Insist on can see hope, encounter problems, sometimes is to die knock, can slowly see hope. Even a day of hope, despair, and so forth.
In the morning, stay in bed for one hours, still do not get up to exercise, if only want to, this one hours I have all day of things done. Talk about doing things in the morning. The main understanding of Intel's previous Tick-tock, processor update rhythm, that is, tick, a tick, mainly update the process, such as from 32nm to 22nm, a tock is the main update architecture, but to the post-Moorish era, it became a three-step strategy, tick, tock, optimization, such as the recent seven-generation core Kabylake, probably illustrates this.
In the afternoon, solved yesterday, the legacy of the problem, one is cuda asynchronous parallel, so that the CPU and GPU parallel execution. Also, the problem with the Intel Haswell microprocessor architecture, currently using a Haswell-based microprocessor i5-4590 to do the experimental platform, want to investigate the application of the Intel microprocessor architecture implementation process, as well as performance evaluation, good luck, No one has yet been found to be doing this thing. Never felt that the microarchitecture was so complicated that a cache conformance protocol had been engaged for an evening before the application scenario was understood. From the beginning of the afternoon, the study of the structure, the evening began to write a little text, finally completed the goal of 1500 words. However, it is clear that I do not know anything about the application of the implementation process on this architecture. This is the task of tomorrow. In general, the solution, the paper can go on to write down.
Completed today's task, is already nine points, it feels very relaxed, I want to investigate their interests have been interested, but there is no solution to the problem, about mobile phones, SOC chip, ISA, to the architecture, then to the operating system, application, implementation process. These I think all is ignorant, the best situation is, I found an article, can perfect answer my question, unfortunately this range is too wide, find such answer probability is too small, so, the answer still need to find. Like, I want to understand the general history of China, I can find the most perfect answer, is a very thick book, the answer still need me to find. Some problems, or need to make their own reasoning, and then learn from others.
In the evening, ran the step, today fog haze again serious, tomorrow do not know can not afford to come.
Dead knock, Dead knock