Hi people, when I launch a program on my Tesla S2050, I have a latency of 2 seconds just run the first function of CUDA, for example cudaMalloc(). If I launch the same program on GeForce 310, on my computer client, this latency there isn’t. Why?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| slowness of first cudaMalloc (K40 vs K20) | 0 | 1010 | October 28, 2015 | |
| slowness of first cudaMalloc (K40 vs K20) | 0 | 671 | October 28, 2015 | |
| slowness of first cudaMalloc (K40 vs K20) | 0 | 768 | October 28, 2015 | |
| Initial delay of 4 seonds on CUDA program before executing on a multi-GPU machine | 1 | 3331 | January 30, 2012 | |
| slowness of first cudaMalloc (K40 vs K20) | 2 | 852 | October 29, 2015 | |
| 5000ms for warm up? | 2 | 1757 | April 6, 2009 | |
| Tesla c2050 maximum runtime | 4 | 6087 | March 25, 2010 | |
| CUDA Driver Problems for Tesla S2050 with SLES 11 sp1 | 1 | 1117 | May 7, 2011 | |
| Trying to reduce delays between kernel launches | 0 | 6639 | January 4, 2011 | |
| Help! First cudaMalloc takes 10 seconds! | 8 | 1496 | February 11, 2012 |