GPU L3: Computation

00:11:10.256,00:11:13.256 Arihant Samar cs18b052: 'da' is equal to physical address in GPU or is page table used? 00:11:57.256,00:12:00.256 NIKAM ASHUTOSH SHASHIKANT ee16b143: Can we access da in 'main' e.g. da[i]? 00:12:03.067,00:12:06.067 Arihant Samar cs18b052: ok sir 00:13:06.962,00:13:09.962 NIKAM ASHUTOSH SHASHIKANT ee16b143: Thank you sir 00:16:13.329,00:16:16.329 Aditya Agrawal CS20S026: What is the difference between using cudaMalloc and cudaMemCpy from Host to Device? 00:16:16.839,00:16:19.839 NIKAM ASHUTOSH SHASHIKANT ee16b143: Sir, don't we need cudaSynchronize() here? 00:17:38.107,00:17:41.107 Nishant Prabhu me17b084: CudaMemcpy done specifically 00:18:20.154,00:18:23.154 Mr Rahul Mastram Verma CS20S038: Sir cudaMemCpy copy values ??? 00:22:37.617,00:22:40.617 Aditya Agrawal CS20S026: When you say cudaMemCpy is over, it means when N integers are copied from GPU to CPU in the example? 00:23:07.921,00:23:10.921 Vignesh S me17b078: Why isnt there a common memory accessible by both CPU and GPU since there is a lot of transfer of data between CPU and GPU in bigger processes ? shouldnt a common memory make the process faster ? 00:23:59.546,00:24:02.546 Rishabh C S ee17b067: @Vignesh, coherency and scalability would become an issue 00:24:04.642,00:24:07.642 Parth S. Shah ee17b059: Sir does the cudaMemcpy go into the queue of each thread? Or does each "store" instruction get assigned to a different thread? 00:31:43.719,00:31:46.719 Parth S. Shah ee17b059: So the cores don't handle this function but its the PCI express? 00:31:55.148,00:31:58.148 Parth S. Shah ee17b059: Thank you Sir 00:37:21.453,00:37:24.453 Akash Reddy A ee17b001: Sir could you please explain why the msg pointer points to the GPU in this case? 00:39:22.708,00:39:25.708 Akash Reddy A ee17b001: Got it sir, thank you 00:40:23.282,00:40:26.282 KARTHIK SURESH ee16b140: Are there CUDA capable iGPUs? 00:41:53.885,00:41:56.885 Mr Rahul Mastram Verma CS20S038: So now days do we have both integrated and discrete GPUs in laptop 00:41:58.548,00:42:01.548 Amalan S EE20D408: Nvidia Jetson 00:46:38.890,00:46:41.890 Parth S. Shah ee17b059: Sir this is mostly about the data. How are the instructions passed from CPU to GPU? Does that happen at compile time? 00:48:43.309,00:48:46.309 Parth S. Shah ee17b059: Hello World? 00:49:22.144,00:49:25.144 Rishabh C S ee17b067: +1 00:49:37.622,00:49:40.622 Suraj Kumar Rai cs20m068: +1 00:49:39.133,00:49:42.133 Shubham Mohan Randive cs20m064: Hello world 00:49:41.680,00:49:44.680 Mr Rahul Mastram Verma CS20S038: +1 00:49:47.150,00:49:50.150 Akash Haridas ae17b020: Each character is incremented by one 00:49:47.686,00:49:50.686 Mansi Choudhary ee17b053: +1 00:49:50.019,00:49:53.019 Jeswant Krishna ae17b030: +1