miner or a gamer like me. I have a specific reason for wanting it beyond just gaming, which NVIDIA GPUs could satisfy just as well (I want it for comparisons in game performance analyses and to do some technical analysis). When someone has a specific goal in mind, higher costs are less an impediment. Consider game prices and how they start expensive, when the people who want to play 'that' game are buying it, but once this group is satisfied, the target shifts to those who want to play games like that, but there is more competition for this group's money so prices have to fall. Vega is new and definitely in that first phase. Mining has disrupted GPU prices in part because it extends the length of that phase significantly, so prices cannot come down for those who just want a good GPU, and not a specific GPU. NVIDIA's less compute-capable GeForce cards receive an advantage here then, because they are just not so good at non-gaming tasks, and the smaller, gaming-focused architecture allows it to be cheaper (and use less power, as some have been attacking AMD for) hitting AMD twice.
Now, here's my guess for the future: the card-only SKU is going to be restocked and sell out almost immediately. This will continue for months, and during those same months the Vega drivers will improve and people will come to realize it is a better architecture than the attacks suggest. People also attacked Ryzen when it launched, and now it is widely seen as a tremendous success for several reasons, so with time will come some clarity. Now those new drivers might make them more interesting to miners too, so that issue will not be solved, but something else is also going to happen. Technically two things will happen, with the public one being the Radeon Packs coming to an end. Some are decrying AMD, claiming that when this happens the card-only SKU will go up in price, but considering this would be stupid economic move and at best a rumor, I suggest we dismiss that concern unless it happens. When that happens, chances are the card-only SKU will keep its current price, but also the manufacturing cost of the GPUs will have come down, so there will be more profit at that MSRP too. Basically, if the Radeon Packs are making more money for AMD, they might be giving them a stop-gap measure to bring in more revenue while the manufacturing process improves, making the lower price more affordable to them. Might take longer than the packs will last, but the packs could change before leaving, and what profit comes in might carry the losses as the price drops.
Be patient. Things will improve. (Also, don't start attacking the unreleased Navi architecture, claiming the same will happen there. We don't know and at least I had a silly idea that would ease this issue if true. In addition to the ability to mass produce one die cheaply for an entire stack, like we see with the Ryzen CPUs through EPYC, my crazy and uninformed idea is that maybe Navi will be able to have additional compute components attached to it. This would mean there could be graphics focused and compute focused versions, allowing AMD to have the different classes of GPUs, like NVIDIA does currently, but without having to produce a lot of different dies. It is a crazy idea, but may be it is possible.)
Huh, that's a long one... sorry about that, but seriously, stop freaking out and attacking AMD. Be patient and see what happens. Chances are it will be less malicious than the hate requires and the harm from the hate will be regrettable.

Ryzen is Coming!

Not here though. Ryzen is coming to the world, but not my house. Can't afford it. Ah well, besides the later I build a new computer, the later the stepping and more refined the process. Good combination for if you want better performance.
Anyway, so much of the news and rumors I have been seeing about it have been very positive, which I am very glad to see. AMD directly competing with Intel again, in multiple segments is a very good thing in many ways. Hopefully Vega too will directly compete with NVIDIA very well in all segments too. Sadly though, because of how the world, and some marketing departments work, superior products in performance, feature, and price can still lose because of misinformation, manipulation, and fanboy-ism. I think AMD will have an easier time challenging Intel than NVIDIA for that last point, but only time will tell.
One idea I had though, but I cannot predict when or if it might happen, is to do some Ryzen Return videos or something on (one of) my YouTube channel(s). The idea is that because some games are a bit much for my A10-5800K (but it still does very well) that if/when I have a more powerful Ryzen CPU to translate the higher performance to videos and share what the difference is like. Not all games would be worth doing this with, but it still could be a fun idea. Plus it would give me a good reason to return to some games. (Any reason is a good reason to replay a good game, right?)
For those of you who are fortunate enough to get a Ryzen CPU, enjoy it. I hope I will be able to join you before too long.

