stracer

If malicious program has access to GPU directly or via some buggy interface, the whole system is at risk. There is no "safe" GPU virtualization like there is with CPUs.

Yeah, why do they change options so often. They should keep some backward compatibility, qemu is not a new project.

why the milestone isn't something classically hard but easily verified

They're far from being able to do any such thing, so posing such milestone would make the field look stagnant and thus be a bad marketing and hurt the money stream. The "milestones" are chosen so that they are plausibly reachable in short time relevant to patrons/investors, because people working on this need to constantly demonstrate progress.

All the manipulations the Americans imagine are contradicted by events.

What does this mean?

the ideology collapsed, economy doesn't perform

Economy doesn't perform, but ideology has collapsed only in minds of ordinary people. Politicians, stakeholders and various media outlets are very much invested, and still push that the current course is the only correct way and the bright green future as designed is unstoppable. Reminds me of the arrogance of the ruling party slogans from before 90s.

There is not much money in it for Intel.

Intel tried to get into GPU-like products 14 years ago. They promote their consumer Arc GPUs since 2022, and still almost nobody wants them. Is there a datacenter Intel GPU that some business wants?

The big money is flowing NVIDIA's way, and even if Intel can make a GPU, it won't be able to divert a big part of the flow, similarly to AMD.

Yeah, but they can saturate fabs and provide income, which they need. Intel can't produce better CPU/GPU products than their competition now. Their design and manufacturing of CPUs has serious problems for years now. The big money in GPUs is already captured by NVIDIA, and it's hard to see how Intel can challenge that - people want NVIDIA and CUDA. So Intel should cut down and focus the remaining bloat and R&D spending on the areas where it's plausible they can get competitive in a reasonable time. That is CPUs, and maybe memory and SSDs - they have X-Point which was great, just marketed and priced wrong.

The idea of another competitive GPU manufacturer is nice. But it is hard to bring into existence. Intel is not in a position to invest lots of money and sustained effort into products for which the market is captured and controlled by a much bigger and more competent company on top of its game. Not even AMD can get more market share, and they are much more competent in the GPU technology. Unless NVIDIA and AMD make serious mistakes, Intel GPUs will remain a 3rd rate product.

"They need to start somewhere in order to break ground"

Intel has big problems and it's not clear they should occupy themselves with this. They should stabilize, and the most plausible way to do that is to cut the weak parts, and get back to what they were good at - performant secure x86_64 CPUs, maybe some new innovative CPUs with low consumption, maybe memory/solid state drives.

Too late, and it has a bad rep. This effort from Intel to sell discrete GPUs is just inertia from old aspirations, won't really help noticeably to save it, as there is not much money in it. Most probably the whole Intel ARC effort will be mothballed, and probably many more will.

you bring up Procter & Gamble,

No, relwin brought it up, and I ridiculed the criterion. I don't get your point at all, sorry.

Why would you ask that? MIC = Military Industrial Complex. The complex is made of both the industry and the military.

If Ryan McBeth defines MIC as just the defense companies, or its existence being conditioned on these companies having greater official revenues than Procter & Gamble, then he either misunderstands the concept, or he is (why?) trying to spread some weak argument for an idea that U.S. does not have an MIC, which is quite comical.

MIC is not just the defense companies, read the name again - it' the complex made of industry and the military. And Eisenhower's point in warning against it is not about revenue of the industry part, but about influence of the whole complex on major decisions.

In 1990's there was a short dip in funding, but since 2000's, its growth caught back on, and it's getting close to a trillion dollars a year. That much money chases a lot of constituency and a lot of power. Millions of people are dependent on it.

There's a certain personality type that is drawn to engineering that believes the whole world can be explained by their simple pet model and that they are smarter than everyone else.

Lots of failed theorists with that personality type/flaw as well.

It's a flop for people who expect that after 20months, they'll get 13%-18% more. Perhaps Steve did not know this "already well known fact" and thus wrote about flop and Intel 11-th gen vibes.

How did you know before benchmarks were out? Did AMD say gaming performance will stagnate? (that would be very stupid thing for them to say).

In which applications is AVX-512 performance decisive? Video editing / 3D modelling?

Yeah, Moore's law seems to culminate and the rising tide with it. We seem to be entering an era where new tech improvements are mostly about integrated coprocessors and specialized workloads, and sometimes lower power.

