oflordal

You can do that on both HIP and cuda through e.g. hipHostMalloc and the cuda equivalent (Not officially supported on the AMD APUs but works in practice). With a discrete GPU the GPU will access memory across PCIe but on an APU it will go full speed to RAM as far as I can tell.

Did anyone evaluate this compared to using api access through an external gui (i.e. continue.dev). For software dev did the cost end up higher? I am thinking this is can be more convenient (and I suppose engineers can more easily use it outside work as a perk). Given practical use across a team will vary you get a lower price when using api and perhaps additional opportunity for scripted use.

Any company that are not allowed to share code with third parties either by internal or customer policy. This should be a lot of companies.

ARM big little (which the M1 is at least somewhat based on) was announced the year before the prediction.

No, this is about HW architectures. While they are likely evolving towards one a other there are tile based (like Imagination and ARM Mali) And immediate mode (Nvidia AMD) that both implement the same APIs (OpenGL, Vulkan etc). All these HW architectures are modern and in use.

It is as much or more about saving power by reducing costly traffic to the external RAM. For the pathological cases where you don't compress you simple don't save power, on average you do.

If you think it is fun, having a hobby (mturk) that is a small net positive can save you a lot of money compared to having a hobby which is a net loss (most hobbies).

Supposedly the step in the middle is SW optimizations?

This is obviously a nice gesture by Nvidia but I always wonder how can it be legal to have different amounts for father/mothers? Are there no laws on discrimination based on sex in the US?

There are games and experiences that people seems to think are interesting enough to play on the GearVR. The PS4 will be more powerful than the phones powering that. Obviously you have to adjust your expectations compared to the Rift but it seems like there can be value for VR platforms that are not as fast as the Rift/Vive minimum specifications.

ARMs customer facing code is 100% verilog and this reflects most of the industry.

The greyscale gradient animations that are meant to show banding are full of gif artifacts so they are not actually smooth gradients.

Its been 10 years since standardization and the eda vendors are still whining about SystemVerilog implementation...

Interfaces was a great addition it serves more uses that simple structs (which were not available before SystemVerilog either). The primary use he doesn't seem to get is that it is a great place to put your interface assertions so all the intent of the interface end up in one place. The way this is typically done when SystemVerilog is only used for testbench/assertions is to put them in a separate file anyway so they don't have to be included in both the sender and receiver.

>Not sure what difference the 4 in 764 It says it is a 4 core device.

Here is a video of Pierre Spray (one of the F-16 designers) discussing the same thing: https://www.youtube.com/watch?v=mxDSiwqM2nw

The testbenches in that directory are supposedly too basic to have been used for development (a few directed test cases per module). I wonder why they wouldn't release the proper testbenches?

KTH in Sweden added a special course that was essentially there to practice for this kind of competitions. "Problem Solving and Programming under Pressure" http://www.kth.se/student/kurser/kurs/DD2458?l=en

I have been thinking this will change as self driving cars becomes more prevalent. Commuting by car is very inefficient since all the time spent driving is wasted (compared to public transport were you can potentially work on the way to work, or cycling/walking were you get exercise). Self driving cars that are sufficiently good to handle the commute most days would change this and allow you to live further away from work in places that lack public transportation.

Säd. Coindidentally the same happened to one of the Oculus founders. http://techcrunch.com/2013/06/01/oculus-vr-co-founder-andrew...

Being from Sweden with its relatively generous parental leave (~18 months for both parents, a fairly common company perk is to get 6 of these with 90% pay) its hard to understand how especially mothers can return to work already after 12 weeks. Will most people in practice work out a deal to take a longer leave and or be forced to quit?

Frankly, I am amazed technical marketing at Semi companies get the specifications right as often as they do. They typically don't get to bother engineering all too much and even if they typically have worked in engineering before they probably didn't work at an architecture level.

Seems entirely plausible this information had not made it from engineering to technical marketing and/or been missed in a review.

Try http://www.robozzle.com for a game in similar vein with a bit more depth.

While this may be of little concern, compared to your PS3 this will use significantly less power. Comapred to your TV is is very possible that the ecosystem for this product will be larger.

Every modern ARM SoC built for thisnpuepose has H.264 hardware video decode (and typically HEVC for that matter).

I have not tried it myself but both Altera and Xilinx can compile opencl for their FPGAs nowadays

On Swedish versions of Word find is moved to Ctrl+b since the word for bold starts with an f. Can get very annoying when mixing versions.

Chalmers univeristy has a HDL built as Haskell modules called Lava: http://www.cse.chalmers.se/edu/year/2012/course/_courses_201... Don't think its used outside of the university though.

HW decoders will not send the deblocking barrier to external memory (typically). Instead you will buffer the required rows until you have decoded the next line of macroblocks. As long as the filter margin are the same this will work fine in itself but you will not have done the buffering at the same place you did deblocking or any of the simpler overlap filters. In that way you will force larger changes to your architecture. Similarly your transform unit is likely not set up to handle this so it requires redesign. Getting to many of these hurdles in can be the death of adoption for the CODEC.

Its hard problem making something unique enough to avoid current patents while still being a nice fit for a haäw pipeline. Things like överlappning transforms may end up creating new minimum values for the vertical context needed.

New itu/MPEG standards typically has lägre involvement from hw companies.

Especially with the url avc.com which many (most?) people would relate to Advanced video codec (another name for H.264) you would have thought this was about compression.