A Little Late and Maybe a Little Political

A little late, but I remembered! What I've decided to write about might be considered political, but I'm not sure. Really it's just going to be me calling a proposal I recently heard about from Bill Gates stupid.
This proposal is to apply a tax on the use of robots used for automation in businesses. According to him this tax should be collected by governments so that it can be directed to programs that would help the elderly and at-need youth. Sounds nice, right? Plenty of things sound nice that end up doing harm and are stupid.
Okay, so let's change the wording a little bit: The proposal is to artificially increase the cost of labor, for certain types of businesses (manufacturing would make greater use of robots than, say, software development). This would also artificially make human hires more competitive, at least until cheaper robots are made until the tax can be made, then the robots win and there is a heavy political battle to not increase the tax, as small companies operating near their margin would be pushed into the red. Increasing costs doesn't hit the big and wealthy, it hits the small and medium. (Decreasing costs artificially also hit the small and medium, once the market corrects, which is what happened in 2009, so that is not a solution.)
That's the problem with trying to artificially increase costs of a type of product, is that eventually someone will find a way to make the product cheaper, and then you're left trying to increase the cost more or even stifle innovation.
Also, a source of cheap labor, like robots, is not really a bad thing. Yes some people will lose jobs, others will move into other positions, managing the robots, even managing new factories. So not every human would be hurt by this. Additionally, some will be able to take advantage of the cheap labor to start their own businesses using it, and assuming they have friends who lost their job, they could be brought in to work as well. Increase the cost of the robot labor, and this cannot happen.
It's the whole, innovation drives innovation concept that has been proven throughout history, naturally combined with human tenacity and desire for pride in work. Stifling it with artificial cost increases is far from a good idea.
On top of that, the idea to just use the government-collected tax to fund anything is a poor idea because those people who have to collect the tax, pass it on to those who distribute it, and then those who do distribute it all need to take a cut out of it, because, you know, they're doing work and deserve a paycheck.
So far we're seeing larger businesses successfully weather the tax while smaller will likely struggle with it, innovation either be stifled or limit undercut the success of the tax, and bureaucrats getting paychecks out of what is meant for the targeted elderly and youth. Unless you are a bureaucrat, or a large business wanting small out of the way, where is the benefit to this idea?
Now, if you actually want to do something to help prevent people from losing their jobs because the robots offer cheaper labor, figure out how to reduce the cost of living, because then the human labor can be less expensive and more competitive. Just don't do it artificially by subsidizing the cost of living, reduce the artificial costs that already increase the cost of living. It is harder to do, but it is the 'rising tides raise all ships' solution.
I think that's enough blathering on for now. (By the way, I'm just a guy, far less wealthy, and economically insulated, than Bill Gates, like most everyone on the planet.)

Did I Miss It?

We've made it through another CES and I am left wondering what I missed. I mean I know some neat stuff was shown off, but nothing got me especially excited. Maybe I am just changing, but I honestly did not have a wow moment with all of the stuff I saw talked about. Maybe I missed it or maybe I just knew enough of what to expect to not be surprised. I knew VR would have a sizable presence, and nothing there was really surprising, and while it is cool to see OLED displays becoming more common, I know enough about the technology to know that has been coming for years.
Did I miss something? Probably. I'm sure I'll have a chance to see it again though.

Years in the Making…

Really, the effort leading up to this was years long…. Yeah, most of the time was not spent working on it, but so what? Everybody takes breaks.

However long ago I asked the question of if it was possible to automatically update the files on my USB drives. After some research, I found the building blocks needed to create such a system and that first system I made was crude, but effective. Since then I have completely redesigned it to operate with different USB devices and different computers, and trust me, those differences can make things complicated.

Well, knowing that the real fun of creating such a system is from sharing it, I wrote a guide explaining how it works and thus how to recreate it. Thanks to the supporting folks at OverclockersClub.com, that guide is now published, for the world to benefit from… or at least those in the world willing to work with batch files, the task scheduler, and possibly environment variables.

