plekter

While in general I am very much in favour of playing both sides of an abstraction, I would argue that RTL is nowhere near the top of the stack.

In the GPU world, you have games, which are built on game engines, which are built against an API, which is implemented in a driver (which contains a compiler!), which communicates through OS abstractions with hardware, which is written in a HDL (this is where you write RTL!), which is then laid out in silicon. Now each of these parts of the stack have ridiculous complexity in them, and there are definitely things in the upper layers of the stack that impact how you want to write your RTL (and vice versa). So if your stack knowledge stops at RTL(which, honestly there is absolutely nothing wrong with!), there is still lots of fun complexity left in the stack.

Modern gpus + drover stack usually had more than one frame in flight. You have to output a frame every 4ms, but you do not need the latency from the start of the application rendering code to the frame being on screen to be 4ms - pipelining is allowed. But keeping that pipelining down to a minimum is also important, as it contributes to input lag which gamers care about.

I think multisampling may be the answer.

For partial rendering all samples must be written out, but for the final one you can resolve(average) them before writeout.

GPU hardware architecture. It's just a never ending stream of fun problems to solve and I love it!

On reading your original post again, what you experienced (both being super exhausted after work, and a shower helping) resonates with my what I felt early on in my career.

Things that helped were - Take care of primary needs during the workday. Eat, drink, bathroom breaks. The basics. I could get so caught up I would forget about those things.

- More experience helps. Work is still incredibly taxing mentally, but as I've grown a bit more battle worn I don't have to second guess whether my work is good enough all the time, which is also exhausting. Being introverted, having been in the same place for a while (not to mention - being remote) also helps, I'm not being bombarded with people I don't know any more since I know them all.

- Having kids was brutal in terms of taking time and energy, but luckily it somehow came with an 'extra gear' to cope with it. Totally worth it though, I can only hope it becomes an option for you with time

- I also found that having a plan for what to eat after work was worthwhile, so I didn't have to think - just cook it and eat it.

- If something work relates is spinning in your head, write yourself an email about it and postpone it until tomorrow.

Not sure if any of this is helpful to you, but worth a shot :)

Kids.

That's it, really, and I love it. Nothing makes me snap out of it like being assigned a role from Paw Patrol by my two year old, and going off on some rescue mission. It's pure bliss.

I totally agree that the existence of these drugs make the world a better place. I also totally agree that there should be significant financial upside to developing drugs like these, after all we do want to incentivise making the world a better place. But surely there must be a middle ground? This seems to be priced by a "what are people willing to pay for this?"(a lot, given its life saving/changing nature) vs "what do we need to charge for this, including a healthy profit for everyone involved". (much less, presumably?) I fundamentally dislike, from the perspective of the patient, the thought that precisely /because/ the drug is desperately needed by someone through no fault of their own, it /deserves/ to be expensive.

The foundation getting a cut of the success and reinvesting in new opportunities sounds good though, thanks for pointing that out!

That particular drug was on the news in Norway this week because it is so incredibly expensive that the state don't want to pay for it. We are talking 20-30k USD per month per patient.

I partially excused that mentally with a "but yeah, research is expensive", but now learning that it was paid for by donations ruins that argument.

The fact that the state don't want to pay for it is of course also debatable..

It works in Firefox on my a32

I don't think they can use jpg directly, that would be a waste of transistors given that the graphics world use other compression formats like etc1, bc, astc and so on.

It is however perfectly possible to decode blocks of JPG on a GPU by using shader code.

Micropolygons. He said on twitter that most of the triangles are pixel sized or smaller.

This approach would fall flat on its face with large primitives:

- The for loop in the rasterizing code would get very large strides between lanes in a warp for the frame buffer memory accesses

- it doesn't parallelize across pixels in a primitive, so at its extreme, a single big triangle would be single threaded..

- A HW rasterizer has a lot more things it needs to do (compute barycentrics, keep around enough state to launch a fragment shader, etc)

- Quad occupancy. Conventional rasterizers go down to 2x2 quads as their unit to work with, because you need some neighbours if you are going to calculate derivatives for mishap selection when doing texturing. The hardware is designed around quads, and all lanes in a quad must come from the same polygon - leaving 3/4 of the hardware downstream from the rasterizer unused with single pixel polygons.

These are problems that conventional rasterizers have, but if you are writing a compute rasterizer you can conveniently skip lots of things.

Are the consequences of the massive energy usage also opt-in?

I'll wager a guess: ASML, the Dutch company that makes the magic machines TSMC(and others) use to make chips.

I agree with that - but if you take an unchanged core and manufacture it at a different node, then you won't see a change in IPC, which in my book makes it questionable to attribute IPC gains to the process node.

Process node changes generally doesn't do anything for IPC - those are generally rather due to microarchitecture improvements, so I doubt the move to 5nm has anything to do with the IPC gain..?

Mobile GPUs (Imagination PowerVR, Qualcomm Adreno, Arm Mali, Apple's GPU) are exactly that.

Some of these are pure gold!

"2:4 And this I pray, that your love may abound yet more and more like a controlled use of shared memory."

I used Meshuggah - Bleed and a carrying harness to get my then 2 month old to sleep. Close enough to white noise, and motivation for daddy to keep rocking. Not sure I want to read research on whether that was a good idea..

