kllrnohj

44 ms·

811.
by kllrnohj 3y ago
Intel had / would have had few significant issues at the time competing on the CPU side of thing. What killed them in phones was the radios. They sucked at them, and buying a discreet chip for it hurt margins and efficiency too much.
812.
by kllrnohj 3y ago
That is simply not true. Kavari (2014) had pointer passing between CPU & GPU. I don't know where you're getting that pages couldn't be shared from but it's not accurate. https://images.anandtech.com/d
813.
by kllrnohj 3y ago
> The M1 chips have a unified memory architecture shared between GPU and CPUs. This is a masterstroke for performance. No it isn't. It's a nice optimization to have, but it doesn't radically change performance. If unified
814.
by kllrnohj 3y ago
RISC-V isn't freedom hardware, there's no equivalent to copyleft.
815.
by kllrnohj 3y ago
For the majority of open source projects the indirect effects are the potential for more eyes, more bugfixes, and more knowledge to spread. RISC-V CPUs don't have any of that as there's no requirement that actual implementation be
816.
by kllrnohj 3y ago
> Remember when Google told everyone to test their Android apps on x86. And then again on MIPS? And then nothing happened. Developers will not waste their time. x86 builds of Android apps & libraries are extremely common to have sinc
817.
by kllrnohj 3y ago
There's an entire industry of consumer products that take massive amounts of compute called "gaming". Apple could have showed off the chips performance at that. Or at least something like photo or video editing?
818.
by kllrnohj 3y ago
> Consumer devices with ARM cores can be had for as low as 2 USD (sensors, breakout boards). Yes, and that is the market where RISC-V is actually appealing. Once you go higher end, the appeal of a no license cost ISA rapidly drops of
819.
by kllrnohj 3y ago
If you're comparing between two products and one is $1500 and the other is $1501, is that $1 really entering your list of pros/cons? No, of course not. It's irrelevant at that point, literally any other difference (even what
820.
by kllrnohj 3y ago
AMD & Intel also pay $0 for licensing costs. For the phone markets & up through servers, licensing costs are a rounding error on what you pay. You're paying for the R&D of actually designing the chip itself, not the ISA, as
821.
by kllrnohj 3y ago
So far all EVs have tanked in resale value.
822.
by kllrnohj 3y ago
> The main downside of NACS is that it doesn't support three-phase The other current downside is it doesn't support up to 1,000V like CCS does, which matters for the currently deployed 800v EVs out there. There's a revis
823.
by kllrnohj 3y ago
Why? The adapters are cheap and plentiful since NACS is using the CCS protocols, so it's just a physical connector difference
824.
by kllrnohj 3y ago
> When these 'base clocks' are listed, would a lower core CPU with a higher base clock run more calculations at its max thread count than a higher core cpu? Say 64 threads on a 32 vore 4Ghz CPU vs a 2.5Ghz 96 core monster. The
825.
by kllrnohj 3y ago
> I would be very surprised if Ryzen goes for 32 cores, 24 is more probable. The rumor is that the CCX is going from 8 to 16 cores, at which point a 24 core ryzen would necessarily be 2 partially-enabled CCXs. That wouldn't make muc
826.
by kllrnohj 3y ago
Citation was in my post if you read it more carefully. Android+ChromeOS is easily the 90%+ of the "Linux" market that cares about having a GUI, and even omitting Android then ChromeOS is still more market share than desktop linux
827.
by kllrnohj 3y ago
I mean, insert any "year of linux on the desktop joke here" In terms of usage numbers, X is very, very dead. Android uses something that looks a lot more like Wayland than X (or arguably Wayland looks a lot more like Android'
828.
by kllrnohj 3y ago
> Isn't C the language (x86_64) processors are designed to be fast for? Nope. They compete on performance in C++ (games mostly), Java (enterprise SKUs, but same core architecture), and JavaScript (browser benchmarks even though raw
829.
by kllrnohj 3y ago
https://developer.nvidia.com/blog/how-access-global-memory-e... SIMT still expects coalesced memory access that's close together otherwise performance falls off a cliff
830.
by kllrnohj 3y ago
DirectX 8 added programmable shaders as we mostly know it today in 2000 with directx 9's shader model 2.0 really solidifying things in 2002. CUDA didn't show up until 2007.
831.
by kllrnohj 3y ago
> I really don't know why they do this though. Because you're looking at absolute numbers and not percentages/relative values. That much cash on hand sounds absurd when you're only looking at the absolute number of it
832.
by kllrnohj 3y ago
> and the papers were not willing enough to call "bullshit" on the front page. The truth was out there but hardly staring you in the face. It kinda was, though. A bunch of lawyer-reacts at the time basically were "this isn
833.
by kllrnohj 3y ago
There's plenty of abstraction libraries available including for example webgpu. Why do you want immediate mode specifically? It's not a very useful balance to strike between just using an actual 2d library vs. learning how shaders
834.
by kllrnohj 3y ago
For some reason you're getting a lot of wrong or just bad replies. But the answer to your question is yes both sRGB/gamma2.2 & log are non-linear, but almost in the opposite direction. gamma2.2 is exponential not logarithmic.
835.
by kllrnohj 3y ago
No, gamma2.2/sRGB is how the pixels are stored on disk, not linear values. Linear is almost never used except as an intermediate for processing where you can throw lots of bits at it (eg, fp16/32 on a GPU when applying effects or
836.
by kllrnohj 3y ago
> However, it reads (to me) as not enough people are playing the game on the mac so they are dropping support. To me that doesn't read as being too hard for them to support mac. I think you're missing the context here that CSGO
837.
by kllrnohj 3y ago
> Is this the same Apple whose new iPhone SOC has hardware ray tracing? IMG has had raytracing in powervr for like a decade, and they just did another "most advanced ever!!!" thing a few years ago ( https://www.imagi
838.
by kllrnohj 3y ago
> Those "native" APIs you mention in most cases just call back to Java APIs under the hood Some do, but most don't.
839.
by kllrnohj 3y ago
> a major part of JVM, JS Engines and even memory allocators, pools and similar is to make this patters less random, more predictable and more local. They don't, though. The JVM memory model doesn't even really allow it. This i
840.
by kllrnohj 3y ago
> but apple’s literal stacking of the LPDDR on the actual die doesn’t? There's pictures of Apple's silicon all over the Internet, didn't think to bother even looking at one? https://cdn.mos.cms.futurecdn.net&#x2

More ›