Enjoy! USB Auto Update Guide

See the Silver Lining, Now Finding the Cloud

I wish I was more interesting. Then I might be writing more stuff for here. Anyway, here's something interesting that is not about me.

I like science in general, so I follow and read a lot of science news, including climate science. To be inflammatory though, you could call it climate propaganda given some of the information being put out there of late. Basically what this information is saying is that claims about global warming climate change are not being thoroughly investigated. First up, biofuels:

Biofuels are these beautiful things, aren't they? You grow a plant, process some of it, and you get fuel to not burn in your car (ethanol is actually bad for your engine, at least in concentrations above ~15%). The combustion of the fuel releases carbon – bad – but growing the plants takes carbon from the atmosphere – good – so everything balances out, right? Actually there is no evidence of this, as was found by a meta-study (meaning it studied other studies). No one has ever bothered to actually measure the total carbon output and input of producing and using biofuels. The studies just assume the output is the same as the input. At the same time as making that assumption though it also counts the carbon plants absorb…. So along with the assumption, they are double-counting the carbon sink. Add onto that lack of scientific rigor the fact that apparently the nitrogen cycle, which could have ever great impacts than the carbon cycle, has largely been ignored when considering biofuels. (That's from another study I don't have the link handy for.)

So why am I talking about this today? After all, that study I link to above is from June. Because I just got another piece of science news that again shows a lack of scientific rigor.

One of the ideas put forth concerning climate change is that it will increase droughts, and according to one measure, that is true. Unfortunately that measure, as scientists are now pointing out, is completely useless for this measurement.The means of measuring drought for this concept is to compare precipitation and evaporation. Well, measuring precipitation is easy with a rain gauge. Measuring evaporation is quite difficult though, because of all the factors involved, including wind, humidity, and temperature. For this drought measurement though, it only considers temperature and assumes that higher temperatures directly lead to more evaporation. So again it is an assumption that is being touted as truth.

When using a different measure of drought, it actually turns out that it was worse in the 1950s and 60s, before global warming was a thing. Wasn't that when people thought we were heading to an ice age and over population was the source of all ills.

For what it's worth though, this is how science has always worked; someone makes a claim, likely without examining all aspects, and someone else comes around to do that investigation. The problem now is the people who try to shape the world on the 'silver-lining science' (that's what I'm calling it) before finding the cloud.

This saddens me…

This is also going to show the kind of nerd I am.

Just read that there is now a plan for Japan to do away with nuclear power by the year 2030. However, this might not happen for odd political reasons. (When the plan may actually become law, there may be a new party in power, which could vote differently.) It still saddens me though because this decision, and similar decisions made in other countries, I believe are not in the best interest of the country but just what people are told they should think they want. Really it seems like everything to do with energy is like that, which probably isn't surprising. People don't care enough about where the electricity comes from, provided it comes every time they turn on their TV, computer, and microwave, so they don't actually understand what is involved, which makes them susceptible to propaganda that is not about energy, but about control.

Nuclear power is clean, it is safe, and the only time it isn't, it's because someone is not doing their job and fulfilling their responsibilities. But so long as the public excepts energy as a 'black-box', this won't change. (Black-box, for this situation, means you do not know what happens inside of it but just that it works when you give it the proper input.) Anyway, rant coming up, so read on, only if you dare!

I've Said It Before and I'll Say It Again

Science fact is more fantastical than science fiction. This is why I ever had the idea to write Sci-FU, so any time I see some science news concerning something I wrote about, I get particularly excited. In this case though, I actually haven't written it yet, but I've been planning to.

Quantum physics at a distance

Physicists at the University of Vienna and the Austrian Academy of Sciences have achieved quantum teleportation over a record distance of 143 km. The experiment is a major step towards satellite-based quantum communication.

One of the events in the sequel I've yet to start formally writing (I've been writing bits and pieces as I have ideas) for Sci-FU is the development and launching of a satellite network that uses quantum mechanical phenomena, like quantum teleportation, to communicate. Looks like we may have such satellites much sooner than the roughly 50 years Sci-FU is in the future. Read on to find my analogy for explaining quantum teleportation.