I live in Northern Norway in eastern Finnmark. This article seems to miss these points

- When you are sufficiently far north, winter means beautiful snowy landscapes, aurora borealis, good stable conditions for cross country skiing, no grey slush. Winter is just straight up better than further south.

- People who live here can move south whenever they feel like it, so those who don't stay by choice. It's not like this region is a hub of great career opportunities and so on, unemployment rates are generally higher than the rest of the country, so people don't get stuck because of work quite as much. So, there is clearly a selection bias towards people who like winter!

Personally I don't think there's too much magic mindset stuff going on. Winter just doesn't suck as much as the author seems to imply. Also, there's plenty of SAD going on so it's not like we are immune..

I do wonder what this means for Arm's Mali GPUs if this goes through.

I'd argue that modern GPUs pipelines are quite different from the API programming models they support, especially tile based renderers do lots of funny things while maintaining the illusion that you are running on the abstract OpenGL pipeline.

So even with GPUs you have a similar situation..

I was in Kenya just after this, when Miguna Miguna went missing for a few days (people feared he had been disappeared,but he "just" got deported). As chance would have it, I was in his home town. Quite heavy demonstrations - makeshift roadblocks with burning tires, and they raided a beer truck, and the police did crowd control by shooting into the air.

I was in Kenya in 2007 as well when the same opposition candidate also lost the election. That was much, much uglier. While I do sympathise somewhat with his followers (the region he is from is much poorer than central Kenya, and this follows tribal lines - they feel they are not being heard, and there is some truth to that), I do think that this candidate is a quite dangerous man to have in power. Any influence he has now is coming from the fact that his followers will tear the country up if he says so - and he did say so back in 2007.

Kenyan politics is a bit of a mess I'm afraid..

So you see it as a general career progression impediment - even the gifted gets "slowed down" progression wise?

Are you talking from personal experience, or guessing? I strongly suspect a lot of this depends heavily on how the talent review process works in the company this takes place in. A gifted engineer will likely get recognized anyways, but an average engineer will need the review process to look beyond face value.

reality is, if your team is 80-90% not remote, forget about career advancement.

That's quite a sweeping statement - and doesn't resonate with me at all. I fit that statistic, but I work in a company that has a mature process for career advancement and that also has multiple sites scattered across the world. I cannot claim that remote work has held me back career-wise. That is, I'm still quite early in my career, maybe the wall comes later.

It has, however, likely steered my career away from management (and thus towards the technical track), which I'll happily admit does not bother me in the slightest :)

Yearly reviews, where high performers get recognized and promoted. My career progression has matched that of my peers, so I have no complaints there.

I've also worked remotely for 4 years now, but in a company where remote workers aren't that common. I absolutely love it - I love the autonomy, the lack of office distractions, etc.

I think the key things for me are

- Love your work, motivation then comes for free

- Communicate a lot, as the article says. Personally I'm a fan of written communication, as I find that is frequently more impactful. Easier to share, and so on.

- 1to1s are also great

- Have someone in office you can talk bullshit with over text chat. That helps a lot with the isolation and staying on top of office gossip.

- Being remote naturally encourages a more independent working style. Being forced to solve problems myself and actually think has been great for me. Not being in office provides the calm and quiet to do exactly that!

- A lot depends on your manager. Mine is trusting and accomodating, and I very much appreciate it.

- Finally, having a good internal network in the company is important. Putting in the travels to get some face time is important, as written communication gets a lot better after having met.

It looks nowhere as good as MRI - but has the advantage of being very quick to do. MRI will also image different things (depending on settings). US will show you interfaces between different kinds of tissue (that's where you reflect part of the US beam - echo!), but MRI has a million settings - and will fundamentally show you water at different densities. What that means in practice depends completely on the acquisition sequence.

The use case in my thesis went something like this

1. Do MRI - use this to plan surgery and to navigate while in surgery 2. Once the patient is on the table, register the MRI image to the physical location of the patient. 3. With various 3d tracked tools, you can then "point" at things in the real world, and have it show up on the MRI image. 4. As you are performing the surgical operation, you want to check "how does it look now?" - and that's where the tracked US probe comes in.

The use case outlined above isn't cost-sensitive, but time-sensitive. Getting an US image takes seconds and can be done mid-operation. Another point is that 3D probes are quite big and clunky - and for AVM-type operations where you crack the skull open, you want to keep that hole as small as possible. Phased US sensors + reconstruction then gives you 3D images with a very small sensor footprint. Another point is that 2D probes have higher resolution, so at least the XY plane looks much better this way.

e: there are some pictures in my thesis: https://brage.bibsys.no/xmlui/handle/11250/253675 The system we used/were developing: https://link.springer.com/article/10.1007/s11548-015-1292-0

I wrote my master's on doing the reconstruction a few years ago. On my shoddy 10 year old laptop GPU I could reconstruct a 3D volume from 800-ish slices in about a second - depending on interpolation - so on modern hardware (or even mobile hardware) I expect this to be doable in real time.

I do wonder about the spatial precision though, we used optical tracking back in my thesis, and that was already noisy. There are techniques using speckle patterns to help with the registration - but that seems like a tough issue to solve. Either way, this is a cool idea and nice tech!

e: Oh, looks like it's orientation only - you'll need a steady hand making the sweep then, since I suppose you can't move the position of the probe, just it's rotation