ryao

61 ms·

931.
by ryao 2y ago
> The inner core is usually thought to be shaped like a ball, but its edges may actually have deformed by 100m or more in height in places, according to Prof John Vidale who led the research. Is this not within the margin of error for ou
932.
by ryao 2y ago
Their promise not to train on the conversations seemed to imply that conversations were private. It is disconcerting that they were not. The extent to which you need to look for loopholes in privacy guarantees these days is amazing.
933.
by ryao 2y ago
The management engine is an embedded processor.
934.
by ryao 2y ago
The first link was supposed to be this: https://github.com/openzfs/zfs/commit/677c6f8457943fe5b56d7a...
935.
by ryao 2y ago
The programming model is that all threads in the warp / thread block run the same instruction (barring masking for branch divergence). Having SIMD instructions at the thread level is a rarity given that the way SIMD is implemented is a
936.
by ryao 2y ago
The PTX ISA for Nvidia GPUs supports 64-bit integer arithmetic: https://docs.nvidia.com/cuda/parallel-thread-execution/index... https://docs.nvidia.com/cuda/parallel-thread-execution/inde
937.
by ryao 2y ago
You can always have the compiler dump the assembly output so you can examine it. I suspect few do that.
938.
by ryao 2y ago
I had suggested the virtual instructions to the RISC-V developers to eliminate branchy cmov assembly, as I am not happy with it either. It is surprising to realize that x86 cores are not making more use of macro-ops fusion, contrary to my e
939.
by ryao 2y ago
For some reason, I feel like this is harder to implement than you expect. The way to find out would be to get a bunch of examples of people doing this “optimizations in shader code, look at the IR generated compared to the optimal version a
940.
by ryao 2y ago
> There's of course no "way" to check, as it's a microarchitectural property. Your best bet is comparing performance of the same code on predictable vs unpredictable branches. I was afraid the answer to my question wo
941.
by ryao 2y ago
You might want to give feedback to the risc-v developers (although it might be too late at this point). What is the way to check if implicit cmov instructions are implemented in the CPU instruction decoder? If AMD did not implement this in
942.
by ryao 2y ago
It just occurred to me that Clang and GCC did not necessarily fail to use conditional moves in your examples. While they failed to use explicit cmov instructions, cmp/jmp 1 instruction/mov is actually an idiom for an implicit cmov
943.
by ryao 2y ago
It depends on whether something runs fast enough. Certain software, such as filesystem drivers in the OS kernel, never is fast enough as there is always someone who wants it to run faster, use less memory, etcetera. Efforts to get it as clo
944.
by ryao 2y ago
You want to pass -mllvm -x86-cmov-converter=false. I assume LLVM has a pass to undo conditional moves on x86 whenever a heuristic determines skipping a calculation by branching is cheaper than doing the calculation and using a conditional m
945.
by ryao 2y ago
Shader compilers tend to be very latency sensitive, so “it takes too long to run” would be a valid reason why it is not done if it is not done.
946.
by ryao 2y ago
A microkernel only gives you scheduling, IPC and virtual memory. You and others get to implement everything else on top of that. That means there are no NUSE or FUSE interfaces and if you want them, you get to implement them as part of user
947.
by ryao 2y ago
Do shader compilers have optimization passes to undo this mistake and if not, could they be added?
948.
by ryao 2y ago
The author addressed that discrepancy: PG&E's current rate structure has urban rate payers subsidize rural rate payers and people who live in wildfire zones in e.g. the Orinda Hills, who need substantial investment in order to
949.
by ryao 2y ago
It depends on whether the software is performance critical. Many of the functions in the OS kernel are performance critical. In many cases in userspace, things are fast enough, rather than performance critical. This is why for example, Java
950.
by ryao 2y ago
I had forgotten that NeXTSTEP went back that far. Thanks for the correction.
951.
by ryao 2y ago
It depends on what part of the GPU driver crashes. The kernel mode part crashing caused countless BSODs on Vista. The switch to WDDM was painful for Windows. Things had improved by the time Windows 7 was made.
952.
by ryao 2y ago
But why linux won? We know now it won, but what is the reason. Tanenbaum was theoraticaly correct. İf HN exist back then, I would argue most devs here would say Minix will last longer, monolithics is indeed an old idea that had been tr
953.
by ryao 2y ago
Do you mean robustness for performance?
954.
by ryao 2y ago
I assume that funding ran out around 2014: https://blog.minix3.org/tag/news/
955.
by ryao 2y ago
While I could go into a long story here about the relative merits of the two designs, suffice it to say that among the people who actually design operating systems, the debate is essentially over. Microkernels have won. The deve
956.
by ryao 2y ago
I miss rote ECMAScript as EMCAScript. ECMA does not sound like a skin disease to me. EMCA might though.
957.
by ryao 2y ago
Here is a novel idea. Just pronounce ECMA as a 2 syllable word using classical Latin pronunciation. Ecma becomes just as catchy as Java when pronounced that way. For English speakers who do not know classical Latin, just read this: Eck-Ma-S
958.
by ryao 2y ago
ECMAScript is what people who want to avoid Oracle’s trademark say. Being anachronistic is irrevelant as long as Oracle has the trademark.
959.
by ryao 2y ago
I doubt they are using exponential time algorithms. They appear to be using Opus, which appears to be O(NlogN) worst case. Opus is also well known to be computationally cheap, like most (all?) audio codecs. It should run fine on the Pentium
960.
by ryao 2y ago
How is that relevant?

More ›