18 Matching Annotations
  1. Oct 2023
    1. This check is required for cases where the number of elements in an array is not evenly divisible by the thread block size, and as a result the number of threads launched by the kernel is larger than the array size

      This is confusing: how the hell can the number of THREADS be greater than the number os elements?

      I mean, unless there is at least one thread per element, this would make no sense. And, even though, this case is apparently absurd, because why would we need one thread PER ELEMENT at all?

    2. do not need to be specified as device variables

      How the hell do specify something as a "device variable"?

      Or does "device variable" just means that this variable is in the device memory? - This makes the most sense, and would be coherent with d_x and d_y becoming device variables after being copied to the device.

  2. Jan 2023
    1. y[i] = x[i] + y[i];

      How is that supposed to be correct? There is no way of knowing if the value of i will not overflow both x and y.

      cudaMallocManaged must be doing some kind of memory space management for this piece of code to be correct.

  3. Mar 2022
    1. Some claimed that under OOP writing tests is harder and it requires extra care to refactor

      Testing OO programs always require testing not only the target object, but also everything it is related to. This imposes some burden on TDD. Refactoring is not harder, but requires more work than usual, since one needs to refactor not only the target object, but everythin else too.

  4. Mar 2021
    1. In the original position, representative individuals are asked to deliberate behind a veil of ignorance about what principles of justice they would choose to regulate their social cooperation and competition

      There are assumptions about behavior. In this case, assume a definition of rationality as defined by Rawls. In the original position, people emulate each notion of justice, e.g., equality, utilitarian, etc.. This emulation leads to distributions/allocation of resources, as determined by each notion of justice. Them, they choose one of these distributions. According to the behavior assumed by Rawls, the notion chosen will be equality

  5. Feb 2021
    1. it's that the only way to understand politics is to embrace a variety of methodological approaches that can compensate for the complications of personality, culture, history, beliefs, and pure chance inherent when human beings vie for power

      Relativist yadda yadda

    2. The trouble with rational choice is not that it exists within the profession alongside other methodologies; the trouble with rational choice is that it is dominating the profession and displacing those other methodologies

      No, it is not

    3. Green and Shapiro then proceeded to apply their critical framework to other major works of rational choice theory on American politics. Time and again they came to the same conclusion: If rational choice theorists had bothered to conduct even rudimentary empirical testing of their elegant theories, they would have discovered that the theories simply don't describe the real world of politics

      Formal modeling is more about contrafactuals, i.e., possible worlds, than our own reality.

    4. But, Green and Shapiro asked, since people in other countries have been known to vote despite the risk of death as a consequence, how can it be said to be "low-cost"

      A battle for who is more nitpicker, RCT dudes or Green and Shapiro, ffs

    5. Political theorists had little patience for a theory that they believed reduced all human behavior to a set of interests and preferences independent of things such as values, culture, or history

      ZzZzZz

    6. and this was in no small part because the entire discipline of political science had developed a severe inferiority complex during the intervening years

      Agreed. Totally agreed

    7. They mostly found their way into second-tier institutions, many of which were best known as engineering schools, such as Cal Tech and Carnegie-Mellon

      Imagine living in an age where CalTech and Carnegie-Mellon are T2

    8. But beneath that question lurks a second issue more important to those of us outside the academy: whether political scientists have an obligation to do work that is not merely interesting as an intellectual enterprise but also helps us govern ourselves.

      This is actually a good comment. Finally

    9. The next generation of political scientists looks increasingly like Shepsle and less like Wilson

      Yeah, sure, the 2000-2010-2020 almost 90% dominance of comparative politics papers would disagree with that

    10. If you ask Shepsle, Krehbiel, or their fellow rational choicers how they've gotten so far so fast, they will tell you it's simply because they are that good—and because they are the only ones in the field who carry out work that qualifies as science

      This "article" is a fucking bravado against RCT for no reason, just because.

    11. The rational choicers believe their quest for universal and logically consistent theories makes them the only true practitioners of political science

      U angry bro? Why so much hate?