Yep, after Zen 3-4/Alder Lake, we're now in another stagnation cycle like Intel Sandy Bridge brought in 2010-2015.

For casual users, a good test if it's time to upgrade is - has the CPU speed, the memory speed doubled? If not, for most people, it's better to keep your old device.

For competitive gamers, smaller gains make sense, but it gets expensive fast.

Steve's reviews are focused on gaming and he probably isn't wrong there. 9700X seems to bring very little for gaming, compared to 7700X. 7700X is almost the same performance, sometimes even better, and cheaper to buy. Most gamers do not care about 10% lower power consumption, especially when they have to pay more upfront for it.

Yes, the better energy efficiency allows 9700X to boost higher to get somewhat higher performance in multi-thread, but for some reason, gaming does not benefit noticeably. Zen 5 seems interesting for laptops and low-power client devices, but so does the new Intel Lunar Lake (for now, at least), so we have to wait for reviews and comparisons there.

has not tested... floating-point computations, cryptography or ML/AI.

That is true, but small fraction of people care about these specialized workloads, it's overhyped in marketing. But if you care about those, I think you may be right that Zen 5 may be more interesting there.

9700X system has materially the same performance and only a little lower consumption than similar 7700X system, which was released 21 months ago. It's stagnation like Intel 2010-2015.

If you wanted to upgrade from Zen3/4, don't get your hopes up...

https://www.youtube.com/watch?v=OF_bMt9fVm0

On average, 9700X is few percent faster than 7700X, and has slightly lower consumption. Upgrading from Zen 3/Zen 4 not warranted.

Is this Russian authorities, or Google themselves, slowly leaving?

Not sure what your point is.

time dilation to make it go fast (from our frame of reference)

The only way to do that is that we move to a place from which the computer processes appear accelerated, e.g. into a strong gravity well. That isn't very useful, because we do not have such a well nearby, as only very dense hypothetical objects can provide it (e.g. black holes), and it would not be compatible with life to move there.

Why is the blog post wildly wrong? He states the seal joints were poorly designed, similar to your statement. His statements seem reasonable and in agreement with Feynman's and others' (Boisjoly) account, which tell a story of corrupt NASA and Thiokol management, who pushed for flying outside the safe temperature window.

Polarization can be measured using polarization filter and light detector, but it is destructive in the usual sense of quantum theory. That is, if the detector after polarization filter clicks, we know the EM field had non-zero component in the direction of the filter, but we do not find out the other components it had before entering the filter.

I meant the message bits carried by the emission, but you're right, the way things are done, more bits are prepared at first, to make the message more resilient against noise, and only then all the bits are modulated onto the EM emission. So while not all emitted bits do come correctly, all data bits can be reconstructed.

Which makes me wonder, whether this way of achieving resilience against changes along the way is necessary. Maybe the data bits can be modulated directly onto analog current, and error resilience can be achieved by remodulating it on the analog level. Then, the emission would have as many bits as there are in the message, error-resilience (modulating, demodulating) would be done strictly using analog electric circuits.

Can you please post another link? This one does not work.

Is there some fundamental limit to the number of bits per photon that can be communicated via EM radiation? I think it does not exist, because photons aren't all equal, we can use very high frequency and X-ray quantum can probably carry much more information than RF quantum.

All transmitted bits are also received, at least when everything works as intended.

A compiled microkernel can restrict, I agree. But if you want to restrict, why would you do a microkernel. It's much easier to justify with a monolith.

One big point of making a microkernel as opposed to making a monolith is that you want to give other developers more freedom to develop, distribute, fix bugs on their own time schedule, while keeping compatibility with new kernels easily. In other words, not wanting to control what and when they do, and what users run on the kernel.

Monolithic kernel like Linux, with its constant stream of little changes in interfaces, creates problems for developers of modules, and this leads to selection - those who keep up and satisfy the kernel requirements get in, those that don't keep up or do not satisfy requirements, can't get in, and are degraded to developing against some old kernel version. This is a kind of control.

Where did I mention microservices? How does the fact current microkernel OSes are being distributed using a monolithic model imply my point is nonsense?

A real microkernel OS with stable interfaces would mean its developers do not want control of all the kernel-like servers, and allow even basic users to easily install and use many alternative user servers, as opposed to Linux kernel services sanctioned and controlled by the Linux distribution kernel maintainers.