Second Life and Virtual Worlds
August 2, 2016 Nvidia released their top of the line video card the Titan-X. There is a previous generation Titan-X. To distinguish between the previous gen and the current generation we have to include the architecture descriptor; Maxwell (previous) or Pascal (new).
How much faster is this Titan-X Pascal than the previous one? Almost twice as fast.
You can see the card is about 30% faster than the GTX1080.
In the last 10 minutes or so of the Third Party Developers’ Meeting the problem with AMD video cards problems came up.
It seems especially the R series of video cards (RX480 etc.) have a problem with OpenGL. See: BUG-20057 – Visual Artifacts with ALM enabled on some AMD graphics cards.
This bug is dependent on screen size. So, you may be able to escape it by changing your viewer window size or screen resolution.
There is a thread in the AMD forum about the problem. It is way techy. See: OpenGL display artifact on W7100.
In building my new computer I needed to verify my calculation of the amount of power required. What size power supply do I need?
By my, mostly in my head, calculation as I selected parts I need 500-550 watts at peak demand. I select a continuous supply rating based on my anticipated peak demand. While power supplies have built in the ability to handle peak loads that exceed their continuous rating and that often gets advertised, that ability usually comes from considering what all the parts can stand when pushed to >100% of their design spec.
Humans can walk all day at 5 mph. But, we can only run at 20 mph for a few minutes. Sort of the same thing with computer power supplies. At some point in time they overheat or an individual component overheats by too much and the unit fails. Some units have anti-self-destruct protection, which reduces the power supplied to save itself. And what does that do to the rest of the computer? Nothing good.
There is debate in the SL Forum as to which video card is a better choice for Second Life™. I’ll try to provide the information needed to make a decision. But, I don’t see a deal maker or breaker that would pull the choice to any one card.
Also there are no good benchmarks, yet, of which new cards work best with Second Life. The following benchmark tests show the cards’ performance are very dependent on how the game is designed.
In the above video he is testing DirectX12 verses Vulkan. Vulkan is the next generation OpenGL. While development on OpenGL has not stopped, it appears to be similar to SL and Sansar. Except I have no doubt Vulkan will be the render engine of choice for most game makers moving on from OpenGL.
Regular readers will know I am in the process of building a new computer capable of handling VR, Sansar, Obduction (from the makers of Myst), and Second Life. I mostly deal with web tech and repair or upgrade of old computers. I do a lot of repair and upgrading of Core2 vintage machines.
So, when I decided to upgrade I started looking for what is best, fast, and cheap. In the process I realized I haven’t been keeping up with the tech. Helping people in the SL Forum I realize there are a lot of people also unaware of the changes in tech. A lot of bad advice is handed out in the forum. So, I am starting a series of articles on the current state of computer technology. Starting with a tutorial on computer memory.
Computer memory is called: RAM – Random Access Memory.
Memory chips like all other tech advances. Memory gets better, meaning faster and using less power. So, can you put faster memory in your existing computer? May be.
A few days ago NVIDIA released the GTX1060. The 1060 is the mid-range video card for gamers. The GTX1080 has been out for some time and it was the top end video graphics card for gamers. Now 25% faster than the 1080 is NVIDIA’s 2016 Titan X (Pascal) Video Card nearing its release date, August 2, 2016. (NowInStock)
The cost of a GTX1080 is in the US$700 to $800 range. The 2016 Titan X is rumored to have a MSRP of $1,200. That is 50% more in cost for 25% more in performance. Not the most GPU per dollar but, it is fast. For the enthusiast this is sort of like comparing a Geo-Metro to a Farrari using cost per pound (kilogram) or top speed. Incomplete comparison.